[gentoo-user] /dev/sda* missing at boot

[gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 1025 bytes --]

Hi, guys

It is a shame, I know, but after several years using Gentoo, it is the  
first time I try to build a kernel without "genkernel".

And now I can't boot to that new kernel, it does not find (and really do  
not have a) /dev/sda* root partition ("real-root"); during the boot it  
stops, complaining about that, gives me the option to get a shell, from  
which I am able to see that there is no /dev/sda* .

I have included everything SATA, so it looks like that is not a kernel  
problem, but a initramfs issue, I guess.

What am I missing?

Thanks a lot
Francisco

PS: my boot partition is sda2, sda3 is a swap partition, and everything  
else is in sda4. sda1 is not used (up to now) and this is my grub.conf :

title Gentoo Linux 2.6.39-gentoo-r3
root (hd0,1)
kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro root=/dev/ram0  
init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32  
nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3

[-- Attachment #2: Type: text/html, Size: 1280 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[András Csányi]

On 18 August 2011 18:59,  <frares@gmail.com> wrote:
> Hi, guys
>
> It is a shame, I know, but after several years using Gentoo, it is the first
> time I try to build a kernel without "genkernel".
>
> And now I can't boot to that new kernel, it does not find (and really do not
> have a) /dev/sda* root partition ("real-root"); during the boot it stops,
> complaining about that, gives me the option to get a shell, from which I am
> able to see that there is no /dev/sda* .
>
> I have included everything SATA, so it looks like that is not a kernel
> problem, but a initramfs issue, I guess.
>
> What am I missing?

Why have you choose this way? I mean, non-genkernel way.

-- 
- -
--  Csanyi Andras (Sayusi Ando)  -- http://sayusi.hu --
http://facebook.com/andras.csanyi
--  ""Trust in God and keep your gunpowder dry!" - Cromwell

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Aug 18, 2011 at 2:59 PM, <frares@gmail.com> wrote:
>
> Hi, guys
>
> It is a shame, I know, but after several years using Gentoo, it is the first time I try to build a kernel without "genkernel".
>
> And now I can't boot to that new kernel, it does not find (and really do not have a) /dev/sda* root partition ("real-root"); during the boot it stops, complaining about that, gives me the option to get a shell, from which I am able to see that there is no /dev/sda* .
>
> I have included everything SATA, so it looks like that is not a kernel problem, but a initramfs issue, I guess.

If you've got a SATA controller, no frills, then all you *really* need
is AHCI. Build that into your kernel if you're worried about having
the right modules in initramfs. You can break it out into a module
later if you like. Opinions differ as to how much stuff should be
broken into modules vs being built-in to the kernel. I tend to build
in everything absolutely needed for boot, myself. Some people build in
just about everything, and some people build in almost nothing.
There's no "right" way for every use case.

Also, check your BIOS to see if it's running your SATA controller in
some kind of IDE emulation mode. If it is, disable that. (Some
motherboards let you choose between "IDE" and "RAID", where "RAID" is
AHCI mode. Others call IDE mode 'legacy', and still others might
actually call the AHCI mode 'AHCI')

Motherboards running SATA controllers in IDE emulation mode is an
incredibly common thing:

17:18 <@IRule> beh
17:18 <@IRule> hda1 turned into sda1
17:19 < shortcircuit> IRule: Turn SCSI-generic support, or did you
switch from legacy to AHCI in your BIOS?
17:20 <@IRule> shortcircuit: quiet, you

--
:wq

Re: Re: [gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 1217 bytes --]

Em 18/08/2011 16:08, András Csányi <sayusi.ando@gmail.com> escreveu:
> On 18 August 2011 18:59, frares@gmail.com> wrote:

> > Hi, guys
> >
> > It is a shame, I know, but after several years using Gentoo, it is the  
> first
> > time I try to build a kernel without "genkernel".
> >
> > And now I can't boot to that new kernel, it does not find (and really  
> do not
> > have a) /dev/sda* root partition ("real-root"); during the boot it  
> stops,
> > complaining about that, gives me the option to get a shell, from which  
> I am
> > able to see that there is no /dev/sda* .
> >
> > I have included everything SATA, so it looks like that is not a kernel
> > problem, but a initramfs issue, I guess.
> >
> > What am I missing?


> Why have you choose this way? I mean, non-genkernel way.


> --
> - -
> -- Csanyi Andras (Sayusi Ando) -- http://sayusi.hu --
> http://facebook.com/andras.csanyi
> -- ""Trust in God and keep your gunpowder dry!" - Cromwell



That's recommended in the new install manual:

http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?style=printable&full=1#book_part1_chap7

Look for item "7c". The alternative way is to use genkernel.

Francisco

[-- Attachment #2: Type: text/html, Size: 1541 bytes --]

Re: Re: [gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 1857 bytes --]

Em 18/08/2011 16:13, Michael Mol <mikemol@gmail.com> escreveu:
> On Thu, Aug 18, 2011 at 2:59 PM, frares@gmail.com> wrote:
> >
> > Hi, guys
> >
> > It is a shame, I know, but after several years using Gentoo, it is the  
> first time I try to build a kernel without "genkernel".
> >
> > And now I can't boot to that new kernel, it does not find (and really  
> do not have a) /dev/sda* root partition ("real-root"); during the boot it  
> stops, complaining about that, gives me the option to get a shell, from  
> which I am able to see that there is no /dev/sda* .
> >
> > I have included everything SATA, so it looks like that is not a kernel  
> problem, but a initramfs issue, I guess.

> If you've got a SATA controller, no frills, then all you *really* need
> is AHCI. Build that into your kernel if you're worried about having
> the right modules in initramfs. You can break it out into a module
> later if you like. Opinions differ as to how much stuff should be
> broken into modules vs being built-in to the kernel. I tend to build
> in everything absolutely needed for boot, myself. Some people build in
> just about everything, and some people build in almost nothing.
> There's no "right" way for every use case.

> Also, check your BIOS to see if it's running your SATA controller in
> some kind of IDE emulation mode. If it is, disable that. (Some
> motherboards let you choose between "IDE" and "RAID", where "RAID" is
> AHCI mode. Others call IDE mode 'legacy', and still others might
> actually call the AHCI mode 'AHCI')

> Motherboards running SATA controllers in IDE emulation mode is an
> incredibly common thing:

> 17:18 beh
> 17:18 hda1 turned into sda1
> 17:19 IRule: Turn SCSI-generic support, or did you
> switch from legacy to AHCI in your BIOS?
> 17:20 shortcircuit: quiet, you

> --

> :wq


Thanks, gonna try it.

Francisco

[-- Attachment #2: Type: text/html, Size: 2307 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Matthew Finkel]

[-- Attachment #1: Type: text/plain, Size: 1508 bytes --]

On Thu, Aug 18, 2011 at 2:59 PM, <frares@gmail.com> wrote:

> Hi, guys
>
> It is a shame, I know, but after several years using Gentoo, it is the
> first time I try to build a kernel without "genkernel".
>
> And now I can't boot to that new kernel, it does not find (and really do
> not have a) /dev/sda* root partition ("real-root"); during the boot it
> stops, complaining about that, gives me the option to get a shell, from
> which I am able to see that there is no /dev/sda* .
>
> I have included everything SATA, so it looks like that is not a kernel
> problem, but a initramfs issue, I guess.
>
> What am I missing?
>
> Thanks a lot
> Francisco
>
> P.S.: my boot partition is sda2, sda3 is a swap partition, and everything
> else is in sda4. sda1 is not used (up to now) and this is my grub.conf :
>
> title Gentoo Linux 2.6.39-gentoo-r3
> root (hd0,1)
> kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro root=/dev/ram0
> init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3



Do you have a block device driver built into the kernel? And what type of
shell are you dropped into when then happens? Is it a single-user mode shell
or grub (or something else entirely)? Also, while you're booted into the
livecd/dvd/usb and you chroot, try lspci -k and check to see what
modules/drivers that lists as installed and see if you have them enabled in
your config.

- Matt

[-- Attachment #2: Type: text/html, Size: 1966 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Mark Knecht]

On Thu, Aug 18, 2011 at 11:59 AM,  <frares@gmail.com> wrote:
> Hi, guys
>
> It is a shame, I know, but after several years using Gentoo, it is the first
> time I try to build a kernel without "genkernel".
>
> And now I can't boot to that new kernel, it does not find (and really do not
> have a) /dev/sda* root partition ("real-root"); during the boot it stops,
> complaining about that, gives me the option to get a shell, from which I am
> able to see that there is no /dev/sda* .
>
> I have included everything SATA, so it looks like that is not a kernel
> problem, but a initramfs issue, I guess.
>
> What am I missing?
>
> Thanks a lot
> Francisco
>
> P.S.: my boot partition is sda2, sda3 is a swap partition, and everything
> else is in sda4. sda1 is not used (up to now) and this is my grub.conf :
>
> title Gentoo Linux 2.6.39-gentoo-r3
> root (hd0,1)
> kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro root=/dev/ram0
> init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3

Maybe I'm missing the obvious here but have you taken a copy of
whatever config file was used/generated by genkernel and used that as
a jumping off point for building your own kernel. kernel's a kernel's
a kernel. What it is capable of doing is in the .config file. If
genkernel doesn't give you a .config file - I've never used genkernel
so I don't know what it does - then assuming you have the feature
turned on you can get the running config using zcat /proc/config.gz.
Save that to a new .config file, put it in the kernel source directory
and you should be good to go.

You can also use zcat /proc/config.gz on the install CD kernel if yuo
boot from that. Save it to a disk and use it as the basis for creating
your own config.

HTH,
Mark

Re: [gentoo-user] /dev/sda* missing at boot

[Graham Murray]

Michael Mol <mikemol@gmail.com> writes:

> Also, check your BIOS to see if it's running your SATA controller in
> some kind of IDE emulation mode. If it is, disable that. (Some
> motherboards let you choose between "IDE" and "RAID", where "RAID" is
> AHCI mode. Others call IDE mode 'legacy', and still others might
> actually call the AHCI mode 'AHCI')

That is if the BIOS will allow you to do so. Some BIOSes, for example
some Dell servers, will only run SATA in emulation mode despite the
chipset supporting AHCI. The only option they give for SATA is
enable/disable.

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 2256 bytes --]

On Friday 19 Aug 2011 03:27:23 Mark Knecht wrote:
> On Thu, Aug 18, 2011 at 11:59 AM,  <frares@gmail.com> wrote:
> > Hi, guys
> > 
> > It is a shame, I know, but after several years using Gentoo, it is the
> > first time I try to build a kernel without "genkernel".
> > 
> > And now I can't boot to that new kernel, it does not find (and really do
> > not have a) /dev/sda* root partition ("real-root"); during the boot it
> > stops, complaining about that, gives me the option to get a shell, from
> > which I am able to see that there is no /dev/sda* .
> > 
> > I have included everything SATA, so it looks like that is not a kernel
> > problem, but a initramfs issue, I guess.
> > 
> > What am I missing?
> > 
> > Thanks a lot
> > Francisco
> > 
> > P.S.: my boot partition is sda2, sda3 is a swap partition, and everything
> > else is in sda4. sda1 is not used (up to now) and this is my grub.conf :
> > 
> > title Gentoo Linux 2.6.39-gentoo-r3
> > root (hd0,1)
> > kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro root=/dev/ram0
> > init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> > nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> > initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3
> 
> Maybe I'm missing the obvious here but have you taken a copy of
> whatever config file was used/generated by genkernel and used that as
> a jumping off point for building your own kernel. kernel's a kernel's
> a kernel. What it is capable of doing is in the .config file. If
> genkernel doesn't give you a .config file - I've never used genkernel
> so I don't know what it does - then assuming you have the feature
> turned on you can get the running config using zcat /proc/config.gz.
> Save that to a new .config file, put it in the kernel source directory
> and you should be good to go.
> 
> You can also use zcat /proc/config.gz on the install CD kernel if yuo
> boot from that. Save it to a disk and use it as the basis for creating
> your own config.

If you no longer use genkernel it is likely that you do not need an initram.  
Build chipset and fs modules into the kernel.  Other drivers you can choose if 
you want to build as modules. 

-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[gentoo-user] Re: /dev/sda* missing at boot

[Nikos Chantziaras]

On 08/18/2011 10:08 PM, András Csányi wrote:
> On 18 August 2011 18:59,<frares@gmail.com>  wrote:
>> Hi, guys
>>
>> It is a shame, I know, but after several years using Gentoo, it is the first
>> time I try to build a kernel without "genkernel".
>>
>> And now I can't boot to that new kernel, it does not find (and really do not
>> have a) /dev/sda* root partition ("real-root"); during the boot it stops,
>> complaining about that, gives me the option to get a shell, from which I am
>> able to see that there is no /dev/sda* .
>>
>> I have included everything SATA, so it looks like that is not a kernel
>> problem, but a initramfs issue, I guess.
>>
>> What am I missing?
>
> Why have you choose this way? I mean, non-genkernel way.

genkernel generates generic (bloated) kernels.

Re: Re: [gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 2190 bytes --]

Em 18/08/2011 23:27, Mark Knecht <markknecht@gmail.com> escreveu:
> On Thu, Aug 18, 2011 at 11:59 AM, frares@gmail.com> wrote:
> > Hi, guys
> >
> > It is a shame, I know, but after several years using Gentoo, it is the  
> first
> > time I try to build a kernel without "genkernel".
> >
> > And now I can't boot to that new kernel, it does not find (and really  
> do not
> > have a) /dev/sda* root partition ("real-root"); during the boot it  
> stops,
> > complaining about that, gives me the option to get a shell, from which  
> I am
> > able to see that there is no /dev/sda* .
> >
> > I have included everything SATA, so it looks like that is not a kernel
> > problem, but a initramfs issue, I guess.
> >
> > What am I missing?
> >
> > Thanks a lot
> > Francisco
> >
> > PS: my boot partition is sda2, sda3 is a swap partition, and everything
> > else is in sda4. sda1 is not used (up to now) and this is my grub.conf :
> >
> > title Gentoo Linux 2.6.39-gentoo-r3
> > root (hd0,1)
> > kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro root=/dev/ram0
> > init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> > nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> > initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3

> Maybe I'm missing the obvious here but have you taken a copy of
> whatever config file was used/generated by genkernel and used that as
> a jumping off point for building your own kernel. kernel's a kernel's
> a kernel. What it is capable of doing is in the .config file. If
> genkernel doesn't give you a .config file - I've never used genkernel
> so I don't know what it does - then assuming you have the feature
> turned on you can get the running config using zcat /proc/config.gz.
> Save that to a new .config file, put it in the kernel source directory
> and you should be good to go.

> You can also use zcat /proc/config.gz on the install CD kernel if yuo
> boot from that. Save it to a disk and use it as the basis for creating
> your own config.

> HTH,
> Mark



That's what I am doing right now. I am using genkernel to have something to  
boot on. Then I will try to find a way to optimize another one.

Thanks
Francisco

[-- Attachment #2: Type: text/html, Size: 2727 bytes --]

Re: Re: [gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 2612 bytes --]

Em 19/08/2011 07:09, Mick <michaelkintzios@gmail.com> escreveu:
> On Friday 19 Aug 2011 03:27:23 Mark Knecht wrote:
> > On Thu, Aug 18, 2011 at 11:59 AM, frares@gmail.com> wrote:
> > > Hi, guys
> > >
> > > It is a shame, I know, but after several years using Gentoo, it is the
> > > first time I try to build a kernel without "genkernel".
> > >
> > > And now I can't boot to that new kernel, it does not find (and really  
> do
> > > not have a) /dev/sda* root partition ("real-root"); during the boot it
> > > stops, complaining about that, gives me the option to get a shell,  
> from
> > > which I am able to see that there is no /dev/sda* .
> > >
> > > I have included everything SATA, so it looks like that is not a kernel
> > > problem, but a initramfs issue, I guess.
> > >
> > > What am I missing?
> > >
> > > Thanks a lot
> > > Francisco
> > >
> > > PS: my boot partition is sda2, sda3 is a swap partition, and  
> everything
> > > else is in sda4. sda1 is not used (up to now) and this is my  
> grub.conf :
> > >
> > > title Gentoo Linux 2.6.39-gentoo-r3
> > > root (hd0,1)
> > > kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro  
> root=/dev/ram0
> > > init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> > > nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> > > initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3
> >
> > Maybe I'm missing the obvious here but have you taken a copy of
> > whatever config file was used/generated by genkernel and used that as
> > a jumping off point for building your own kernel. kernel's a kernel's
> > a kernel. What it is capable of doing is in the .config file. If
> > genkernel doesn't give you a .config file - I've never used genkernel
> > so I don't know what it does - then assuming you have the feature
> > turned on you can get the running config using zcat /proc/config.gz.
> > Save that to a new .config file, put it in the kernel source directory
> > and you should be good to go.
> >
> > You can also use zcat /proc/config.gz on the install CD kernel if yuo
> > boot from that. Save it to a disk and use it as the basis for creating
> > your own config.

> If you no longer use genkernel it is likely that you do not need an  
> initram.
> Build chipset and fs modules into the kernel. Other drivers you can  
> choose if
> you want to build as modules.

> --
> Regards,
> Mick


I the case I don't need a initram, I guess that the grub line for parameter  
passing to the kernel would be empty. Am I wrong?

I was just looking on how to build my own initram. What is it supposed to  
do anyway?

Thanks
Francisco

[-- Attachment #2: Type: text/html, Size: 3309 bytes --]

Re: Re: [gentoo-user] /dev/sda* missing at boot

[Gregory Woodbury]

[-- Attachment #1: Type: text/plain, Size: 3382 bytes --]

On Fri, Aug 19, 2011 at 9:12 AM, <frares@gmail.com> wrote:

> Em 19/08/2011 07:09, Mick <michaelkintzios@gmail.com> escreveu:
> > On Friday 19 Aug 2011 03:27:23 Mark Knecht wrote:
>
> > > On Thu, Aug 18, 2011 at 11:59 AM,  frares@gmail.com> wrote:
> > > > Hi, guys
> > > >
> > > > It is a shame, I know, but after several years using Gentoo, it is
> the
> > > > first time I try to build a kernel without "genkernel".
> > > >
> > > > And now I can't boot to that new kernel, it does not find (and really
> do
> > > > not have a) /dev/sda* root partition ("real-root"); during the boot
> it
> > > > stops, complaining about that, gives me the option to get a shell,
> from
> > > > which I am able to see that there is no /dev/sda* .
> > > >
> > > > I have included everything SATA, so it looks like that is not a
> kernel
> > > > problem, but a initramfs issue, I guess.
> > > >
> > > > What am I missing?
> > > >
> > > > Thanks a lot
> > > > Francisco
> > > >
> > > > P.S.: my boot partition is sda2, sda3 is a swap partition, and
> everything
> > > > else is in sda4. sda1 is not used (up to now) and this is my
> grub.conf :
> > > >
> > > > title Gentoo Linux 2.6.39-gentoo-r3
> > > > root (hd0,1)
> > > > kernel /boot/kernel-genkernel-x86_64-2.6.39-gentoo-r3 ro
> root=/dev/ram0
> > > > init=/linuxrc real_root=/dev/sda4 vga=0x318 video=uvesafb:1024x768-32
> > > > nodevfs udev devfs=nomount quiet CONSOLE=/dev/tty1
> > > > initrd /boot/initramfs-genkernel-x86_64-2.6.39-gentoo-r3
> > >
> > > Maybe I'm missing the obvious here but have you taken a copy of
> > > whatever config file was used/generated by genkernel and used that as
> > > a jumping off point for building your own kernel. kernel's a kernel's
> > > a kernel. What it is capable of doing is in the .config file. If
> > > genkernel doesn't give you a .config file - I've never used genkernel
> > > so I don't know what it does - then assuming you have the feature
> > > turned on you can get the running config using zcat /proc/config.gz.
> > > Save that to a new .config file, put it in the kernel source directory
> > > and you should be good to go.
> > >
> > > You can also use zcat /proc/config.gz on the install CD kernel if yuo
> > > boot from that. Save it to a disk and use it as the basis for creating
> > > your own config.
> >
> > If you no longer use genkernel it is likely that you do not need an
> initram.
> > Build chipset and fs modules into the kernel.  Other drivers you can
> choose if
> > you want to build as modules.
>
> I the case I don't need a initram, I guess that the grub line for parameter
> passing to the kernel would be empty. Am I wrong?
>
> I was just looking on how to build my own initram. What is it supposed to
> do anyway?
>

The initramfs is a container for modules and stuff need to bring up the
system before the mounts of
/ and /boot.    If all the drivers are built-in to the kernel (or at least
the minimum required drivers are built-in)
then the initramfs isn't necessary.

Passing parameters to the kernel is a different issue entirely.

My grub.conf line is:

                kernel /vmlinuz-3.0.3-gentoo root=/dev/sda2
pata_it821x.noraid=1

with the pata_it821x driver built-in for the kenel to find a set of older
IDE drives on the IT8212 card I have installed.

IIRC the initramfs is built with the mkinitrd command.  I haven't had to use
it so I could be wrong.

[-- Attachment #2: Type: text/html, Size: 4300 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri 19 August 2011 13:12:25 frares@gmail.com did opine thusly:
> I the case I don't need a initram, I guess that the grub line for
> parameter   passing to the kernel would be empty. Am I wrong?

Yes.

Using or not using kernel parameters has nothing to do with whether 
you use an initramfs or not.

It's the initrd line in grub you do not need.

> 
> I was just looking on how to build my own initram. What is it
> supposed to   do anyway?

First, it's initramfs (the previous incarnation was initrd). You 
should use the correct name.

An initramfs is a filesystems image stored on disk in a place that 
grub can find. It contains a kernel, essential drivers and other bits 
and pieces. When booting, grub finds the image, bangs it into memory 
and instructs the cpu to start executing at a known point.

Why is this useful?

For Gentoo it usually isn't (there are times when it is - see below). 
Binary distros like Ubuntu and Fedora absolutely require this. These 
distros do not know what hardware you have and what drivers you 
require, so they supply drivers for everything. But Ubuntu cannot 
possibly compile into the kernel every possible driver you might need 
to boot as the list would be huge (every known floppy, CD, USB, every 
known MFM, IDE, SATA, SCSI, netboot, Fibre and more driver for a 
start), so what they do instead is probe the hardware at boot time, 
find out what you have, and load the driver modules you DO need.

This is the problem. The kernel wants to load disk drivers so that it 
can access the disk and continue booting. Where are the drivers? Well, 
they are on the disk. Oops, circular problem.

The difficulty is not finding and loading drivers, it's how do you get 
the disk driver off the disk before you have the disk driver in 
memory? (think chicken and egg here).

An initramfs solves this nicely. Grub shoved a disk image into memory 
when it booted. The kernel knows how to access it's memory it doesn't 
need a driver for that. And now the files containing the needed 
drivers are on a virtual disk *in memory*. The kernel loads them, and 
can now access the real physical disks.

Lots more complicated stuff then happens, like getting rid of the 
virtual filesystem from the initramfs and mounting the real filesystem 
from disk at /, but that's beyond the scope of this mail.

Gentoo mostly doesn't need any of this because you do know your 
hardware and can just compile your disk drivers into the kernel - this 
is the very thing that Ubuntu cannot do.

Some Gentoo users still need an initramfs, such as booting off drives 
in a RAID configuration. They need the RAID drivers first to read the 
disks so use an initramfs to fix this little problem exactly as Ubunut 
fixes their problem.

Make sense?


-- 
alan dot mckinnon at gmail dot com

Re: Re: [gentoo-user] /dev/sda* missing at boot

[frares]

[-- Attachment #1: Type: text/plain, Size: 3090 bytes --]

Em 19/08/2011 10:48, Alan McKinnon <alan.mckinnon@gmail.com> escreveu:
> On Fri 19 August 2011 13:12:25 frares@gmail.com did opine thusly:

> > I the case I don't need a initram, I guess that the grub line for
> > parameter passing to the kernel would be empty. Am I wrong?

> Yes.

> Using or not using kernel parameters has nothing to do with whether
> you use an initramfs or not.

> It's the initrd line in grub you do not need.

> >
> > I was just looking on how to build my own initram. What is it
> > supposed to do anyway?

> First, it's initramfs (the previous incarnation was initrd). You
> should use the correct name.

> An initramfs is a filesystems image stored on disk in a place that
> grub can find. It contains a kernel, essential drivers and other bits
> and pieces. When booting, grub finds the image, bangs it into memory
> and instructs the cpu to start executing at a known point.

> Why is this useful?

> For Gentoo it usually isn't (there are times when it is - see below).
> Binary distros like Ubuntu and Fedora absolutely require this. These
> distros do not know what hardware you have and what drivers you
> require, so they supply drivers for everything. But Ubuntu cannot
> possibly compile into the kernel every possible driver you might need
> to boot as the list would be huge (every known floppy, CD, USB, every
> known MFM, IDE, SATA, SCSI, netboot, Fibre and more driver for a
> start), so what they do instead is probe the hardware at boot time,
> find out what you have, and load the driver modules you DO need.

> This is the problem. The kernel wants to load disk drivers so that it
> can access the disk and continue booting. Where are the drivers? Well,
> they are on the disk. Oops, circular problem.

> The difficulty is not finding and loading drivers, it's how do you get
> the disk driver off the disk before you have the disk driver in
> memory? (think chicken and egg here).

> An initramfs solves this nicely. Grub shoved a disk image into memory
> when it booted. The kernel knows how to access it's memory it doesn't
> need a driver for that. And now the files containing the needed
> drivers are on a virtual disk *in memory*. The kernel loads them, and
> can now access the real physical disks.

> Lots more complicated stuff then happens, like getting rid of the
> virtual filesystem from the initramfs and mounting the real filesystem
> from disk at /, but that's beyond the scope of this mail.

> Gentoo mostly doesn't need any of this because you do know your
> hardware and can just compile your disk drivers into the kernel - this
> is the very thing that Ubuntu cannot do.

> Some Gentoo users still need an initramfs, such as booting off drives
> in a RAID configuration. They need the RAID drivers first to read the
> disks so use an initramfs to fix this little problem exactly as Ubunut
> fixes their problem.

> Make sense?


Completely! Thanks a lot.

So I guess that my problem is to find an appropriate pair of driver and  
hard disk operating mode.

> --
> alan dot mckinnon at gmail dot com


Thanks again
Francisco

[-- Attachment #2: Type: text/html, Size: 3768 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri 19 August 2011 15:06:46 frares@gmail.com did opine thusly:
> > Some Gentoo users still need an initramfs, such as booting off
> > drives in a RAID configuration. They need the RAID drivers
> > first to read the disks so use an initramfs to fix this little
> > problem exactly as Ubunut fixes their problem.
> > 
> > Make sense?
> 
> Completely! Thanks a lot.
> 
> So I guess that my problem is to find an appropriate pair of driver
> and   hard disk operating mode.

That's right. Don't forget the filesystem drivers. For example if you 
mount an ext3 partition at / and it's on a SATA drive, then you need 
SATA and ext3 drivers in the kernel (not as modules). 


-- 
alan dot mckinnon at gmail dot com

[gentoo-user] Re: /dev/sda* missing at boot

[Francesco Talamona]

On Friday 19 August 2011, Nikos Chantziaras wrote:
> On 08/18/2011 10:08 PM, András Csányi wrote:
> > On 18 August 2011 18:59,<frares@gmail.com>  wrote:
> >> Hi, guys
> >> 
> >> It is a shame, I know, but after several years using Gentoo, it is
> >> the first time I try to build a kernel without "genkernel".
> >> 
> >> And now I can't boot to that new kernel, it does not find (and
> >> really do not have a) /dev/sda* root partition ("real-root");
> >> during the boot it stops, complaining about that, gives me the
> >> option to get a shell, from which I am able to see that there is
> >> no /dev/sda* .
> >> 
> >> I have included everything SATA, so it looks like that is not a
> >> kernel problem, but a initramfs issue, I guess.
> >> 
> >> What am I missing?
> > 
> > Why have you choose this way? I mean, non-genkernel way.
> 
> genkernel generates generic (bloated) kernels.

This is a generalization, not entirely true:

genkernel --no-clean --no-mrproper --kerneldir=blabla all

With the above command, for example, you can provide your own .config 
and genkernel will do exactly as you wish.

Cheers
	Francesco


-- 
Linux Version 3.0.0-gentoo, Compiled #3 SMP PREEMPT Fri Aug 5 21:02:22 
CEST 2011
Two 1GHz AMD Athlon 64 X2 Processors, 4GB RAM, 4021.84 Bogomips Total
aemaeth

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Gregory Woodbury wrote:
>
> The initramfs is a container for modules and stuff need to bring up 
> the system before the mounts of
> / and /boot.    If all the drivers are built-in to the kernel (or at 
> least the minimum required drivers are built-in)
> then the initramfs isn't necessary.
>
> Passing parameters to the kernel is a different issue entirely.
>
> My grub.conf line is:
>
>                 kernel /vmlinuz-3.0.3-gentoo root=/dev/sda2 
> pata_it821x.noraid=1
>
> with the pata_it821x driver built-in for the kenel to find a set of 
> older IDE drives on the IT8212 card I have installed.
>
> IIRC the initramfs is built with the mkinitrd command.  I haven't had 
> to use it so I could be wrong.

Update with new info.  With udev needing some things in /usr, and /var, 
you will need a init* if /usr and /var is not on / in the near future.  
Yea, real neat.  Some need it already just depends on what is installed 
from what I read.

Dale

:-)  ;-)

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1290 bytes --]

On Friday 19 Aug 2011 23:08:06 Dale wrote:
> Gregory Woodbury wrote:
> > The initramfs is a container for modules and stuff need to bring up
> > the system before the mounts of
> > / and /boot.    If all the drivers are built-in to the kernel (or at
> > least the minimum required drivers are built-in)
> > then the initramfs isn't necessary.
> > 
> > Passing parameters to the kernel is a different issue entirely.
> > 
> > My grub.conf line is:
> >                 kernel /vmlinuz-3.0.3-gentoo root=/dev/sda2
> > 
> > pata_it821x.noraid=1
> > 
> > with the pata_it821x driver built-in for the kenel to find a set of
> > older IDE drives on the IT8212 card I have installed.
> > 
> > IIRC the initramfs is built with the mkinitrd command.  I haven't had
> > to use it so I could be wrong.
> 
> Update with new info.  With udev needing some things in /usr, and /var,
> you will need a init* if /usr and /var is not on / in the near future.
> Yea, real neat.  Some need it already just depends on what is installed
> from what I read.

Give us a link please Dale.

2/3 of my boxen have both /usr and/var on separate partitions and I never had 
to use initramfs (other than boot splash - or whatever it happens to be called 
this month).

-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Mick wrote:
> On Friday 19 Aug 2011 23:08:06 Dale wrote:
>    
>> Gregory Woodbury wrote:
>>      
>>> The initramfs is a container for modules and stuff need to bring up
>>> the system before the mounts of
>>> / and /boot.    If all the drivers are built-in to the kernel (or at
>>> least the minimum required drivers are built-in)
>>> then the initramfs isn't necessary.
>>>
>>> Passing parameters to the kernel is a different issue entirely.
>>>
>>> My grub.conf line is:
>>>                  kernel /vmlinuz-3.0.3-gentoo root=/dev/sda2
>>>
>>> pata_it821x.noraid=1
>>>
>>> with the pata_it821x driver built-in for the kenel to find a set of
>>> older IDE drives on the IT8212 card I have installed.
>>>
>>> IIRC the initramfs is built with the mkinitrd command.  I haven't had
>>> to use it so I could be wrong.
>>>        
>> Update with new info.  With udev needing some things in /usr, and /var,
>> you will need a init* if /usr and /var is not on / in the near future.
>> Yea, real neat.  Some need it already just depends on what is installed
>> from what I read.
>>      
> Give us a link please Dale.
>
> 2/3 of my boxen have both /usr and/var on separate partitions and I never had
> to use initramfs (other than boot splash - or whatever it happens to be called
> this month).
>
>    

It was discussed on -dev so far.  This is the subject line:

"Warn users not to do separate /usr partition without proper initramfs 
in the handbook?"

I think it will apply to /var to at some point.  I think it sucks.  I 
have /var on a separate partition and want to put /usr on one to but not 
now.

I think it can be found on gmane.com.  Again, it is on -dev and yes I 
raised my objections to this but it is UPSTREAM from Gentoo.  Dang 
Fedora or something.  ;-)

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat 20 August 2011 02:17:06 Dale did opine thusly:
> >> Update with new info.  With udev needing some things in /usr,
> >> and /var, you will need a init* if /usr and /var is not on /
> >> in the near future. Yea, real neat.  Some need it already
> >> just depends on what is installed from what I read.
> >>
> >>      
> >>
> > Give us a link please Dale.
> > 
> > 2/3 of my boxen have both /usr and/var on separate partitions
> > and I never had to use initramfs (other than boot splash - or
> > whatever it happens to be called this month).
> >
> >    
> 
> It was discussed on -dev so far.  This is the subject line:
> 
> "Warn users not to do separate /usr partition without proper
> initramfs  in the handbook?"
> 
> I think it will apply to /var to at some point.  I think it
> sucks.  I  have /var on a separate partition and want to put /usr
> on one to but not now.

Eh? That's fucking braindead. It also violates everything udev ever 
intended to do.

/usr and /var on separate partitions, plus a custom kernel without an 
initramfs is *exactly* the most common use case for Gentoo....


-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> On Sat 20 August 2011 02:17:06 Dale did opine thusly:
>    
>> It was discussed on -dev so far.  This is the subject line:
>>
>> "Warn users not to do separate /usr partition without proper
>> initramfs  in the handbook?"
>>
>> I think it will apply to /var to at some point.  I think it
>> sucks.  I  have /var on a separate partition and want to put /usr
>> on one to but not now.
>>      
> Eh? That's fucking braindead. It also violates everything udev ever
> intended to do.
>
> /usr and /var on separate partitions, plus a custom kernel without an
> initramfs is *exactly* the most common use case for Gentoo....
>
>
>    

I wish you could convince the devs of that.  I already have /var on its 
own and was planning to put /usr on its own.  I'm not now tho.  Looks 
like /, /boot, /home and that's it for the OS part.  It downright sucks.

You planning to explain this to the devs?  Maybe you will have better 
luck.  They said it is a dev from Fedora that started this . . . . 
crap.  ;-)

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat 20 August 2011 03:48:18 Dale did opine thusly:
> Alan McKinnon wrote:
> > On Sat 20 August 2011 02:17:06 Dale did opine thusly:
> >> It was discussed on -dev so far.  This is the subject line:
> >> 
> >> "Warn users not to do separate /usr partition without proper
> >> initramfs  in the handbook?"
> >> 
> >> I think it will apply to /var to at some point.  I think it
> >> sucks.  I  have /var on a separate partition and want to put
> >> /usr on one to but not now.
> > 
> > Eh? That's fucking braindead. It also violates everything udev
> > ever intended to do.
> > 
> > /usr and /var on separate partitions, plus a custom kernel
> > without an initramfs is *exactly* the most common use case for
> > Gentoo....
> I wish you could convince the devs of that.  I already have /var on
> its own and was planning to put /usr on its own.  I'm not now tho. 
> Looks like /, /boot, /home and that's it for the OS part.  It
> downright sucks.
> 
> You planning to explain this to the devs?  Maybe you will have
> better luck.  They said it is a dev from Fedora that started this .
> . . . crap.  ;-)


I'll spend some time I don't have reading the archives, then see.


-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] /dev/sda* missing at boot

[Gregory Woodbury]

[-- Attachment #1: Type: text/plain, Size: 712 bytes --]

On Sat, Aug 20, 2011 at 4:48 AM, Dale <rdalek1967@gmail.com> wrote:


> I wish you could convince the devs of that.  I already have /var on its own
> and was planning to put /usr on its own.  I'm not now tho.  Looks like /,
> /boot, /home and that's it for the OS part.  It downright sucks.
>
> You planning to explain this to the devs?  Maybe you will have better luck.
>  They said it is a dev from Fedora that started this . . . . crap.  ;-)


What's more, the Fedora dev who did this sh** believes that it's a
"religious issue" and refuses to discuss it. (personal experience)
He ignores the historical reasons, the advantages of separate partitions,
and even signed off on a bugzilla discussion as WONTFIX.

[-- Attachment #2: Type: text/html, Size: 1068 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1525 bytes --]

On Saturday 20 Aug 2011 09:57:46 Alan McKinnon wrote:
> On Sat 20 August 2011 03:48:18 Dale did opine thusly:
> > Alan McKinnon wrote:
> > > On Sat 20 August 2011 02:17:06 Dale did opine thusly:
> > >> It was discussed on -dev so far.  This is the subject line:
> > >> 
> > >> "Warn users not to do separate /usr partition without proper
> > >> initramfs  in the handbook?"
> > >> 
> > >> I think it will apply to /var to at some point.  I think it
> > >> sucks.  I  have /var on a separate partition and want to put
> > >> /usr on one to but not now.
> > > 
> > > Eh? That's fucking braindead. It also violates everything udev
> > > ever intended to do.
> > > 
> > > /usr and /var on separate partitions, plus a custom kernel
> > > without an initramfs is *exactly* the most common use case for
> > > Gentoo....
> > 
> > I wish you could convince the devs of that.  I already have /var on
> > its own and was planning to put /usr on its own.  I'm not now tho.
> > Looks like /, /boot, /home and that's it for the OS part.  It
> > downright sucks.
> > 
> > You planning to explain this to the devs?  Maybe you will have
> > better luck.  They said it is a dev from Fedora that started this .
> > . . . crap.  ;-)
> 
> I'll spend some time I don't have reading the archives, then see.

The very reason I use Gentoo is BECAUSE I don't like RHL.  :(

I still have to use CentOS on a server and curse every time it won't work like 
Gentoo.  Can't they just leave us alone?
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sat, 20 Aug 2011 03:48:18 -0500, Dale wrote about "Re:
[gentoo-user] /dev/sda* missing at boot":

[snip]
>I wish you could convince the devs of that.  I already have /var on
>its own and was planning to put /usr on its own.  I'm not now tho.
>Looks like /, /boot, /home and that's it for the OS part.  It
>downright sucks.

I have also been following the discussion on gentoo-dev, although I
currently only lurk there.  I was going to register and post with a
suggestion that everything should be on the root partition; that way we
could rename it C: and be compliant with the "industry standard".

However, it gets worse: one cannot safely fsck a partition or logical
volume once it has been mounted.  As things currently stand, there are
no statically linked fsck modules for ext2/3/4, as static linkage was
dropped from e2fsprogs about 3 years ago.  This means for fsck to run
inside an initramfs or intrd, the image will have to contain glibc,
libpthread and a whole slew of other large libraries in order to run
e2fsck with dynamic linkage.  The initramfs will end up being *many*
times larger than the kernel itself. [On my systems, the vmlinuz file
is only about 1.8 megs, and glibc alone makes that look really puny.]

Welcome to progress.
- -- 
Regards,

Dave  [RLU #314465]
======================================================================
dwnoon@ntlworld.com (David W Noon)
======================================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (GNU/Linux)

iEYEARECAAYFAk5Pr8QACgkQRQ2Fs59Psv+S8ACeMadMIjobzT61nCWoVrlqz0Pz
t50AoLZ83Jgw16BIWg7CD2tCb8hrdRzf
=gDyo
-----END PGP SIGNATURE-----

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1466 bytes --]

On Saturday 20 Aug 2011 13:59:42 David W Noon wrote:
> On Sat, 20 Aug 2011 03:48:18 -0500, Dale wrote about "Re:
> [gentoo-user] /dev/sda* missing at boot":
> 
> [snip]
> 
> >I wish you could convince the devs of that.  I already have /var on
> >its own and was planning to put /usr on its own.  I'm not now tho.
> >Looks like /, /boot, /home and that's it for the OS part.  It
> >downright sucks.
> 
> I have also been following the discussion on gentoo-dev, although I
> currently only lurk there.  I was going to register and post with a
> suggestion that everything should be on the root partition; that way we
> could rename it C: and be compliant with the "industry standard".
> 
> However, it gets worse: one cannot safely fsck a partition or logical
> volume once it has been mounted.  As things currently stand, there are
> no statically linked fsck modules for ext2/3/4, as static linkage was
> dropped from e2fsprogs about 3 years ago.  This means for fsck to run
> inside an initramfs or intrd, the image will have to contain glibc,
> libpthread and a whole slew of other large libraries in order to run
> e2fsck with dynamic linkage.  The initramfs will end up being *many*
> times larger than the kernel itself. [On my systems, the vmlinuz file
> is only about 1.8 megs, and glibc alone makes that look really puny.]
> 
> Welcome to progress.

This is madness.  Is there anything we can do to stop it?
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Pandu Poluan]

Welllll...

... the Gentoo project can always fork e2fsprogs ...

... but who will maintain it, then?

Rgds,


On 2011-08-20, Mick <michaelkintzios@gmail.com> wrote:
> On Saturday 20 Aug 2011 13:59:42 David W Noon wrote:
>> On Sat, 20 Aug 2011 03:48:18 -0500, Dale wrote about "Re:
>> [gentoo-user] /dev/sda* missing at boot":
>>
>> [snip]
>>
>> >I wish you could convince the devs of that.  I already have /var on
>> >its own and was planning to put /usr on its own.  I'm not now tho.
>> >Looks like /, /boot, /home and that's it for the OS part.  It
>> >downright sucks.
>>
>> I have also been following the discussion on gentoo-dev, although I
>> currently only lurk there.  I was going to register and post with a
>> suggestion that everything should be on the root partition; that way we
>> could rename it C: and be compliant with the "industry standard".
>>
>> However, it gets worse: one cannot safely fsck a partition or logical
>> volume once it has been mounted.  As things currently stand, there are
>> no statically linked fsck modules for ext2/3/4, as static linkage was
>> dropped from e2fsprogs about 3 years ago.  This means for fsck to run
>> inside an initramfs or intrd, the image will have to contain glibc,
>> libpthread and a whole slew of other large libraries in order to run
>> e2fsck with dynamic linkage.  The initramfs will end up being *many*
>> times larger than the kernel itself. [On my systems, the vmlinuz file
>> is only about 1.8 megs, and glibc alone makes that look really puny.]
>>
>> Welcome to progress.
>
> This is madness.  Is there anything we can do to stop it?
> --
> Regards,
> Mick
>


-- 
--
Pandu E Poluan - IT Optimizer
My website: http://pandu.poluan.info/

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat 20 August 2011 14:29:15 Mick did opine thusly:
> On Saturday 20 Aug 2011 13:59:42 David W Noon wrote:
> > On Sat, 20 Aug 2011 03:48:18 -0500, Dale wrote about "Re:
> > [gentoo-user] /dev/sda* missing at boot":
> > 
> > [snip]
> > 
> > >I wish you could convince the devs of that.  I already have
> > >/var on its own and was planning to put /usr on its own.  I'm
> > >not now tho. Looks like /, /boot, /home and that's it for the
> > >OS part.  It downright sucks.
> > 
> > I have also been following the discussion on gentoo-dev,
> > although I currently only lurk there.  I was going to register
> > and post with a suggestion that everything should be on the
> > root partition; that way we could rename it C: and be compliant
> > with the "industry standard".
> > 
> > However, it gets worse: one cannot safely fsck a partition or
> > logical volume once it has been mounted.  As things currently
> > stand, there are no statically linked fsck modules for
> > ext2/3/4, as static linkage was dropped from e2fsprogs about 3
> > years ago.  This means for fsck to run inside an initramfs or
> > intrd, the image will have to contain glibc, libpthread and a
> > whole slew of other large libraries in order to run e2fsck with
> > dynamic linkage.  The initramfs will end up being *many* times
> > larger than the kernel itself. [On my systems, the vmlinuz file
> > is only about 1.8 megs, and glibc alone makes that look really
> > puny.]
> > 
> > Welcome to progress.
> 
> This is madness.  Is there anything we can do to stop it?

Fork.


-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 806 bytes --]

On Sat, 20 Aug 2011 20:58:53 +0700, Pandu Poluan wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On 2011-08-20, Mick <michaelkintzios@gmail.com> wrote:
[snip]
> > This is madness.  Is there anything we can do to stop it?

[top posting corrected]
> Welllll...
> 
> ... the Gentoo project can always fork e2fsprogs ...
> 
> ... but who will maintain it, then?

I will be working on this next week.  I hope to resurrect the old
e2fsck.static program by late in the week, as the Makefile recipes
seem to be still there, but the target is no longer on the list.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> I'll spend some time I don't have reading the archives, then see. 


Well, they are about to move some openrc things to /usr or at least that 
was the way it was leaning earlier.  The thread is titled "rfc: using 
/libexec" if you are interested.  If you still don't like the idea of 
not being able to have /usr on a separate partition, may want to speak 
up soon.  Once openrc stuff gets moved there, you won't boot with a 
separate /usr anymore.

What is happening to Linux nowadays?  :/

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[William Hubbs]

[-- Attachment #1: Type: text/plain, Size: 1006 bytes --]

On Tue, Sep 06, 2011 at 06:06:35PM -0500, Dale wrote:
> Alan McKinnon wrote:
> > I'll spend some time I don't have reading the archives, then see. 
> 
> 
> Well, they are about to move some openrc things to /usr or at least that 
> was the way it was leaning earlier.  The thread is titled "rfc: using 
> /libexec" if you are interested.  If you still don't like the idea of 
> not being able to have /usr on a separate partition, may want to speak 
> up soon.  Once openrc stuff gets moved there, you won't boot with a 
> separate /usr anymore.

I'm not quite sure what is going to happen with this yet. If we do move
openrc to /usr, there will be a way provided to boot with separate /usr.

> What is happening to Linux nowadays?  :/

In a nutshell, it is because of udev rules running things in /usr. That
forces /usr to be available as part of the early boot sequence.

This is definitely not a choice that the gentoo disto made; it is coming
from several upstreams.

William


[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

William Hubbs wrote:
> On Tue, Sep 06, 2011 at 06:06:35PM -0500, Dale wrote:
>> Alan McKinnon wrote:
>>> I'll spend some time I don't have reading the archives, then see.
>>
>> Well, they are about to move some openrc things to /usr or at least that
>> was the way it was leaning earlier.  The thread is titled "rfc: using
>> /libexec" if you are interested.  If you still don't like the idea of
>> not being able to have /usr on a separate partition, may want to speak
>> up soon.  Once openrc stuff gets moved there, you won't boot with a
>> separate /usr anymore.
> I'm not quite sure what is going to happen with this yet. If we do move
> openrc to /usr, there will be a way provided to boot with separate /usr.

Actually, I would prefer the /run option and put all the stuff needed 
for booting in there.  Just keep it small.  I was planning to separate 
/usr but I don't want the hassle of a init* system.  So, instead of just 
creating a new partition for /usr, I got to redo the whole drive setup.  
If I do end up with the init* anyway, I may as well use lvm too.  I sure 
hate to jump into all that at once tho.  Murphy's Law just loves the 
crap out of me.


>
>> What is happening to Linux nowadays?  :/
> In a nutshell, it is because of udev rules running things in /usr. That
> forces /usr to be available as part of the early boot sequence.
>
> This is definitely not a choice that the gentoo disto made; it is coming
> from several upstreams.
>
> William
>

I know it is from upstream but it still tastes really bad.   ;-)

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dan Johansson]

> > This is definitely not a choice that the gentoo disto made; it is coming
> > from several upstreams.
> 
> I know it is from upstream but it still tastes really bad.   ;-)

I can only agree!
I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
This will be a MAJOR issue if /usr needs to be on /.

Just my 2 cents.
-- 
Dan Johansson, <http://www.dmj.nu>
***************************************************
This message is printed on 100% recycled electrons!
***************************************************

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 1:23 PM, Dan Johansson <Dan.Johansson@dmj.nu> wrote:
>> > This is definitely not a choice that the gentoo disto made; it is coming
>> > from several upstreams.
>>
>> I know it is from upstream but it still tastes really bad.   ;-)
>
> I can only agree!
> I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
> This will be a MAJOR issue if /usr needs to be on /.

It is my understanding that /usr does *not* need to be on /, only that
if you do, you will need an initramfs. Look at
http://article.gmane.org/gmane.linux.gentoo.devel/72275 and the thread
that followed it.

After reading that, and other similar threads, I still don't
understand the benefits of a separated /usr. Mounting it read-only
seems the only sensible one, and then I think is better to go all the
way and mount / read-only.

But that's just me: anyway, anyone will be able to keep /usr in
another partition if so they desire it. They will only need to use an
initramfs.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Wed, Sep 7, 2011 at 1:52 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 1:23 PM, Dan Johansson <Dan.Johansson@dmj.nu> wrote:
>>> > This is definitely not a choice that the gentoo disto made; it is coming
>>> > from several upstreams.
>>>
>>> I know it is from upstream but it still tastes really bad.   ;-)
>>
>> I can only agree!
>> I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
>> This will be a MAJOR issue if /usr needs to be on /.
>
> It is my understanding that /usr does *not* need to be on /, only that
> if you do, you will need an initramfs. Look at
> http://article.gmane.org/gmane.linux.gentoo.devel/72275 and the thread
> that followed it.

...and now I know that my entirely UUID-driven fstab may stop working,
if they choose not to add that particular "tweak/improvement".



-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 2:09 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 1:52 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Wed, Sep 7, 2011 at 1:23 PM, Dan Johansson <Dan.Johansson@dmj.nu> wrote:
>>>> > This is definitely not a choice that the gentoo disto made; it is coming
>>>> > from several upstreams.
>>>>
>>>> I know it is from upstream but it still tastes really bad.   ;-)
>>>
>>> I can only agree!
>>> I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
>>> This will be a MAJOR issue if /usr needs to be on /.
>>
>> It is my understanding that /usr does *not* need to be on /, only that
>> if you do, you will need an initramfs. Look at
>> http://article.gmane.org/gmane.linux.gentoo.devel/72275 and the thread
>> that followed it.
>
> ...and now I know that my entirely UUID-driven fstab may stop working,
> if they choose not to add that particular "tweak/improvement".

...or you could, you know, use the genkernel generated initramfs, or
dracut. Anyway, probably UUIDs and labels will be added to the minimal
initramfs (it is my undrestanding it's kinda easy to do). The
important thing is that it will be still supported.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Wed, Sep 7, 2011 at 2:28 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 2:09 PM, Michael Mol <mikemol@gmail.com> wrote:
>> On Wed, Sep 7, 2011 at 1:52 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>> On Wed, Sep 7, 2011 at 1:23 PM, Dan Johansson <Dan.Johansson@dmj.nu> wrote:
>>>>> > This is definitely not a choice that the gentoo disto made; it is coming
>>>>> > from several upstreams.
>>>>>
>>>>> I know it is from upstream but it still tastes really bad.   ;-)
>>>>
>>>> I can only agree!
>>>> I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
>>>> This will be a MAJOR issue if /usr needs to be on /.
>>>
>>> It is my understanding that /usr does *not* need to be on /, only that
>>> if you do, you will need an initramfs. Look at
>>> http://article.gmane.org/gmane.linux.gentoo.devel/72275 and the thread
>>> that followed it.
>>
>> ...and now I know that my entirely UUID-driven fstab may stop working,
>> if they choose not to add that particular "tweak/improvement".
>
> ...or you could, you know, use the genkernel generated initramfs, or
> dracut. Anyway, probably UUIDs and labels will be added to the minimal
> initramfs (it is my undrestanding it's kinda easy to do). The
> important thing is that it will be still supported.

I use the proprietary NVidia drivers, so genkernel went away very
early in my system's lifetime. I hadn't heard about dracut until
today. I still don't know anything about it, really.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 3:07 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 2:28 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Wed, Sep 7, 2011 at 2:09 PM, Michael Mol <mikemol@gmail.com> wrote:
>>> On Wed, Sep 7, 2011 at 1:52 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>>> On Wed, Sep 7, 2011 at 1:23 PM, Dan Johansson <Dan.Johansson@dmj.nu> wrote:
>>>>>> > This is definitely not a choice that the gentoo disto made; it is coming
>>>>>> > from several upstreams.
>>>>>>
>>>>>> I know it is from upstream but it still tastes really bad.   ;-)
>>>>>
>>>>> I can only agree!
>>>>> I am having /usr on a LVM volume on all systems (Gentoo and non Gentoo).
>>>>> This will be a MAJOR issue if /usr needs to be on /.
>>>>
>>>> It is my understanding that /usr does *not* need to be on /, only that
>>>> if you do, you will need an initramfs. Look at
>>>> http://article.gmane.org/gmane.linux.gentoo.devel/72275 and the thread
>>>> that followed it.
>>>
>>> ...and now I know that my entirely UUID-driven fstab may stop working,
>>> if they choose not to add that particular "tweak/improvement".
>>
>> ...or you could, you know, use the genkernel generated initramfs, or
>> dracut. Anyway, probably UUIDs and labels will be added to the minimal
>> initramfs (it is my undrestanding it's kinda easy to do). The
>> important thing is that it will be still supported.
>
> I use the proprietary NVidia drivers, so genkernel went away very
> early in my system's lifetime. I hadn't heard about dracut until
> today. I still don't know anything about it, really.

It's another initramfs creator. Since I use systemd and wanted to try
plymouth, I started to use it. In my case it works, and I get a really
nice splash screen at boot time.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Michael Mol writes:

> I use the proprietary NVidia drivers, so genkernel went away very
> early in my system's lifetime.

Huh? What does genkernel have to do with NVidia drivers?

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Wed, Sep 7, 2011 at 3:19 PM, Alex Schuster <wonko@wonkology.org> wrote:
> Michael Mol writes:
>
>> I use the proprietary NVidia drivers, so genkernel went away very
>> early in my system's lifetime.
>
> Huh? What does genkernel have to do with NVidia drivers?

genkernel included nouvou, which conflicted with the NVidia
proprietary drivers at the time.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 3:24 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 3:19 PM, Alex Schuster <wonko@wonkology.org> wrote:
>> Michael Mol writes:
>>
>>> I use the proprietary NVidia drivers, so genkernel went away very
>>> early in my system's lifetime.
>>
>> Huh? What does genkernel have to do with NVidia drivers?
>
> genkernel included nouvou, which conflicted with the NVidia
> proprietary drivers at the time.

I'm pretty sure (but could be wrong, I haven't used genkernel in ages)
that there is a way to blacklist some modules and force others.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 583 bytes --]

On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:

> After reading that, and other similar threads, I still don't
> understand the benefits of a separated /usr.

Putting it on a logical volume is one advantage, allowing /usr to be
resized should the need arise.

> Mounting it read-only
> seems the only sensible one, and then I think is better to go all the
> way and mount / read-only.

Putting /etc on a read-only filesystem seems a really bad idea.


-- 
Neil Bothwick

Q: What's the proper plural of a 'Net-connected Windows machine?
A: A Botnet

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 6:54 PM, Neil Bothwick <neil@digimed.co.uk> wrote:
> On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:
>
>> After reading that, and other similar threads, I still don't
>> understand the benefits of a separated /usr.
>
> Putting it on a logical volume is one advantage, allowing /usr to be
> resized should the need arise.

Why not allow / to be resized entirely? You probably will take the
machine off-line anyway.

>> Mounting it read-only
>> seems the only sensible one, and then I think is better to go all the
>> way and mount / read-only.
>
> Putting /etc on a read-only filesystem seems a really bad idea.

mount -o remount,rw /
emerge --sync && emerge -uDNv world
dispatch-conf
mount -o remount,ro /

Or, if you only want to modify some configuration file (which in a
sane environment doesn't happen that often):

mount -o remount,rw /
adduser fulano ...
mount -o remount,ro /

Again, I don't see the reason for a separated /usr. But *again*, if
that's what you want, you will be able to do it. You will just need an
initramfs.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Wed, Sep 7, 2011 at 7:04 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 6:54 PM, Neil Bothwick <neil@digimed.co.uk> wrote:
>> On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:
>>
>>> After reading that, and other similar threads, I still don't
>>> understand the benefits of a separated /usr.
>>
>> Putting it on a logical volume is one advantage, allowing /usr to be
>> resized should the need arise.
>
> Why not allow / to be resized entirely? You probably will take the
> machine off-line anyway.

A few months ago, I had to recover a live Debian machine which had
massive filesystem corruption in /usr; the hard drive it was on had
begun going bad, and it was taking out /usr slowly.

I wound up being able to recover by doing a full reinstall of all
packages on the live system after mounting /usr into a freshly-mkfs'd
new lvm volume. If I'd taken the system offline, it would have been
much more difficult.

(As it was, I was shocked it worked)

>
>>> Mounting it read-only
>>> seems the only sensible one, and then I think is better to go all the
>>> way and mount / read-only.
>>
>> Putting /etc on a read-only filesystem seems a really bad idea.
>
> mount -o remount,rw /
> emerge --sync && emerge -uDNv world
> dispatch-conf
> mount -o remount,ro /
>
> Or, if you only want to modify some configuration file (which in a
> sane environment doesn't happen that often):
>
> mount -o remount,rw /
> adduser fulano ...
> mount -o remount,ro /

So, no hobbyists? Operating a 'sane' environment at home isn't how
I've taught myself Linux. In a production environment, sure; having
everything possible be read-only is nice, from a security standpoint.

>
> Again, I don't see the reason for a separated /usr. But *again*, if
> that's what you want, you will be able to do it. You will just need an
> initramfs.

Yeah, great. Used to be, I could configure needed components to be
built-ins in the kernel.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1620 bytes --]

On Wed, 7 Sep 2011 19:04:17 -0400, Canek Peláez Valdés wrote:

> > Putting it on a logical volume is one advantage, allowing /usr to be
> > resized should the need arise.  
> 
> Why not allow / to be resized entirely? You probably will take the
> machine off-line anyway.

Because you can't boot from an LV, so you'd than need a separate /boot
and an initramfs. Without LVM, you are unlikely to be able to resize /
or /usr as it is not usually the last partition on the drive.

> >> Mounting it read-only
> >> seems the only sensible one, and then I think is better to go all the
> >> way and mount / read-only.  
> >
> > Putting /etc on a read-only filesystem seems a really bad idea.  
> 
> mount -o remount,rw /
> emerge --sync && emerge -uDNv world
> dispatch-conf
> mount -o remount,ro /
> 
> Or, if you only want to modify some configuration file (which in a
> sane environment doesn't happen that often):
> 
> mount -o remount,rw /
> adduser fulano ...
> mount -o remount,ro /

This is longer than the init script needed in an initramfs. I wonder what
problems you'd have when booting as the kernel tries to update the likes
of /etc/mtab on a read-only fs.

> Again, I don't see the reason for a separated /usr.

That doesn't mean there aren't several valid reasons to do so.

> But *again*, if
> that's what you want, you will be able to do it. You will just need an
> initramfs.

I neither have nor need one at the moment, which means this update will
break my system.


-- 
Neil Bothwick

Last words of a Windows user: = Where do I have to click now? - There?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 1564 bytes --]

On Wed, 7 Sep 2011 23:54:57 +0100, Neil Bothwick wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:
> 
> > After reading that, and other similar threads, I still don't
> > understand the benefits of a separated /usr.
> 
> Putting it on a logical volume is one advantage, allowing /usr to be
> resized should the need arise.

More than this, one can put /usr on a stripe set so that /usr/bin
and /usr/lib, two of the directories with the highest I/O traffic, can
be made more performant.  But this requires LVM, RAID or some blend of
both.  This, in turn, precludes that it be merged with /, unless the
initramfs grows even more to handle those extra DASD management
facilities.

The more I think about this merge of / and /usr, the dumber I think the
idea is.  As I wrote in an earlier message on this list, the initramfs
will be many times larger than the kernel itself.  Indeed, my /boot
partition is only 32 MiB, and that will be too small to contain all the
extra libraries and programs to run the initramfs script.

> > Mounting it read-only
> > seems the only sensible one, and then I think is better to go all
> > the way and mount / read-only.
> 
> Putting /etc on a read-only filesystem seems a really bad idea.

To say the least.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

David W Noon wrote:
> The more I think about this merge of / and /usr, the dumber I think 
> the idea is. As I wrote in an earlier message on this list, the 
> initramfs will be many times larger than the kernel itself. Indeed, my 
> /boot partition is only 32 MiB, and that will be too small to contain 
> all the extra libraries and programs to run the initramfs script.

If you don't like the idea, better hope over to -dev and help come up 
with ideas.  Actually, you may have to go hunt down the Fedora dev that 
came up with this brilliant idea.  The Gentoo devs are just having to 
follow along since the Fedora dev isn't listening.  Personally, I hope 
he runs into a real hair pulling issue that makes him regret this.  Then 
again, he is using a binary distro which uses a init* anyway so he may 
not care either way.  Should we wish for a nice rm command to get put 
into his init* thingy and erase his stuff?  Evil ain't I?   LOL

What's the emoticon thingy for choking a person?  This work?  :-@

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 7:39 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 7:04 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Wed, Sep 7, 2011 at 6:54 PM, Neil Bothwick <neil@digimed.co.uk> wrote:
>>> On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:
>>>
>>>> After reading that, and other similar threads, I still don't
>>>> understand the benefits of a separated /usr.
>>>
>>> Putting it on a logical volume is one advantage, allowing /usr to be
>>> resized should the need arise.
>>
>> Why not allow / to be resized entirely? You probably will take the
>> machine off-line anyway.
>
> A few months ago, I had to recover a live Debian machine which had
> massive filesystem corruption in /usr; the hard drive it was on had
> begun going bad, and it was taking out /usr slowly.
>
> I wound up being able to recover by doing a full reinstall of all
> packages on the live system after mounting /usr into a freshly-mkfs'd
> new lvm volume. If I'd taken the system offline, it would have been
> much more difficult.

You can always remount / in another LVM module. Really, what's so
especial about /usr?

> (As it was, I was shocked it worked)
>
>>
>>>> Mounting it read-only
>>>> seems the only sensible one, and then I think is better to go all the
>>>> way and mount / read-only.
>>>
>>> Putting /etc on a read-only filesystem seems a really bad idea.
>>
>> mount -o remount,rw /
>> emerge --sync && emerge -uDNv world
>> dispatch-conf
>> mount -o remount,ro /
>>
>> Or, if you only want to modify some configuration file (which in a
>> sane environment doesn't happen that often):
>>
>> mount -o remount,rw /
>> adduser fulano ...
>> mount -o remount,ro /
>
> So, no hobbyists? Operating a 'sane' environment at home isn't how
> I've taught myself Linux. In a production environment, sure; having
> everything possible be read-only is nice, from a security standpoint.

Then why you want another partition for /usr?

>>
>> Again, I don't see the reason for a separated /usr. But *again*, if
>> that's what you want, you will be able to do it. You will just need an
>> initramfs.
>
> Yeah, great. Used to be, I could configure needed components to be
> built-ins in the kernel.

Yeah, first time I installed Linux, it required 512 Mb (if I installed
X), and 16 Mb of memmory. Change happens. I welcome it happily,
because that's how we progress and get even better stuff.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 7:55 PM, Neil Bothwick <neil@digimed.co.uk> wrote:
> On Wed, 7 Sep 2011 19:04:17 -0400, Canek Peláez Valdés wrote:
>
>> > Putting it on a logical volume is one advantage, allowing /usr to be
>> > resized should the need arise.
>>
>> Why not allow / to be resized entirely? You probably will take the
>> machine off-line anyway.
>
> Because you can't boot from an LV, so you'd than need a separate /boot
> and an initramfs. Without LVM, you are unlikely to be able to resize /
> or /usr as it is not usually the last partition on the drive.

So, you guys want a separated /usr, but don't want a separate /boot. Awesome.

>> >> Mounting it read-only
>> >> seems the only sensible one, and then I think is better to go all the
>> >> way and mount / read-only.
>> >
>> > Putting /etc on a read-only filesystem seems a really bad idea.
>>
>> mount -o remount,rw /
>> emerge --sync && emerge -uDNv world
>> dispatch-conf
>> mount -o remount,ro /
>>
>> Or, if you only want to modify some configuration file (which in a
>> sane environment doesn't happen that often):
>>
>> mount -o remount,rw /
>> adduser fulano ...
>> mount -o remount,ro /
>
> This is longer than the init script needed in an initramfs. I wonder what
> problems you'd have when booting as the kernel tries to update the likes
> of /etc/mtab on a read-only fs.

# ls -l /etc/mtab
lrwxrwxrwx 1 root root 17 Dec  1  2010 /etc/mtab -> /proc/self/mounts

I don't see any problem at all.

>> Again, I don't see the reason for a separated /usr.
>
> That doesn't mean there aren't several valid reasons to do so.

I didn't say they were invalid, I say that *I* don't see the reason to
separate /usr. The arguments exposed just don't convice me. But
anyway, you will be able to do it with an initramfs.

>> But *again*, if
>> that's what you want, you will be able to do it. You will just need an
>> initramfs.
>
> I neither have nor need one at the moment, which means this update will
> break my system.

Then don't update. Wanna keep up with upstream? Then accept that
sometimes you will need to change your setup, and change how you do
stuff.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 9:37 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> On Wed, 7 Sep 2011 23:54:57 +0100, Neil Bothwick wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> On Wed, 7 Sep 2011 13:52:22 -0400, Canek Peláez Valdés wrote:
>>
>> > After reading that, and other similar threads, I still don't
>> > understand the benefits of a separated /usr.
>>
>> Putting it on a logical volume is one advantage, allowing /usr to be
>> resized should the need arise.
>
> More than this, one can put /usr on a stripe set so that /usr/bin
> and /usr/lib, two of the directories with the highest I/O traffic, can
> be made more performant.  But this requires LVM, RAID or some blend of
> both.  This, in turn, precludes that it be merged with /, unless the
> initramfs grows even more to handle those extra DASD management
> facilities.
>
> The more I think about this merge of / and /usr, the dumber I think the
> idea is.  As I wrote in an earlier message on this list, the initramfs
> will be many times larger than the kernel itself.  Indeed, my /boot
> partition is only 32 MiB, and that will be too small to contain all the
> extra libraries and programs to run the initramfs script.

I don't see any problem with an initramfs larger than the kernel. It
will handle a lot of stuff. But if you don't want to change your /boot
partition, then don't upgrade to new kernels.

Change happens.

>> > Mounting it read-only
>> > seems the only sensible one, and then I think is better to go all
>> > the way and mount / read-only.
>>
>> Putting /etc on a read-only filesystem seems a really bad idea.
>
> To say the least.

It works, and it makes life easier for upstream. Which are the ones
writting the code.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Canek Peláez Valdés wrote:
> Then don't update. Wanna keep up with upstream? Then accept that 
> sometimes you will need to change your setup, and change how you do 
> stuff. Regards. 

This is so like something I have told folks about windoze.  Awesome !  
To think I stayed away from windoze because of the freedom Linux gives a 
user just to find out now, its not as different as I thought.  :-(

Yeppie !!

Dale

:-)  :-)

[gentoo-user] Re: /dev/sda* missing at boot

[Alberto Luaces]

Neil Bothwick writes:

> On Wed, 7 Sep 2011 19:04:17 -0400, Canek Peláez Valdés wrote:
>
>> > Putting it on a logical volume is one advantage, allowing /usr to be
>> > resized should the need arise.  
>> 
>> Why not allow / to be resized entirely? You probably will take the
>> machine off-line anyway.
>
> Because you can't boot from an LV, so you'd than need a separate /boot
> and an initramfs. Without LVM, you are unlikely to be able to resize /
> or /usr as it is not usually the last partition on the drive.
>

Isn't it possible now with grub2?. I think it can read inside LVM
partitions. I have not personally tested it but I have a RAID system
with no separate /boot partition — a close case.

-- 
Alberto

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1391 bytes --]

On Wed, 7 Sep 2011 23:30:16 -0400, Canek Peláez Valdés wrote:

> > Because you can't boot from an LV, so you'd than need a separate /boot
> > and an initramfs. Without LVM, you are unlikely to be able to resize /
> > or /usr as it is not usually the last partition on the drive.
> 
> So, you guys want a separated /usr, but don't want a separate /boot.
> Awesome.

I want as much as possible on LVM, and no initramfs. A small / and
everything else on LVM fulfils that need.

> >> Again, I don't see the reason for a separated /usr.
> >
> > That doesn't mean there aren't several valid reasons to do so.
> 
> I didn't say they were invalid, I say that *I* don't see the reason to
> separate /usr. The arguments exposed just don't convice me. But
> anyway, you will be able to do it with an initramfs.

No one is saying YOU should change your preferred setup. Please do us the
same courtesy. Accept that we may have thought quite hard about what we
want, what we need and how best to achieve it. How would you feel if you
were told that you had to separate /usr, install extra packages, go
through extra configuration steps and introduce more points of failure,
just to do what you are already doing?

> Then don't update.

I can't decide whether this comment is arrogant or ingenuous.


-- 
Neil Bothwick

Today's subliminal message is:                  .

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
> > The more I think about this merge of / and /usr, the dumber I think the
> > idea is.  As I wrote in an earlier message on this list, the initramfs
> > will be many times larger than the kernel itself.  Indeed, my /boot
> > partition is only 32 MiB, and that will be too small to contain all the
> > extra libraries and programs to run the initramfs script.
> 
> I don't see any problem with an initramfs larger than the kernel. It
> will handle a lot of stuff. But if you don't want to change your /boot
> partition, then don't upgrade to new kernels.

How about accepting the fact, that there are a lot of things out there "you 
don't see"? Get over it. People have told a lot of valid reasons. They might 
not seem valid to you, but that's not their problem.
Have you *ever* thought about machines, that are not x86 or x86_64?
Here's an intersting read:
http://permalink.gmane.org/gmane.linux.gentoo.devel/72769

> Change happens.

That's right. And sometimes these changes are simply bad ideas.

> >> > Mounting it read-only
> >> > seems the only sensible one, and then I think is better to go all
> >> > the way and mount / read-only.
> >> 
> >> Putting /etc on a read-only filesystem seems a really bad idea.
> > 
> > To say the least.
> 
> It works, and it makes life easier for upstream. Which are the ones
> writting the code.

Hu? There's one upstream writing all the code for all the stuff we use? That's 
news to me.

> Regards.

Regards,
Michael

[gentoo-user] Re: /dev/sda* missing at boot

[Alberto Luaces]

Alberto Luaces writes:

> Neil Bothwick writes:
>
>> On Wed, 7 Sep 2011 19:04:17 -0400, Canek Peláez Valdés wrote:
>>
>>> > Putting it on a logical volume is one advantage, allowing /usr to be
>>> > resized should the need arise.  
>>> 
>>> Why not allow / to be resized entirely? You probably will take the
>>> machine off-line anyway.
>>
>> Because you can't boot from an LV, so you'd than need a separate /boot
>> and an initramfs. Without LVM, you are unlikely to be able to resize /
>> or /usr as it is not usually the last partition on the drive.
>>
>
> Isn't it possible now with grub2?. I think it can read inside LVM
> partitions. I have not personally tested it but I have a RAID system
> with no separate /boot partition — a close case.

Ah, you mean without initramfs.

-- 
Alberto

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Wed, 7 Sep 2011 23:23:45 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> > I wound up being able to recover by doing a full reinstall of all
> > packages on the live system after mounting /usr into a
> > freshly-mkfs'd new lvm volume. If I'd taken the system offline, it
> > would have been much more difficult.  
> 
> You can always remount / in another LVM module. Really, what's so
> especial about /usr?

Don't get me started. Oh, wait, you just did.

Right, here goes:

An initramfs is optional becuase i can disable it in the kernel. I
would like to keep that optional.

FHS says I can have /usr on a separate partition and I would like to
keep that because it is a good idea.

FHS says I can mount /usr read-only if I choose, which is also a good
idea. On a shared jumphost with 570 concurrent users it's actually a
VERY GOOD ODEA and I'd rather not lose that facility thankyouverymuch.

I do not need, want nor can I find a valid reason to *require* an
initramfs. Systems boot just fine without them.

FHS says I can have the minimal software and tools to effect a system
repair on / and put then entirety of user-space on /usr. Everything
involved in this thread runs early in the boot process and I fail to
find a single convincing reason why /usr is involved at all. Anything
required at this point can simply be put into /bin, /sbin and /lib{,64}
which one will note is exactly how we have been doing it all along.

This whole mess has every indication of a singular maintainer who
cannot be bothered taking other people's needs into account and
foisting off his own personal preferences onto an entire ecosystem.

I think such people should take note of how Torvalds works and emulate
him as opposed to emulating say Drepper as a role-model for good
project mantainership practice.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Wed, Sep 7, 2011 at 11:39 PM, Dale <rdalek1967@gmail.com> wrote:
> Canek Peláez Valdés wrote:
>>
>> Then don't update. Wanna keep up with upstream? Then accept that sometimes
>> you will need to change your setup, and change how you do stuff. Regards.
>
> This is so like something I have told folks about windoze.  Awesome !  To
> think I stayed away from windoze because of the freedom Linux gives a user
> just to find out now, its not as different as I thought.  :-(

But the freedom is still there. The freedom to either keep your system
as it is (don't upgrade), or to modify the source code to suit your
own needs.

Just don't expect from upstream to maintain code for each and every
possible configuration. It gets really complex really really really
fast.

Upstream (either Gentoo, or the kernel, or udev, or all of them) will
decide to support only a subset of all possible configurations and it
will mark them as supported. Don't aprove of that? Then maintain it
yourself (which you have the freedom to do), or keep up with the
change.

Freedom doesn't equals to "give me everything I want, and the way I
want it". The freedom we have is "here is this set of programs, and we
support this set of configurations; if you don't like it, here is also
the source code". Which is light years better than in Windows or MacOS
X.

> Yeppie !!

Yipi indeed.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 3:59 AM, Neil Bothwick <neil@digimed.co.uk> wrote:
> On Wed, 7 Sep 2011 23:30:16 -0400, Canek Peláez Valdés wrote:
>
>> > Because you can't boot from an LV, so you'd than need a separate /boot
>> > and an initramfs. Without LVM, you are unlikely to be able to resize /
>> > or /usr as it is not usually the last partition on the drive.
>>
>> So, you guys want a separated /usr, but don't want a separate /boot.
>> Awesome.
>
> I want as much as possible on LVM, and no initramfs. A small / and
> everything else on LVM fulfils that need.

If you really want it, go ahead and code the support for it. The code
for every single piece on the stack is available, and you have the
right to modify it if you don't like upstream decisions.

Upstream(s) don't need or have to care about what I (or you, or
anybody) wants. They (for reasons that we may or may not agree) will
decide what set of configurations will be supported. Don't like the
decision? Use the source, Luke.

>> >> Again, I don't see the reason for a separated /usr.
>> >
>> > That doesn't mean there aren't several valid reasons to do so.
>>
>> I didn't say they were invalid, I say that *I* don't see the reason to
>> separate /usr. The arguments exposed just don't convice me. But
>> anyway, you will be able to do it with an initramfs.
>
> No one is saying YOU should change your preferred setup. Please do us the
> same courtesy.

I do: I don't want you to change your setup. Absolutely nobody is
saying that. I'm truly sorry if thats the way I sounded (my first
language is not English).

However, if you don't want to change your setup, you will have two
options: either you don't upgrade (which is feasible, if you limit
yourself to security upgrades), or you write the code necessary for
your particular configuration. We (neither you nor I) can force the
developers (being udev, kernel, Gentoo or whatever) to write the code
supporting configuration X or Y, no matter how "rational" it seems to
you (or me).

When upstream wants to changes stuff, I change my setup and relearn
how to do it with the new technology/options. It happens all the time:
we moved from devfs to udev, from OSS to ALSA (and then PulseAudio),
from ipchains to iptables... I could go on and on.

> want, what we need and how best to achieve it. How would you feel if you
> were told that you had to separate /usr, install extra packages, go
> through extra configuration steps and introduce more points of failure,
> just to do what you are already doing?

I will do it. I have done it. And the reason is that the kernel churns
out a new version every 2 or 3 months, GNOME gets a new version every
six months, and everything gets (usually) better so fast thatt you
want to keep ip. And the only way to keep up with development, is
sometimes you need to change your setup.

Of course there are different setups. My laptop I keep it updated
every week. Sometimes shit gets wrong: I fix it, no problem. A
production server only gets security updates, and I keep a twin setup
ready if something goes wrong updating. If an update is intrusive
(say, I need to change /usr to the same partition as /), then I do it
first on the twin, and when it's ready I change them and do the same
in the primary.

And life goes on.

>> Then don't update.
>
> I can't decide whether this comment is arrogant or ingenuous.

It's called "being in production". You don't upgrade your kernel in
production, unless there is a security flaw that affects you. If there
is a security flaw in NFS, and you don't use NFS, you don't upgrade.
You don't upgrade Apache unless there is a security flaw. You don't
upgrade *anything* if it's working.

Want the new features? Guess what? You need to use the supported
setups, or write the code for your particular setup.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:09 AM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
>> > The more I think about this merge of / and /usr, the dumber I think the
>> > idea is.  As I wrote in an earlier message on this list, the initramfs
>> > will be many times larger than the kernel itself.  Indeed, my /boot
>> > partition is only 32 MiB, and that will be too small to contain all the
>> > extra libraries and programs to run the initramfs script.
>>
>> I don't see any problem with an initramfs larger than the kernel. It
>> will handle a lot of stuff. But if you don't want to change your /boot
>> partition, then don't upgrade to new kernels.
>
> How about accepting the fact, that there are a lot of things out there "you
> don't see"? Get over it. People have told a lot of valid reasons. They might
> not seem valid to you, but that's not their problem.

Relax man, I keep saying that is *I* who don't see a valid reason.
That doesn't mean there is no valid reason; I thought that went
without saying. Sorry if it sounded like I was invalidating all you
guys reasons.

My primary point was that, I *you* have your reasons to keep a
separated /usr, then by all means do it. You will only need an
initramfs.

> Have you *ever* thought about machines, that are not x86 or x86_64?
> Here's an intersting read:
> http://permalink.gmane.org/gmane.linux.gentoo.devel/72769

No, I haven't thought about them, because I don't use them. What it
has to do with anything?

>> Change happens.
>
> That's right. And sometimes these changes are simply bad ideas.

If so you think, then write the code to support the *really good* ideas.

>> >> > Mounting it read-only
>> >> > seems the only sensible one, and then I think is better to go all
>> >> > the way and mount / read-only.
>> >>
>> >> Putting /etc on a read-only filesystem seems a really bad idea.
>> >
>> > To say the least.
>>
>> It works, and it makes life easier for upstream. Which are the ones
>> writting the code.
>
> Hu? There's one upstream writing all the code for all the stuff we use? That's
> news to me.

Well, in this case by "upstream" I was meaning the Gentoo devs.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 10:51 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Wed, Sep 7, 2011 at 11:39 PM, Dale <rdalek1967@gmail.com> wrote:
>> Canek Peláez Valdés wrote:
>>>
>>> Then don't update. Wanna keep up with upstream? Then accept that sometimes
>>> you will need to change your setup, and change how you do stuff. Regards.
>>
>> This is so like something I have told folks about windoze.  Awesome !  To
>> think I stayed away from windoze because of the freedom Linux gives a user
>> just to find out now, its not as different as I thought.  :-(
>
> But the freedom is still there. The freedom to either keep your system
> as it is (don't upgrade), or to modify the source code to suit your
> own needs.

Please don't ever, ever, ever recommend not upgrading as a reasonable
long-term strategy. I don't like to think about how many security
problems exist in systems I'm familiar with because "not upgrading"
was the more convenient route.

The other side of what you're saying, "show me the code," is
reasonable. And if there's only one upstream maintainer who's got what
feels like the entire Linux community over a barrel on this, that
seems like a really good idea in principle.

> Just don't expect from upstream to maintain code for each and every
> possible configuration. It gets really complex really really really
> fast.

See also: LibreOffice requiring CUPS discussion earlier this week. No
surprises there, and it's understandable.

Still, I think I understand the complexity of what we're talking
about, yet it feels like the developer has a serious case of "my use
cases are the most valid ones, and I want to simplify udev's problem
space in favor of that."

As long as (and only as long as) udev isn't required for a server to
well and correctly, that's almost reasonable. That almost puts it in
the same class as DBus. (See the discussion from *last* week.)

Perhaps udev's problem is that it's too complex, as a result of having
too large a problem scope.

> Upstream (either Gentoo, or the kernel, or udev, or all of them) will
> decide to support only a subset of all possible configurations and it
> will mark them as supported. Don't aprove of that? Then maintain it
> yourself (which you have the freedom to do), or keep up with the
> change.
>
> Freedom doesn't equals to "give me everything I want, and the way I
> want it". The freedom we have is "here is this set of programs, and we
> support this set of configurations; if you don't like it, here is also
> the source code". Which is light years better than in Windows or MacOS

Code or GTFO. Classic FL/OSS fare. (Admittedly the best solution we've
found so far)

...

I remember devfs, and that it was rejected in favor of udev because
some things belong in userspace. udev, as far as I understand, udev
listens to hotplug events and performs actions in response. Perhaps an
alternate implementation is in order.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 6:11 AM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Wed, 7 Sep 2011 23:23:45 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> > I wound up being able to recover by doing a full reinstall of all
>> > packages on the live system after mounting /usr into a
>> > freshly-mkfs'd new lvm volume. If I'd taken the system offline, it
>> > would have been much more difficult.
>>
>> You can always remount / in another LVM module. Really, what's so
>> especial about /usr?
>
> Don't get me started. Oh, wait, you just did.

Uh oh :D

> Right, here goes:
>
> An initramfs is optional becuase i can disable it in the kernel. I
> would like to keep that optional.

udev at some point was optional, then it wasn't. Right now initramfs
is optional primarily because of embedded systems.

Change happens, I repeat.

> FHS says I can have /usr on a separate partition and I would like to
> keep that because it is a good idea.

FHS is dead: for years we didn't hear from it, and it was until a few
months that some activity was registered from it. For practical
reasons, it's dead, and nobody follows it completely (where in FHS is
/usr/libexec? do you use /srv like FHS says?)

But, if you think /usr in a separate partition is a good idea, then by
all means write the code for it.

> FHS says I can mount /usr read-only if I choose, which is also a good
> idea. On a shared jumphost with 570 concurrent users it's actually a
> VERY GOOD ODEA and I'd rather not lose that facility thankyouverymuch.

Then don't loose it. Just use an initramfs.

> I do not need, want nor can I find a valid reason to *require* an
> initramfs. Systems boot just fine without them.

Then either restrain yourself to security updates (which may be a good
idea if you support a server for 570 concurrent users), or write the
code to support a separated /usr without initramfs.

> FHS says I can have the minimal software and tools to effect a system
> repair on / and put then entirety of user-space on /usr. Everything
> involved in this thread runs early in the boot process and I fail to
> find a single convincing reason why /usr is involved at all. Anything
> required at this point can simply be put into /bin, /sbin and /lib{,64}
> which one will note is exactly how we have been doing it all along.

If it is so easy, then write the code to do it.

> This whole mess has every indication of a singular maintainer who
> cannot be bothered taking other people's needs into account and
> foisting off his own personal preferences onto an entire ecosystem.

And he magically convinces his distribution (and ours) to follow
through? Man, he must be really powerful. I don't think that it is
even possible that *maybe* some of his reasoning actually makes sense.

> I think such people should take note of how Torvalds works and emulate
> him as opposed to emulating say Drepper as a role-model for good
> project mantainership practice.

The people that writes the code, gets shit done. Code talks.

We always have the option to write the code ourselves, and get shit
done the way we want it. Don't want to do that? Then accept that you
will need to follow what the writers of the code decide.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 11:15 AM, Michael Mol <mikemol@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 10:51 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Wed, Sep 7, 2011 at 11:39 PM, Dale <rdalek1967@gmail.com> wrote:
>>> Canek Peláez Valdés wrote:
>>>>
>>>> Then don't update. Wanna keep up with upstream? Then accept that sometimes
>>>> you will need to change your setup, and change how you do stuff. Regards.
>>>
>>> This is so like something I have told folks about windoze.  Awesome !  To
>>> think I stayed away from windoze because of the freedom Linux gives a user
>>> just to find out now, its not as different as I thought.  :-(
>>
>> But the freedom is still there. The freedom to either keep your system
>> as it is (don't upgrade), or to modify the source code to suit your
>> own needs.
>
> Please don't ever, ever, ever recommend not upgrading as a reasonable
> long-term strategy. I don't like to think about how many security
> problems exist in systems I'm familiar with because "not upgrading"
> was the more convenient route.

Please refer to my other responses: of course you need to upgrade
security flaws. But in production, that's it.

> The other side of what you're saying, "show me the code," is
> reasonable. And if there's only one upstream maintainer who's got what
> feels like the entire Linux community over a barrel on this, that
> seems like a really good idea in principle.

I agree.

>> Just don't expect from upstream to maintain code for each and every
>> possible configuration. It gets really complex really really really
>> fast.
>
> See also: LibreOffice requiring CUPS discussion earlier this week. No
> surprises there, and it's understandable.
>
> Still, I think I understand the complexity of what we're talking
> about, yet it feels like the developer has a serious case of "my use
> cases are the most valid ones, and I want to simplify udev's problem
> space in favor of that."
>
> As long as (and only as long as) udev isn't required for a server to
> well and correctly, that's almost reasonable. That almost puts it in
> the same class as DBus. (See the discussion from *last* week.)
>
> Perhaps udev's problem is that it's too complex, as a result of having
> too large a problem scope.

I don't agree: udev is really small and (for what it does) really
simple. Not having udev (or dbus for that matter) is going back ten or
fifteen years. Both projects are prime examples of what I see as
*progress* in Linux.

>> Upstream (either Gentoo, or the kernel, or udev, or all of them) will
>> decide to support only a subset of all possible configurations and it
>> will mark them as supported. Don't aprove of that? Then maintain it
>> yourself (which you have the freedom to do), or keep up with the
>> change.
>>
>> Freedom doesn't equals to "give me everything I want, and the way I
>> want it". The freedom we have is "here is this set of programs, and we
>> support this set of configurations; if you don't like it, here is also
>> the source code". Which is light years better than in Windows or MacOS
>
> Code or GTFO. Classic FL/OSS fare. (Admittedly the best solution we've
> found so far)

It's not the best: it's the only one. Because as long as someone does
not write code, he can argue until his throat is dry but nobody will
really care.

> I remember devfs, and that it was rejected in favor of udev because
> some things belong in userspace. udev, as far as I understand, udev
> listens to hotplug events and performs actions in response. Perhaps an
> alternate implementation is in order.

Like I have been saying: by all means, write it if so you think. But
the kernel devs went through three iterations (classical Unix /dev,
devfs, and finally udev) to get to this point. I really don't think a
new implementation is neither necessary nor required.

But maybe I'm wrong, and some hotshot programmer will get it really
right (TM) next time. I really dubt it, though.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1057 bytes --]

On Thu, 8 Sep 2011 11:15:40 -0400, Michael Mol wrote:

> Perhaps udev's problem is that it's too complex, as a result of having
> too large a problem scope.

The problem, AIUI, is the udev can run any programs specified in the
rules files, and they may not be available before /usr is mounted. If the
problem is only restricted to udev, one solution may be to have a
two-stage startup. First the /dev hierarchy is created, then later on,
after filesystems have been mounted, the monitor is started and rules
executed.

As no solution comparable with this is being proposed, I can only assume
that there is a lot more to it. I'll probably switch to a separate /boot
and / on a LV with an initramfs, but that doesn't mean I have to like
this being forced on me. The only reason I am considering it now is that
I recently set up a system like this and can still remember most of what
I did (I hope).


-- 
Neil Bothwick

Barth's Distinction:
There are two types of people: those who divide people into two types, and
those who don't.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 11:13:58 schrieb Canek Peláez Valdés:
> On Thu, Sep 8, 2011 at 4:09 AM, Michael Schreckenbauer <grimlog@gmx.de> 
wrote:
> > Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
> >> I don't see any problem with an initramfs larger than the kernel. It
> >> will handle a lot of stuff. But if you don't want to change your /boot
> >> partition, then don't upgrade to new kernels.
> > 
> > How about accepting the fact, that there are a lot of things out there
> > "you don't see"? Get over it. People have told a lot of valid reasons.
> > They might not seem valid to you, but that's not their problem.
> 
> Relax man, I keep saying that is *I* who don't see a valid reason.
> That doesn't mean there is no valid reason; I thought that went
> without saying. Sorry if it sounded like I was invalidating all you
> guys reasons.
> 
> My primary point was that, I *you* have your reasons to keep a
> separated /usr, then by all means do it. You will only need an
> initramfs.

That's the point. You *need* an initramfs. You know KISS?

> > Have you *ever* thought about machines, that are not x86 or x86_64?
> > Here's an intersting read:
> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769
> 
> No, I haven't thought about them, because I don't use them. What it
> has to do with anything?

Well, I linked a mail. MIPS is mentioned. As I read it, there are cases with 
MIPS, where the initramfs *has* to be built into the kernel *and* the kernel-
image is size restricted. That's the problem with an initramfs bigger than the 
kernel itself.

> >> Change happens.
> > 
> > That's right. And sometimes these changes are simply bad ideas.
> 
> If so you think, then write the code to support the *really good* ideas.

Ah. Criticism is only allowed, if you are writing the code. Not in my world, 
sorry.

> >> >> > Mounting it read-only
> >> >> > seems the only sensible one, and then I think is better to
> >> >> > go all
> >> >> > the way and mount / read-only.
> >> >> 
> >> >> Putting /etc on a read-only filesystem seems a really bad idea.
> >> > 
> >> > To say the least.
> >> 
> >> It works, and it makes life easier for upstream. Which are the ones
> >> writting the code.
> > 
> > Hu? There's one upstream writing all the code for all the stuff we use?
> > That's news to me.
> 
> Well, in this case by "upstream" I was meaning the Gentoo devs.

Not all of the gentoo-devs are in favour of the idea.

> Regards.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 16:58:22 schrieb Neil Bothwick:
> On Thu, 8 Sep 2011 11:15:40 -0400, Michael Mol wrote:
> > Perhaps udev's problem is that it's too complex, as a result of having
> > too large a problem scope.
 
> The problem, AIUI, is the udev can run any programs specified in the
> rules files, and they may not be available before /usr is mounted.

Funny thing is, devfs was removed, because of "unfixable race-conditions" 
(among other things iirc). What else is this then?
An initramfs is not a proper fix for this design flaw, imo.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 12:06 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Donnerstag, 8. September 2011, 11:13:58 schrieb Canek Peláez Valdés:
>> On Thu, Sep 8, 2011 at 4:09 AM, Michael Schreckenbauer <grimlog@gmx.de>
> wrote:
>> > Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
>> >> I don't see any problem with an initramfs larger than the kernel. It
>> >> will handle a lot of stuff. But if you don't want to change your /boot
>> >> partition, then don't upgrade to new kernels.
>> >
>> > How about accepting the fact, that there are a lot of things out there
>> > "you don't see"? Get over it. People have told a lot of valid reasons.
>> > They might not seem valid to you, but that's not their problem.
>>
>> Relax man, I keep saying that is *I* who don't see a valid reason.
>> That doesn't mean there is no valid reason; I thought that went
>> without saying. Sorry if it sounded like I was invalidating all you
>> guys reasons.
>>
>> My primary point was that, I *you* have your reasons to keep a
>> separated /usr, then by all means do it. You will only need an
>> initramfs.
>
> That's the point. You *need* an initramfs. You know KISS?

If it's so "simple", write the code for support the option of not
having an initramfs. If it's not that simple, then what KISS are we
talking about?

>> > Have you *ever* thought about machines, that are not x86 or x86_64?
>> > Here's an intersting read:
>> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769
>>
>> No, I haven't thought about them, because I don't use them. What it
>> has to do with anything?
>
> Well, I linked a mail. MIPS is mentioned. As I read it, there are cases with
> MIPS, where the initramfs *has* to be built into the kernel *and* the kernel-
> image is size restricted. That's the problem with an initramfs bigger than the
> kernel itself.

That's a MIPS restriction. Then with MIPS you will need to put /usr in
/, and problem solved. But no, everyone wants everything, and exactly
the way they want.

Well, then they will need to write the code to support it, because no
developer is forced to support every single architecture in the whole
damn world, in every possible configuration available.

>> >> Change happens.
>> >
>> > That's right. And sometimes these changes are simply bad ideas.
>>
>> If so you think, then write the code to support the *really good* ideas.
>
> Ah. Criticism is only allowed, if you are writing the code. Not in my world,
> sorry.

By all means, criticize as much as you want. What I meant by "If so
you think, then write the code to support the *really good* ideas" is
that you have the *option* to do that. You can of course complain
forever: that will not mean that anybody (and in particular the
developers) will listen.

>> >> >> > Mounting it read-only
>> >> >> > seems the only sensible one, and then I think is better to
>> >> >> > go all
>> >> >> > the way and mount / read-only.
>> >> >>
>> >> >> Putting /etc on a read-only filesystem seems a really bad idea.
>> >> >
>> >> > To say the least.
>> >>
>> >> It works, and it makes life easier for upstream. Which are the ones
>> >> writting the code.
>> >
>> > Hu? There's one upstream writing all the code for all the stuff we use?
>> > That's news to me.
>>
>> Well, in this case by "upstream" I was meaning the Gentoo devs.
>
> Not all of the gentoo-devs are in favour of the idea.

Of course not. But, as with anything Open Source related, the ones
that write the support code will prevail. The complainers (if they
only complain) will not change anything.

My point is: if everything would be the other way around, and the
Gentoo (or kernel, or udev) developers decided that the True One Way
(TM) to do things were to separate / and /usr, I would do it. I did it
when me moved from ipchains to iptables, and that was particularly
painful because every single damn script just stopped working.

But such is life: i didn't write the code. If I wanted to keep up with
development, I needed to change my way of doing things. I have rolled
with the change every single time since I started to use Linux in 1996
(damn, I'm old), and sometimes it bite you in the ass in the long run
(hello HAL!)

But most of the times is for a good reason, and everything kinda
improves. And since I'm not writing code, just taking advantage of
getting it for free (as in beer and as in speech), I usually trust
developers. It usually pays off.

Of course, sometimes it doesn't (hello devfs!), but what are you going
to do? Look a gift horse in the mouth?

In the long term, trusting the developers usually it's the way to go.
Been here a long time, I stick to my guns. Don't like it? Well,
complain if you want, but if you don't writing some code it would
probably be for nothing.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 3022 bytes --]

On Wed, 7 Sep 2011 23:33:35 -0400, Canek Peláez Valdés wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Wed, Sep 7, 2011 at 9:37 PM, David W Noon <dwnoon@ntlworld.com>
> wrote:
[snip]
> > The more I think about this merge of / and /usr, the dumber I think
> > the idea is.  As I wrote in an earlier message on this list, the
> > initramfs will be many times larger than the kernel itself.
> >  Indeed, my /boot partition is only 32 MiB, and that will be too
> > small to contain all the extra libraries and programs to run the
> > initramfs script.
> 
> I don't see any problem with an initramfs larger than the kernel. It
> will handle a lot of stuff. But if you don't want to change your /boot
> partition, then don't upgrade to new kernels.

It is not the kernel that is the problem.  It is udev.

I expect to switch my simpler systems away from udev to mdev.  This
loses some functionality of udev, but that isn't needed on the simpler
hardware configurations.  So mdev could be the simplest solution to the
design flaws creeping into udev.

A very real problem with a large initramfs/initrd is maintaining the
software embedded in the image file.  If it contains duplicates of
e2fsck, reiserfsck, glibc, libpthread, etc., then these typically need
to be upgraded whenever the primary copy is upgraded.  The bigger the
initramfs becomes, the bigger the maintenance headache it inflicts.

> Change happens.

I think a more appropriate observation is: change is inevitable, but
progress isn't.

> >> > Mounting it read-only
> >> > seems the only sensible one, and then I think is better to go all
> >> > the way and mount / read-only.
> >>
> >> Putting /etc on a read-only filesystem seems a really bad idea.
> >
> > To say the least.
> 
> It works,

Putting /etc on a read-only mount works??  I take it you don't run any
database servers.  Every time I add a new database to PostgreSQL it
requires (for my needs) at least 1 new tablespace be created with its
own mount point.  This requires me to add at least 1 line to /etc/fstab
so that the new tablespace(s) is/are mounted before PostgreSQL starts
after a re-boot.  This becomes impossible if /etc is read-only.

Similarly, /etc/mtab needs to remain writeable, as symlinking it
to /proc/mounts (or /proc/self/mounts) won't always work for programs
that parse /etc/mtab.  This is because /proc/mounts contains additional
mount options that are fairly Linux-specific, whereas /etc/mtab should
be vanilla UNIX.

> and it makes life easier for upstream. Which are the ones
> writting the code.

It allows people developing udev scripts to use programs and
libraries that are not [currently] on rootfs inside their scripts.  If I
don't use those scripts, I don't care.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 12:11 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Donnerstag, 8. September 2011, 16:58:22 schrieb Neil Bothwick:
>> On Thu, 8 Sep 2011 11:15:40 -0400, Michael Mol wrote:
>> > Perhaps udev's problem is that it's too complex, as a result of having
>> > too large a problem scope.
>
>> The problem, AIUI, is the udev can run any programs specified in the
>> rules files, and they may not be available before /usr is mounted.
>
> Funny thing is, devfs was removed, because of "unfixable race-conditions"
> (among other things iirc). What else is this then?
> An initramfs is not a proper fix for this design flaw, imo.

Then design the correct solution and implement it. If it's technically
sound, it will prevail. I think it's a rather complicated problem with
a non trivial solution, but the code is there if you feel like give it
a try.

devfs was replaced by udev primarily because devfs shoved a lot stuff
in the kernel (the rules to create the devide nods) that belongs in
users pace. And I agree: the rules to determine what devices nodes
gets created by what hardware (that nowadays hardware appears and
disappears almost randomly from the kernel point of view), belongs in
user space. And guess what? Then to boot you need a minimal user
space. And the fool-proof way to get one before mounting anything, is
to put it in an initramfs.

Embedded systems don't need that, and deal with it in a particular way
(device-trees and other stuff). Desktop and servers can use (and I
think should use) an initramfs (yes, servers too, especially with
eSATA and similar things). The kernel devs have been moving in that
direction for a long time.

I would not be surprised that the option to not have initramfs will be
removed from the kernel in the future, unless you select
CONFIG_EMBEDDED=y.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 12:44 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> On Wed, 7 Sep 2011 23:33:35 -0400, Canek Peláez Valdés wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> On Wed, Sep 7, 2011 at 9:37 PM, David W Noon <dwnoon@ntlworld.com>
>> wrote:
> [snip]
>> > The more I think about this merge of / and /usr, the dumber I think
>> > the idea is.  As I wrote in an earlier message on this list, the
>> > initramfs will be many times larger than the kernel itself.
>> >  Indeed, my /boot partition is only 32 MiB, and that will be too
>> > small to contain all the extra libraries and programs to run the
>> > initramfs script.
>>
>> I don't see any problem with an initramfs larger than the kernel. It
>> will handle a lot of stuff. But if you don't want to change your /boot
>> partition, then don't upgrade to new kernels.
>
> It is not the kernel that is the problem.  It is udev.
>
> I expect to switch my simpler systems away from udev to mdev.  This
> loses some functionality of udev, but that isn't needed on the simpler
> hardware configurations.  So mdev could be the simplest solution to the
> design flaws creeping into udev.

Maybe. I would not bet on it, but any new technical experiment is
worth trying, I believe. I will stick with the kernel-blessed option
of udev, though.

> A very real problem with a large initramfs/initrd is maintaining the
> software embedded in the image file.  If it contains duplicates of
> e2fsck, reiserfsck, glibc, libpthread, etc., then these typically need
> to be upgraded whenever the primary copy is upgraded.  The bigger the
> initramfs becomes, the bigger the maintenance headache it inflicts.

Dracut automatizes this. Is a non-problem.

>> Change happens.
>
> I think a more appropriate observation is: change is inevitable, but
> progress isn't.

To progress you need to try (and sometimes fail). udev is the third
iteration to handle /dev in the kernel; maybe mdev is a better option,
but I'm highly sceptical.

>> >> > Mounting it read-only
>> >> > seems the only sensible one, and then I think is better to go all
>> >> > the way and mount / read-only.
>> >>
>> >> Putting /etc on a read-only filesystem seems a really bad idea.
>> >
>> > To say the least.
>>
>> It works,
>
> Putting /etc on a read-only mount works??  I take it you don't run any
> database servers.  Every time I add a new database to PostgreSQL it
> requires (for my needs) at least 1 new tablespace be created with its
> own mount point.  This requires me to add at least 1 line to /etc/fstab
> so that the new tablespace(s) is/are mounted before PostgreSQL starts
> after a re-boot.  This becomes impossible if /etc is read-only.

mount -o remount,rw /
do stuff...
mount -o remount,ro /

Really, I don't see the problem.

> Similarly, /etc/mtab needs to remain writeable, as symlinking it
> to /proc/mounts (or /proc/self/mounts) won't always work for programs
> that parse /etc/mtab.  This is because /proc/mounts contains additional
> mount options that are fairly Linux-specific, whereas /etc/mtab should
> be vanilla UNIX.

I really, really don't care about non-Linux systems. But that's me,
anyone else can use wathever they want. Just don't expect everyone to
be happy with the lowest common feature set.

Having said that, which programs do you use that need to parse mtab?

>> and it makes life easier for upstream. Which are the ones
>> writting the code.
>
> It allows people developing udev scripts to use programs and
> libraries that are not [currently] on rootfs inside their scripts.  If I
> don't use those scripts, I don't care.

And rightly so. Then by all means try mdev and other solutions. Maybe
one (or all) of them will result in the dominant technology in the
future. Maybe it will allow you to not use an initramfs, and ignore
udev, and to keep /usr separated from /.

Just don't expect the limited resources of the Gentoo devs to be able
to test and support your special configuration.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 12:34:50 schrieb Canek Peláez Valdés:
> On Thu, Sep 8, 2011 at 12:06 PM, Michael Schreckenbauer <grimlog@gmx.de> 
wrote:
> > Am Donnerstag, 8. September 2011, 11:13:58 schrieb Canek Peláez Valdés:
> >> On Thu, Sep 8, 2011 at 4:09 AM, Michael Schreckenbauer
> >> <grimlog@gmx.de>
> > 
> > wrote:
> >> > Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
> >> >> I don't see any problem with an initramfs larger than the
> >> >> kernel. It
> >> >> will handle a lot of stuff. But if you don't want to change your
> >> >> /boot partition, then don't upgrade to new kernels.
> >> > 
> >> > How about accepting the fact, that there are a lot of things out
> >> > there
> >> > "you don't see"? Get over it. People have told a lot of valid
> >> > reasons.
> >> > They might not seem valid to you, but that's not their problem.
> >> 
> >> Relax man, I keep saying that is *I* who don't see a valid reason.
> >> That doesn't mean there is no valid reason; I thought that went
> >> without saying. Sorry if it sounded like I was invalidating all you
> >> guys reasons.
> >> 
> >> My primary point was that, I *you* have your reasons to keep a
> >> separated /usr, then by all means do it. You will only need an
> >> initramfs.
> > 
> > That's the point. You *need* an initramfs. You know KISS?
> 
> If it's so "simple", write the code for support the option of not
> having an initramfs. If it's not that simple, then what KISS are we
> talking about?

We already *have* the situation of not requiring initramfs for separate /usr. 
Mission accomplished.
It's the upcoming change, that violates KISS. If udev cannot work properly 
with separate /usr, fix udev not the FS-hierarchy. What next? Put /home into 
initramfs, because udev decides it cannot work without /home mounted?

> >> > Have you *ever* thought about machines, that are not x86 or
> >> > x86_64?
> >> > Here's an intersting read:
> >> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769
> >> 
> >> No, I haven't thought about them, because I don't use them. What it
> >> has to do with anything?
> > 
> > Well, I linked a mail. MIPS is mentioned. As I read it, there are cases
> > with MIPS, where the initramfs *has* to be built into the kernel *and*
> > the kernel- image is size restricted. That's the problem with an
> > initramfs bigger than the kernel itself.
> 
> That's a MIPS restriction. Then with MIPS you will need to put /usr in
> /, and problem solved.

Solved? You call "no separate /usr on MIPS" a solution?
How about existing installations? Ah, yes, don't upgrade.

> But no, everyone wants everything, and exactly
> the way they want.

It works now.

> Well, then they will need to write the code to support it, because no
> developer is forced to support every single architecture in the whole
> damn world, in every possible configuration available.
> 
> >> >> Change happens.
> >> > 
> >> > That's right. And sometimes these changes are simply bad ideas.
> >> 
> >> If so you think, then write the code to support the *really good*
> >> ideas.
> > 
> > Ah. Criticism is only allowed, if you are writing the code. Not in my
> > world, sorry.
> 
> By all means, criticize as much as you want. What I meant by "If so
> you think, then write the code to support the *really good* ideas" is
> that you have the *option* to do that. You can of course complain
> forever: that will not mean that anybody (and in particular the
> developers) will listen.

Not listening to users is a very bad idea.

> Of course not. But, as with anything Open Source related, the ones
> that write the support code will prevail. The complainers (if they
> only complain) will not change anything.

You keep talking about "complainers". I'd say, we discuss things, as do the 
gentoo-devs on their list.

> My point is: if everything would be the other way around, and the
> Gentoo (or kernel, or udev) developers decided that the True One Way
> (TM) to do things were to separate / and /usr, I would do it. I did it
> when me moved from ipchains to iptables, and that was particularly
> painful because every single damn script just stopped working.

Ah yes. What option was lost, when this switch happend?
Nobody (I think) complains about some config changes. It's the removal of sane 
and valid options.

> But such is life: i didn't write the code. If I wanted to keep up with
> development, I needed to change my way of doing things. I have rolled
> with the change every single time since I started to use Linux in 1996
> (damn, I'm old), and sometimes it bite you in the ass in the long run
> (hello HAL!)
> 
> But most of the times is for a good reason, and everything kinda
> improves. And since I'm not writing code, just taking advantage of
> getting it for free (as in beer and as in speech), I usually trust
> developers. It usually pays off.

How's needing an initramsfs for separate /usr an improvement?

> Of course, sometimes it doesn't (hello devfs!), but what are you going
> to do? Look a gift horse in the mouth?
> 
> In the long term, trusting the developers usually it's the way to go.
> Been here a long time, I stick to my guns. Don't like it? Well,
> complain if you want, but if you don't writing some code it would
> probably be for nothing.

Yeah, "probably", that's why we discuss things.

> Regards.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 12:45:47 schrieb Canek Peláez Valdés:
> On Thu, Sep 8, 2011 at 12:11 PM, Michael Schreckenbauer <grimlog@gmx.de> 
wrote:
> > Am Donnerstag, 8. September 2011, 16:58:22 schrieb Neil Bothwick:
> >> On Thu, 8 Sep 2011 11:15:40 -0400, Michael Mol wrote:
> >> > Perhaps udev's problem is that it's too complex, as a result of
> >> > having
> >> > too large a problem scope.
> >> 
> >> The problem, AIUI, is the udev can run any programs specified in the
> >> rules files, and they may not be available before /usr is mounted.
> > 
> > Funny thing is, devfs was removed, because of "unfixable
> > race-conditions"
> > (among other things iirc). What else is this then?
> > An initramfs is not a proper fix for this design flaw, imo.
> 
> Then design the correct solution and implement it. If it's technically
> sound, it will prevail. I think it's a rather complicated problem with
> a non trivial solution, but the code is there if you feel like give it
> a try.

Where did I write, that I am in the position to write such a beast?
I only take the freedom to name this a design flaw in udev.
It needs things from userspace, which are not yet available at the point it 
requests them. An initramsfs is a workaround for this, not a proper fix.

> Regards.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:01 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Donnerstag, 8. September 2011, 12:34:50 schrieb Canek Peláez Valdés:
>> On Thu, Sep 8, 2011 at 12:06 PM, Michael Schreckenbauer <grimlog@gmx.de>
> wrote:
>> > Am Donnerstag, 8. September 2011, 11:13:58 schrieb Canek Peláez Valdés:
>> >> On Thu, Sep 8, 2011 at 4:09 AM, Michael Schreckenbauer
>> >> <grimlog@gmx.de>
>> >
>> > wrote:
>> >> > Am Mittwoch, 7. September 2011, 23:33:35 schrieb Canek Peláez Valdés:
>> >> >> I don't see any problem with an initramfs larger than the
>> >> >> kernel. It
>> >> >> will handle a lot of stuff. But if you don't want to change your
>> >> >> /boot partition, then don't upgrade to new kernels.
>> >> >
>> >> > How about accepting the fact, that there are a lot of things out
>> >> > there
>> >> > "you don't see"? Get over it. People have told a lot of valid
>> >> > reasons.
>> >> > They might not seem valid to you, but that's not their problem.
>> >>
>> >> Relax man, I keep saying that is *I* who don't see a valid reason.
>> >> That doesn't mean there is no valid reason; I thought that went
>> >> without saying. Sorry if it sounded like I was invalidating all you
>> >> guys reasons.
>> >>
>> >> My primary point was that, I *you* have your reasons to keep a
>> >> separated /usr, then by all means do it. You will only need an
>> >> initramfs.
>> >
>> > That's the point. You *need* an initramfs. You know KISS?
>>
>> If it's so "simple", write the code for support the option of not
>> having an initramfs. If it's not that simple, then what KISS are we
>> talking about?
>
> We already *have* the situation of not requiring initramfs for separate /usr.
> Mission accomplished.
> It's the upcoming change, that violates KISS. If udev cannot work properly
> with separate /usr, fix udev not the FS-hierarchy. What next? Put /home into
> initramfs, because udev decides it cannot work without /home mounted?

Then don't upgrade. Keep doing only security updates.

Want the new cool stuff? Roll with the change.

>> >> > Have you *ever* thought about machines, that are not x86 or
>> >> > x86_64?
>> >> > Here's an intersting read:
>> >> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769
>> >>
>> >> No, I haven't thought about them, because I don't use them. What it
>> >> has to do with anything?
>> >
>> > Well, I linked a mail. MIPS is mentioned. As I read it, there are cases
>> > with MIPS, where the initramfs *has* to be built into the kernel *and*
>> > the kernel- image is size restricted. That's the problem with an
>> > initramfs bigger than the kernel itself.
>>
>> That's a MIPS restriction. Then with MIPS you will need to put /usr in
>> /, and problem solved.
>
> Solved? You call "no separate /usr on MIPS" a solution?
> How about existing installations? Ah, yes, don't upgrade.

There you go.

>> But no, everyone wants everything, and exactly
>> the way they want.
>
> It works now.

Exactly, and if you don't upgrade, it will work as long as you want.

>> Well, then they will need to write the code to support it, because no
>> developer is forced to support every single architecture in the whole
>> damn world, in every possible configuration available.
>>
>> >> >> Change happens.
>> >> >
>> >> > That's right. And sometimes these changes are simply bad ideas.
>> >>
>> >> If so you think, then write the code to support the *really good*
>> >> ideas.
>> >
>> > Ah. Criticism is only allowed, if you are writing the code. Not in my
>> > world, sorry.
>>
>> By all means, criticize as much as you want. What I meant by "If so
>> you think, then write the code to support the *really good* ideas" is
>> that you have the *option* to do that. You can of course complain
>> forever: that will not mean that anybody (and in particular the
>> developers) will listen.
>
> Not listening to users is a very bad idea.

No, they listen to users. They just don't listen too every user,
because that's impossible. Maybe I'm wrong, but I think your setup is
in the minority of use-cases. Who they should listen to?

>> Of course not. But, as with anything Open Source related, the ones
>> that write the support code will prevail. The complainers (if they
>> only complain) will not change anything.
>
> You keep talking about "complainers".

If someone complains and doesn't code, it's a complainer. By
definition. If someone complains and code, it's creating alternative
technologies.

> I'd say, we discuss things, as do the
> gentoo-devs on their list.

I agree. I'm subscribed to both.

>> My point is: if everything would be the other way around, and the
>> Gentoo (or kernel, or udev) developers decided that the True One Way
>> (TM) to do things were to separate / and /usr, I would do it. I did it
>> when me moved from ipchains to iptables, and that was particularly
>> painful because every single damn script just stopped working.
>
> Ah yes. What option was lost, when this switch happend?
> Nobody (I think) complains about some config changes. It's the removal of sane
> and valid options.

You cannot keep *EVERY* option supported. It's impossible. They grow a
lot really fast. You have to mark some things as "not supported".
Don't like it? Try alternative technologies.

>> But such is life: i didn't write the code. If I wanted to keep up with
>> development, I needed to change my way of doing things. I have rolled
>> with the change every single time since I started to use Linux in 1996
>> (damn, I'm old), and sometimes it bite you in the ass in the long run
>> (hello HAL!)
>>
>> But most of the times is for a good reason, and everything kinda
>> improves. And since I'm not writing code, just taking advantage of
>> getting it for free (as in beer and as in speech), I usually trust
>> developers. It usually pays off.
>
> How's needing an initramsfs for separate /usr an improvement?

With the initramfs you can do a lot of really cool stuff. I know it's
shallow, but I really like my plymouth-based splash screen.

>> Of course, sometimes it doesn't (hello devfs!), but what are you going
>> to do? Look a gift horse in the mouth?
>>
>> In the long term, trusting the developers usually it's the way to go.
>> Been here a long time, I stick to my guns. Don't like it? Well,
>> complain if you want, but if you don't writing some code it would
>> probably be for nothing.
>
> Yeah, "probably", that's why we discuss things.

Again, we can discuss (or complain) until the sun is red. As long as
we don't give code back, it's basically academic.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:11 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Donnerstag, 8. September 2011, 12:45:47 schrieb Canek Peláez Valdés:
>> On Thu, Sep 8, 2011 at 12:11 PM, Michael Schreckenbauer <grimlog@gmx.de>
> wrote:
>> > Am Donnerstag, 8. September 2011, 16:58:22 schrieb Neil Bothwick:
>> >> On Thu, 8 Sep 2011 11:15:40 -0400, Michael Mol wrote:
>> >> > Perhaps udev's problem is that it's too complex, as a result of
>> >> > having
>> >> > too large a problem scope.
>> >>
>> >> The problem, AIUI, is the udev can run any programs specified in the
>> >> rules files, and they may not be available before /usr is mounted.
>> >
>> > Funny thing is, devfs was removed, because of "unfixable
>> > race-conditions"
>> > (among other things iirc). What else is this then?
>> > An initramfs is not a proper fix for this design flaw, imo.
>>
>> Then design the correct solution and implement it. If it's technically
>> sound, it will prevail. I think it's a rather complicated problem with
>> a non trivial solution, but the code is there if you feel like give it
>> a try.
>
> Where did I write, that I am in the position to write such a beast?
> I only take the freedom to name this a design flaw in udev.

And I only take the freedom to tell you that if you don't like the
design, you have the option of improved (or completely replace it) and
implement such design.

> It needs things from userspace, which are not yet available at the point it
> requests them. An initramsfs is a workaround for this, not a proper fix.

Again, it's a complex problem. What do you think is the proper fix?

I don't the kernel/Gentoo/udev devs are making this as a workaround. I
see it as a policy decision. Maybe I'm wrong, but I really trust the
devs.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-08 05:23, Canek Peláez Valdés wrote:

> Yeah, first time I installed Linux, it required 512 Mb (if I installed
> X), and 16 Mb of memmory. Change happens. I welcome it happily,
> because that's how we progress and get even better stuff.

Change is sometimes good, sometimes bad. Something that used to work
(and still works on other *nix:es) but doesn't because someone doesn't
think it should work that way is... well, what can I say? This is a
regression, there's no other way to put it.

Also, you say you welcome change... If you could have all of todays
functionality in a GNU/Linux OS in 512MB/16MB, wouldn't that be much
better than needing several GB (in both hard drive space and memory)?
You could put all that space (hard drive/mem) to good use instead... Or
are you one of those that thinks change is good just for the sake of...
change?

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-08 16:51, Canek Peláez Valdés wrote:

> But the freedom is still there. The freedom to either keep your system
> as it is (don't upgrade)
            ^^^^^^^^^^^^^
You do realise that this is quite valid for Windows (and all other OS's
in existence)? At least so far...

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 1:18 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 1:01 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
>>
>> We already *have* the situation of not requiring initramfs for separate /usr.
>> Mission accomplished.
>> It's the upcoming change, that violates KISS. If udev cannot work properly
>> with separate /usr, fix udev not the FS-hierarchy. What next? Put /home into
>> initramfs, because udev decides it cannot work without /home mounted?
>
> Then don't upgrade. Keep doing only security updates.

And, eventually, those security updates will stop coming. Just
pointing this out.

>> It works now.
>
> Exactly, and if you don't upgrade, it will work as long as you want.

See my remark on security updates not continuing indefinitely.

>> Not listening to users is a very bad idea.
>
> No, they listen to users. They just don't listen too every user,
> because that's impossible. Maybe I'm wrong, but I think your setup is
> in the minority of use-cases. Who they should listen to?
>
>> You keep talking about "complainers".
>
> If someone complains and doesn't code, it's a complainer. By
> definition. If someone complains and code, it's creating alternative
> technologies.
>
>> I'd say, we discuss things, as do the
>> gentoo-devs on their list.
>
> I agree. I'm subscribed to both.
>
>> Yeah, "probably", that's why we discuss things.
>
> Again, we can discuss (or complain) until the sun is red. As long as
> we don't give code back, it's basically academic.

This isn't a discussion. This is a bunch of people offering
displeasure, ideas and/or thoughts, and one person saying, "hey,
nothing I can do. I trust the devs."

A discussion is when there's an interchange of ideas, arguments,
counterarguments, and the fleshing out of a new framework of thought.
That kind of point/counterpoint is *vital* for architectural
foresight. All I keep reading from you is, "if you think that will
work, go write it." *No* writing for a problem of this scope is
warranted without some extensive discussion, noting of edge cases and
planning around the same.

People have been pointing out edge cases, use cases which are being
disregarded, etc, and pretty much all they're getting back is "I don't
see those as valid." Granted, you're kinda painting a target on your
back by being the only one defending upstream's decision here, but
when someone pointed at an already-existing alternative, you simply
said, "I doubt that'll be the solution."

As for me, this will be a royal inconvenience, and may require the
rebuilding of my primary machine. Still, I can deal. It'll mean
learning how to build initramfs*, how to make sure it contains the
needed tools, and probably a half-dozen other things I didn't even see
coming when I set up this box last fall.

* I've avoided it for ten years it was a grossly unnecessary
complexity for my systems. Now it sounds like it'll become a
necessity.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 1:35 PM, pk <peterk2@coolmail.se> wrote:
> On 2011-09-08 16:51, Canek Peláez Valdés wrote:
>
>> But the freedom is still there. The freedom to either keep your system
>> as it is (don't upgrade)
>            ^^^^^^^^^^^^^
> You do realise that this is quite valid for Windows (and all other OS's
> in existence)? At least so far...

Don't get *me* started. My _day job_ is C++/MFC on Windows. _Please_
upgrade, you'll make my life much easier.

Outdated operating systems make baby coder cry.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 2781 bytes --]

On Thu, 8 Sep 2011 12:56:44 -0400, Canek Peláez Valdés wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Thu, Sep 8, 2011 at 12:44 PM, David W Noon <dwnoon@ntlworld.com>
> wrote:
[snip]
> > I expect to switch my simpler systems away from udev to mdev.  This
> > loses some functionality of udev, but that isn't needed on the
> > simpler hardware configurations.  So mdev could be the simplest
> > solution to the design flaws creeping into udev.
> 
> Maybe. I would not bet on it, but any new technical experiment is
> worth trying, I believe. I will stick with the kernel-blessed option
> of udev, though.

I don't know if the kernel offers any particular blessing to any
hotplug handler.

> > A very real problem with a large initramfs/initrd is maintaining the
> > software embedded in the image file.  If it contains duplicates of
> > e2fsck, reiserfsck, glibc, libpthread, etc., then these typically
> > need to be upgraded whenever the primary copy is upgraded.  The
> > bigger the initramfs becomes, the bigger the maintenance headache
> > it inflicts.
> 
> Dracut automatizes this. Is a non-problem.

If dracut actually worked ...

[snip]
> mount -o remount,rw /
> do stuff...
> mount -o remount,ro /
> 
> Really, I don't see the problem.

During the "do stuff" phase, /usr is also writeable, which is
undesirable on production systems.  That's the *original* problem with
merging a read-only /usr with /. [We seem to be going in circles with
this one.]

> > Similarly, /etc/mtab needs to remain writeable, as symlinking it
> > to /proc/mounts (or /proc/self/mounts) won't always work for
> > programs that parse /etc/mtab.  This is because /proc/mounts
> > contains additional mount options that are fairly Linux-specific,
> > whereas /etc/mtab should be vanilla UNIX.
> 
> I really, really don't care about non-Linux systems. But that's me,
> anyone else can use wathever they want. Just don't expect everyone to
> be happy with the lowest common feature set.
> 
> Having said that, which programs do you use that need to parse mtab?

I have about 6 or 7 backup jobs that run during the night and
parse /etc/mtab to see if they need to place a copy of the backup onto
an external medium.  These examine the mount options and don't
understand the non-standard options offered by Linux in /proc/mounts.

[snip]
> Just don't expect the limited resources of the Gentoo devs to be able
> to test and support your special configuration.

:-))

They already don't do that.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-08 19:47, Michael Mol wrote:

> Outdated operating systems make baby coder cry.

I like making baby coders cry... >:->

;-)

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:30 PM, pk <peterk2@coolmail.se> wrote:
> On 2011-09-08 05:23, Canek Peláez Valdés wrote:
>
>> Yeah, first time I installed Linux, it required 512 Mb (if I installed
>> X), and 16 Mb of memmory. Change happens. I welcome it happily,
>> because that's how we progress and get even better stuff.
>
> Change is sometimes good, sometimes bad. Something that used to work
> (and still works on other *nix:es) but doesn't because someone doesn't
> think it should work that way is... well, what can I say? This is a
> regression, there's no other way to put it.

The world is never black or white. Under several definitions,
*nothing* has stopped working. You just need to do extra stuff to keep
it working.

> Also, you say you welcome change... If you could have all of todays
> functionality in a GNU/Linux OS in 512MB/16MB, wouldn't that be much
> better than needing several GB (in both hard drive space and memory)?
> You could put all that space (hard drive/mem) to good use instead... Or
> are you one of those that thinks change is good just for the sake of...
> change?

No, I understand that today's functionality in Linux requires more
hard drive space, more memory and faster CPU's. I don't like change
for change's sake, my whole point is that the benefits we get from
following upstream overweights the problems.

(Of course, again, is not black and white: the kernel devs, I trust
almost withouth doubt. GNOME devs I trust less. An unknown hacker with
a new project I trust almost nothing.)

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:35 PM, pk <peterk2@coolmail.se> wrote:
> On 2011-09-08 16:51, Canek Peláez Valdés wrote:
>
>> But the freedom is still there. The freedom to either keep your system
>> as it is (don't upgrade)
>            ^^^^^^^^^^^^^
> You do realise that this is quite valid for Windows (and all other OS's
> in existence)? At least so far...

Yeah, but not the second part that you conveniently omitted: the
freedom to modify the code.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:45 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 1:18 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Thu, Sep 8, 2011 at 1:01 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
>>>
>>> We already *have* the situation of not requiring initramfs for separate /usr.
>>> Mission accomplished.
>>> It's the upcoming change, that violates KISS. If udev cannot work properly
>>> with separate /usr, fix udev not the FS-hierarchy. What next? Put /home into
>>> initramfs, because udev decides it cannot work without /home mounted?
>>
>> Then don't upgrade. Keep doing only security updates.
>
> And, eventually, those security updates will stop coming. Just
> pointing this out.

Agree: but then again, that's the price you pay for not rolling with
the change. If you are tenacious enough, you could roll your own
security updates.

>>> It works now.
>>
>> Exactly, and if you don't upgrade, it will work as long as you want.
>
> See my remark on security updates not continuing indefinitely.
>
>>> Not listening to users is a very bad idea.
>>
>> No, they listen to users. They just don't listen too every user,
>> because that's impossible. Maybe I'm wrong, but I think your setup is
>> in the minority of use-cases. Who they should listen to?
>>
>>> You keep talking about "complainers".
>>
>> If someone complains and doesn't code, it's a complainer. By
>> definition. If someone complains and code, it's creating alternative
>> technologies.
>>
>>> I'd say, we discuss things, as do the
>>> gentoo-devs on their list.
>>
>> I agree. I'm subscribed to both.
>>
>>> Yeah, "probably", that's why we discuss things.
>>
>> Again, we can discuss (or complain) until the sun is red. As long as
>> we don't give code back, it's basically academic.
>
> This isn't a discussion. This is a bunch of people offering
> displeasure, ideas and/or thoughts, and one person saying, "hey,
> nothing I can do. I trust the devs."

I disagree: I'm not saying trust the devs and shut up. I'm trying to
explain why *I* trust the devs and why *I* think is for the best for
someone else to do it. Nobody needs to agree with me, of course, I'm
just trying to explain my POV.

> A discussion is when there's an interchange of ideas, arguments,
> counterarguments, and the fleshing out of a new framework of thought.
> That kind of point/counterpoint is *vital* for architectural
> foresight. All I keep reading from you is, "if you think that will
> work, go write it." *No* writing for a problem of this scope is
> warranted without some extensive discussion, noting of edge cases and
> planning around the same.

Sorry you read that way; I keep trying to explain why I don't find a
separated /usr a necessity and why I don't think an initramfs is such
a big problem.

> People have been pointing out edge cases, use cases which are being
> disregarded, etc, and pretty much all they're getting back is "I don't
> see those as valid."

And I have  tried to explain why it's not economically feasible to
support every architecture and every set of configurations. Yeah, the
only two solutions I see is either roll up with the change, or
maintain it yourself.

That people don't like the answer doesn't mean is not true.

> Granted, you're kinda painting a target on your
> back by being the only one defending upstream's decision here

Someone has to. I've been using Gentoo almost eight years now, and I
usually don't participate in the dicussions, but I have seen in the
last years a trend to criticize the devs without actually considering
the alternatives.

Sometimes the devs do stupid things; but most of the times they really
think and come to a solution. And the affected users usually just see
how that solution affects them in the short term, instead of trying to
see the big picture and how affects the whole distribution, the
community, and the technological path that Linux is following.

>, but
> when someone pointed at an already-existing alternative, you simply
> said, "I doubt that'll be the solution."

And I doubt it. But I also said that they are more than welcome to try
wathever they want. I think the way I think; that's the whole point of
me trying to communicate it here.

> As for me, this will be a royal inconvenience, and may require the
> rebuilding of my primary machine. Still, I can deal. It'll mean
> learning how to build initramfs*, how to make sure it contains the
> needed tools, and probably a half-dozen other things I didn't even see
> coming when I set up this box last fall.

I compile my own kernels (no genkernel for me). I don't use modules
(except for the stupid scsi_wait_scan), and I didn't used an initramfs
until I started using systemd. The arguments for using it convinced
me, and I made the switch in all my machines.

I don't see why it would require to "rebuild" your primary machine,
but I don't know your configuration. I know that any sane
configuration would only require to install dracut and modify a line
in grub. Maybe a kernel rebuild.

> * I've avoided it for ten years it was a grossly unnecessary
> complexity for my systems. Now it sounds like it'll become a
> necessity.

Apparently. But is not "grossly" unnecessary: udev need it, and udev
solves a kinda complicated problem. Maybe mdev can solve it in a
simpler way.

But again, I doubt it.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 1:47 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 1:35 PM, pk <peterk2@coolmail.se> wrote:
>> On 2011-09-08 16:51, Canek Peláez Valdés wrote:
>>
>>> But the freedom is still there. The freedom to either keep your system
>>> as it is (don't upgrade)
>>            ^^^^^^^^^^^^^
>> You do realise that this is quite valid for Windows (and all other OS's
>> in existence)? At least so far...
>
> Don't get *me* started. My _day job_ is C++/MFC on Windows. _Please_
> upgrade, you'll make my life much easier.
>
> Outdated operating systems make baby coder cry.

I already mentioned that you update security flaws.

And again, that's only if you resist the change.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 2:05 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> On Thu, 8 Sep 2011 12:56:44 -0400, Canek Peláez Valdés wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> On Thu, Sep 8, 2011 at 12:44 PM, David W Noon <dwnoon@ntlworld.com>
>> wrote:
> [snip]
>> > I expect to switch my simpler systems away from udev to mdev.  This
>> > loses some functionality of udev, but that isn't needed on the
>> > simpler hardware configurations.  So mdev could be the simplest
>> > solution to the design flaws creeping into udev.
>>
>> Maybe. I would not bet on it, but any new technical experiment is
>> worth trying, I believe. I will stick with the kernel-blessed option
>> of udev, though.
>
> I don't know if the kernel offers any particular blessing to any
> hotplug handler.

udev is the device manager for the Linux kernel. It replaced devfs.

It's related, but doesn't (necessarily) need to be the same that the
user space part.

Yeah, udev is mandatory in the kernel, unless you use a traditional
/dev directory.

>> > A very real problem with a large initramfs/initrd is maintaining the
>> > software embedded in the image file.  If it contains duplicates of
>> > e2fsck, reiserfsck, glibc, libpthread, etc., then these typically
>> > need to be upgraded whenever the primary copy is upgraded.  The
>> > bigger the initramfs becomes, the bigger the maintenance headache
>> > it inflicts.
>>
>> Dracut automatizes this. Is a non-problem.
>
> If dracut actually worked ...

What doesn't work for you?

> [snip]
>> mount -o remount,rw /
>> do stuff...
>> mount -o remount,ro /
>>
>> Really, I don't see the problem.
>
> During the "do stuff" phase, /usr is also writeable, which is
> undesirable on production systems.  That's the *original* problem with
> merging a read-only /usr with /. [We seem to be going in circles with
> this one.]

It's the same when you upgrade the system. If you don't allow rw in
/user *ever*, then you are not allowed to upgrade. Which I was chewed
up because I said it was an alternative.

>> > Similarly, /etc/mtab needs to remain writeable, as symlinking it
>> > to /proc/mounts (or /proc/self/mounts) won't always work for
>> > programs that parse /etc/mtab.  This is because /proc/mounts
>> > contains additional mount options that are fairly Linux-specific,
>> > whereas /etc/mtab should be vanilla UNIX.
>>
>> I really, really don't care about non-Linux systems. But that's me,
>> anyone else can use wathever they want. Just don't expect everyone to
>> be happy with the lowest common feature set.
>>
>> Having said that, which programs do you use that need to parse mtab?
>
> I have about 6 or 7 backup jobs that run during the night and
> parse /etc/mtab to see if they need to place a copy of the backup onto
> an external medium.  These examine the mount options and don't
> understand the non-standard options offered by Linux in /proc/mounts.

Really? You cannot grep -v those options to another file and make the
jobs read this other file?

In my experience that sounds like a problem with the jobs. But hey,
that only was an option (making / ro). You have several options,
already discussed.

And again, it's not the regular use-case. The devs don't need to worry
about your use-case (or my use-case); they worry about the global
picture (or at least should).

You can either roll with that or maintain it yourself. If I were you,
I would try to change those jobs. Otherwise, I would use an initramfs.
Either way, I would fix it.

> [snip]
>> Just don't expect the limited resources of the Gentoo devs to be able
>> to test and support your special configuration.
>
> :-))
>
> They already don't do that.

Well, then you already know what to do.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 3:00 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 1:45 PM, Michael Mol <mikemol@gmail.com> wrote:
>> This isn't a discussion. This is a bunch of people offering
>> displeasure, ideas and/or thoughts, and one person saying, "hey,
>> nothing I can do. I trust the devs."
>
> I disagree: I'm not saying trust the devs and shut up. I'm trying to
> explain why *I* trust the devs and why *I* think is for the best for
> someone else to do it. Nobody needs to agree with me, of course, I'm
> just trying to explain my POV.



>
>> A discussion is when there's an interchange of ideas, arguments,
>> counterarguments, and the fleshing out of a new framework of thought.
>> That kind of point/counterpoint is *vital* for architectural
>> foresight. All I keep reading from you is, "if you think that will
>> work, go write it." *No* writing for a problem of this scope is
>> warranted without some extensive discussion, noting of edge cases and
>> planning around the same.
>
> Sorry you read that way; I keep trying to explain why I don't find a
> separated /usr a necessity and why I don't think an initramfs is such
> a big problem.

The rhythmic baseline of your explanation is "and you can use an
initramfs for that." I haven't seen much in argument for why initramfs
isn't a problem apart from your expectation that the *kernel* will
eventually require it, and dracut theoretically solve the problem of
updating initrd. (Which sounds nice, but it's yet another moving part
in the system, and another point of failure.)

>> People have been pointing out edge cases, use cases which are being
>> disregarded, etc, and pretty much all they're getting back is "I don't
>> see those as valid."
>
> And I have  tried to explain why it's not economically feasible to
> support every architecture and every set of configurations.

Which I think has been firmly explained in at least two different
threads; dev time is not infinite, sure.

> Yeah, the only two solutions I see is either roll up with the change, or
> maintain it yourself.

If those are the only two solutions you see, I suspect you're not
looking hard enough. If udev's problem is that it's running arbitrary
programs, then perhaps a recognizable constraint needs to be made on
what programs udev should run. I thought the primary reason we had
/{,s}bin and /usr/{,s}bin was to differentiate between system-vital
programs and other programs.

> That people don't like the answer doesn't mean is not true.

There's also no solid evidence that it's true, either. I remember when
sysfs came about. Linux Journal's diff -u column described it as a
herculean effort accomplishing something the majority of kernel devs
thought impossible or not worth the time. I rather like sysfs, but at
one time, it was the thing very few people thought was in the possible
set of solutions.

>> Granted, you're kinda painting a target on your
>> back by being the only one defending upstream's decision here
>
> Someone has to. I've been using Gentoo almost eight years now, and I
> usually don't participate in the dicussions, but I have seen in the
> last years a trend to criticize the devs without actually considering
> the alternatives.

People are beginning to consider alternatives, but you've shot them
down without offering improvements, suggesting adjustments or even
pointing out specific flaws. As a result, you come across as something
akin to a fanboy with your mind already made up, which just seems
*bizarre*.

>
> Sometimes the devs do stupid things; but most of the times they really
> think and come to a solution. And the affected users usually just see
> how that solution affects them in the short term, instead of trying to
> see the big picture and how affects the whole distribution, the
> community, and the technological path that Linux is following.

For being irritated that people aren't seeing it, you haven't been
clarifying it much.

>>, but
>> when someone pointed at an already-existing alternative, you simply
>> said, "I doubt that'll be the solution."
>
> And I doubt it. But I also said that they are more than welcome to try
> wathever they want. I think the way I think; that's the whole point of
> me trying to communicate it here.

That much is appreciated, but you didn't say *why* you doubt that'll
be the solution. Downvotes aren't debate, nor are they discussion.
They're expressions of dissatisfaction.

>> As for me, this will be a royal inconvenience, and may require the
>> rebuilding of my primary machine. Still, I can deal. It'll mean
>> learning how to build initramfs*, how to make sure it contains the
>> needed tools, and probably a half-dozen other things I didn't even see
>> coming when I set up this box last fall.
>
> I compile my own kernels (no genkernel for me). I don't use modules
> (except for the stupid scsi_wait_scan), and I didn't used an initramfs
> until I started using systemd. The arguments for using it convinced
> me, and I made the switch in all my machines.
>
> I don't see why it would require to "rebuild" your primary machine,
> but I don't know your configuration. I know that any sane
> configuration would only require to install dracut and modify a line
> in grub. Maybe a kernel rebuild.

My / isn't large enough to hold /usr.

>> * I've avoided it for ten years it was a grossly unnecessary
>> complexity for my systems. Now it sounds like it'll become a
>> necessity.
>
> Apparently. But is not "grossly" unnecessary: udev need it, and udev
> solves a kinda complicated problem. Maybe mdev can solve it in a
> simpler way.
>
> But again, I doubt it.

One question: Why? Are there specific technical reasons to your
doubts, or is it simply confidence that the devs are more likely to
have made a good choice than a bad choice?

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 3:01 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 1:47 PM, Michael Mol <mikemol@gmail.com> wrote:
>> On Thu, Sep 8, 2011 at 1:35 PM, pk <peterk2@coolmail.se> wrote:
>>>            ^^^^^^^^^^^^^
>>> You do realise that this is quite valid for Windows (and all other OS's
>>> in existence)? At least so far...
>>
>> Don't get *me* started. My _day job_ is C++/MFC on Windows. _Please_
>> upgrade, you'll make my life much easier.
>>
>> Outdated operating systems make baby coder cry.
>
> I already mentioned that you update security flaws.
>
> And again, that's only if you resist the change.

Chill; that was a side-thread joke. :)

I was looking at XP->Vista->Win7, anyway. The newer operating systems'
kernels have features which would make my life much easier, if I could
rely on their presence.


-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 11:13:58 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> > Have you *ever* thought about machines, that are not x86 or x86_64?
> > Here's an intersting read:
> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769  
> 
> No, I haven't thought about them, because I don't use them. What it
> has to do with anything?

Linux runs on many many many more arches than just x86 and amd64. IIRC
it's about 29 in total.

Do you agree that the needs and requirements of all those other arches
might be important to everything else?

You keep mentioning "it will all work if you just use an initramfs".
Did it occur to you that that statement is the entire problem and
demonstrates the problem nicely? I do not have an initramfs, do not
need one, see no need to have one and have not yet seen a valid
technical reason for why having one is ideal. My gentoo systems do not
run binary distros, I have no need for a generic mechanism designed to
cope with any hardware Fedora might happen to find itself booting on,
hardware that the devs have no idea of when they compile their distros.
I, on the other hand, already know everything I need to know about my
hardware for the purposes of booting, running udev and building a valid
kernel that fits my needs.

Tell me again what it is that validly requires me to switch to some new
way of doing things?

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 3:37 PM, Michael Mol <mikemol@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 3:00 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Thu, Sep 8, 2011 at 1:45 PM, Michael Mol <mikemol@gmail.com> wrote:
>>> This isn't a discussion. This is a bunch of people offering
>>> displeasure, ideas and/or thoughts, and one person saying, "hey,
>>> nothing I can do. I trust the devs."
>>
>> I disagree: I'm not saying trust the devs and shut up. I'm trying to
>> explain why *I* trust the devs and why *I* think is for the best for
>> someone else to do it. Nobody needs to agree with me, of course, I'm
>> just trying to explain my POV.
>>
>>> A discussion is when there's an interchange of ideas, arguments,
>>> counterarguments, and the fleshing out of a new framework of thought.
>>> That kind of point/counterpoint is *vital* for architectural
>>> foresight. All I keep reading from you is, "if you think that will
>>> work, go write it." *No* writing for a problem of this scope is
>>> warranted without some extensive discussion, noting of edge cases and
>>> planning around the same.
>>
>> Sorry you read that way; I keep trying to explain why I don't find a
>> separated /usr a necessity and why I don't think an initramfs is such
>> a big problem.
>
> The rhythmic baseline of your explanation is "and you can use an
> initramfs for that." I haven't seen much in argument for why initramfs
> isn't a problem apart from your expectation that the *kernel* will
> eventually require it, and dracut theoretically solve the problem of
> updating initrd. (Which sounds nice, but it's yet another moving part
> in the system, and another point of failure.)

No, I think you haven't been reading carefully enough. Again:

1. In 2011, we need a dynamic /dev tree. I'm not going to argue why.
2. udev, successor of devfs, which was successor of the classical /dev
tree, after years of design and development iterations, solves the
problem. It's not perfect, but I think that is as close as it could
be, for the problem it tries to solve, and with the feature set it
has.
3. udev needs either an initramfs, because it needs an early user
space, or a /usr inside /.

From this 3 points, I make my conclusion: keep up with the changes, or
code an alternative (that includes using something like mdev).

>>> People have been pointing out edge cases, use cases which are being
>>> disregarded, etc, and pretty much all they're getting back is "I don't
>>> see those as valid."
>>
>> And I have  tried to explain why it's not economically feasible to
>> support every architecture and every set of configurations.
>
> Which I think has been firmly explained in at least two different
> threads; dev time is not infinite, sure.

Agree.

>> Yeah, the only two solutions I see is either roll up with the change, or
>> maintain it yourself.
>
> If those are the only two solutions you see, I suspect you're not
> looking hard enough. If udev's problem is that it's running arbitrary
> programs, then perhaps a recognizable constraint needs to be made on
> what programs udev should run. I thought the primary reason we had
> /{,s}bin and /usr/{,s}bin was to differentiate between system-vital
> programs and other programs.

That enters perflectly in my other solution: code an alternative. That
includes changing (or trying to) udev.

>> That people don't like the answer doesn't mean is not true.
>
> There's also no solid evidence that it's true, either.

Prove me wrong, or find the code that proves me wrong. If it's that
important to you, do it. My whole point is that it's easier (and in
the long term more beneficial) to roll with the changes.

> I remember when
> sysfs came about. Linux Journal's diff -u column described it as a
> herculean effort accomplishing something the majority of kernel devs
> thought impossible or not worth the time. I rather like sysfs, but at
> one time, it was the thing very few people thought was in the possible
> set of solutions.

I don't see the connection. If someone actually goes and writes the
code for an alternative to udev (or /usr inside /, or an initramfs),
then it enters my second alternative.

>>> Granted, you're kinda painting a target on your
>>> back by being the only one defending upstream's decision here
>>
>> Someone has to. I've been using Gentoo almost eight years now, and I
>> usually don't participate in the dicussions, but I have seen in the
>> last years a trend to criticize the devs without actually considering
>> the alternatives.
>
> People are beginning to consider alternatives, but you've shot them
> down without offering improvements, suggesting adjustments or even
> pointing out specific flaws. As a result, you come across as something
> akin to a fanboy with your mind already made up, which just seems
> *bizarre*.

I don't really care if people mind me a fanboi: I know I'm not. I'm
old enoguh to be a fanold, though.

Seriously, maybe I'm not communicating my point clearly. The only
alternative mentioned (I think: please correct me if I'm wrong) is
mdev... which I didn't shoot down, I just mentioned that I don't think
will solve the problem. And if you have followed the development of
Linux as I have, then you know it is a *really* hard problem, and that
udev is the result of many man-hours of thinking, designing and
implementing. That's why I don't give to much hope to a project I have
never heard about.

That doesn't mean I'm not mistaken, of course.

>>
>> Sometimes the devs do stupid things; but most of the times they really
>> think and come to a solution. And the affected users usually just see
>> how that solution affects them in the short term, instead of trying to
>> see the big picture and how affects the whole distribution, the
>> community, and the technological path that Linux is following.
>
> For being irritated that people aren't seeing it, you haven't been
> clarifying it much.

When did I say I was being irritated? I'm only trying to express my POV.

>>>, but
>>> when someone pointed at an already-existing alternative, you simply
>>> said, "I doubt that'll be the solution."
>>
>> And I doubt it. But I also said that they are more than welcome to try
>> wathever they want. I think the way I think; that's the whole point of
>> me trying to communicate it here.
>
> That much is appreciated, but you didn't say *why* you doubt that'll
> be the solution. Downvotes aren't debate, nor are they discussion.
> They're expressions of dissatisfaction.

Why should I be dissatisfied? The development of Linux goes exactly in
the direction *I* want, I don't complain about any of the required
changes discussed.

Why I doubt it I answered some paragraphs above.

>>> As for me, this will be a royal inconvenience, and may require the
>>> rebuilding of my primary machine. Still, I can deal. It'll mean
>>> learning how to build initramfs*, how to make sure it contains the
>>> needed tools, and probably a half-dozen other things I didn't even see
>>> coming when I set up this box last fall.
>>
>> I compile my own kernels (no genkernel for me). I don't use modules
>> (except for the stupid scsi_wait_scan), and I didn't used an initramfs
>> until I started using systemd. The arguments for using it convinced
>> me, and I made the switch in all my machines.
>>
>> I don't see why it would require to "rebuild" your primary machine,
>> but I don't know your configuration. I know that any sane
>> configuration would only require to install dracut and modify a line
>> in grub. Maybe a kernel rebuild.
>
> My / isn't large enough to hold /usr.

I don't understand. You said you were to try building an initramfs,
and if that's the case, then you can keep /usr separated.

If you will put /usr in / (and hence rebuild your system), then why
you said you would learn how to built an initramfs?

>>> * I've avoided it for ten years it was a grossly unnecessary
>>> complexity for my systems. Now it sounds like it'll become a
>>> necessity.
>>
>> Apparently. But is not "grossly" unnecessary: udev need it, and udev
>> solves a kinda complicated problem. Maybe mdev can solve it in a
>> simpler way.
>>
>> But again, I doubt it.
>
> One question: Why? Are there specific technical reasons to your
> doubts, or is it simply confidence that the devs are more likely to
> have made a good choice than a bad choice?

I answered that already (actually, in that paragraph). But again: udev
is not trivial, and it solves a (far from) trivial problem. If some
developers think they can outsmart the kernel devs, please, lets try
it. Maybe they will.

But I'm not holding my breath.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 08 Sep 2011 19:11:04 +0200
Michael Schreckenbauer <grimlog@gmx.de> wrote:

> > Then design the correct solution and implement it. If it's
> > technically sound, it will prevail. I think it's a rather
> > complicated problem with a non trivial solution, but the code is
> > there if you feel like give it a try.  
> 
> Where did I write, that I am in the position to write such a beast?
> I only take the freedom to name this a design flaw in udev.
> It needs things from userspace, which are not yet available at the
> point it requests them. An initramsfs is a workaround for this, not a
> proper fix.

If that is the argument from the udev devs you just quoted, then I do
not understand it at all.

Why can there not be a restriction that udev may only run code in the
traditional / space (i.e. it will not attempt to run code in the /usr
or /home spaces)?

Device nodes are a root function; root is the only user that should
dictate how device nodes are created; root is the only user that can
normally write to / and thereby create udev's rules and rulesets.

In what valid way does access to /usr become something that udev may be
required to support?

Not arguing with *you* here Michael, just wondering about the validity
of the position you quoted

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 3:48 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 8 Sep 2011 11:13:58 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> > Have you *ever* thought about machines, that are not x86 or x86_64?
>> > Here's an intersting read:
>> > http://permalink.gmane.org/gmane.linux.gentoo.devel/72769
>>
>> No, I haven't thought about them, because I don't use them. What it
>> has to do with anything?
>
> Linux runs on many many many more arches than just x86 and amd64. IIRC
> it's about 29 in total.
>
> Do you agree that the needs and requirements of all those other arches
> might be important to everything else?

I'm not saying it's not important, I'm saying that the needs of the
less used archs cannot slow down the development of the most popular
ones. Again, is an economic reason: we don't have enough devs.

> You keep mentioning "it will all work if you just use an initramfs".
> Did it occur to you that that statement is the entire problem and
> demonstrates the problem nicely?

No, because it's not a solution decided because of laziness, which
many here seem to think it is.

> I do not have an initramfs, do not
> need one, see no need to have one and have not yet seen a valid
> technical reason for why having one is ideal.

It's not "ideal" (I don't think anybody has said that). Almost nothing
is "ideal" in computer science.

Maybe it's not enough for you, but I repeat: we need dynamic /dev
trees, udev giveus that, the udev code lives in user space, we need an
early user space => initramfs.

> My gentoo systems do not
> run binary distros, I have no need for a generic mechanism designed to
> cope with any hardware Fedora might happen to find itself booting on,
> hardware that the devs have no idea of when they compile their distros.

Hey, I compile all my modules inside my kernels. That has nothing to
do with udev, because you can connect via USB or eSATA *any* hardware
into your computer, and the /dev tree needs to update dynamically.

Maybe *you* don't want that, and that's fine: but the majority of
users do want that. Your use-case is not the most important one in the
whole world.

> I, on the other hand, already know everything I need to know about my
> hardware for the purposes of booting, running udev and building a valid
> kernel that fits my needs.

Me too. And yet, I use an initramfs for the pretty plymouth splash screen.

> Tell me again what it is that validly requires me to switch to some new
> way of doing things?

Nothing: if you don't like it, don't use it (or use an alternative),
or change it.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:05 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 08 Sep 2011 19:11:04 +0200
> Michael Schreckenbauer <grimlog@gmx.de> wrote:
>
>> > Then design the correct solution and implement it. If it's
>> > technically sound, it will prevail. I think it's a rather
>> > complicated problem with a non trivial solution, but the code is
>> > there if you feel like give it a try.
>>
>> Where did I write, that I am in the position to write such a beast?
>> I only take the freedom to name this a design flaw in udev.
>> It needs things from userspace, which are not yet available at the
>> point it requests them. An initramsfs is a workaround for this, not a
>> proper fix.
>
> If that is the argument from the udev devs you just quoted, then I do
> not understand it at all.
>
> Why can there not be a restriction that udev may only run code in the
> traditional / space (i.e. it will not attempt to run code in the /usr
> or /home spaces)?
>
> Device nodes are a root function; root is the only user that should
> dictate how device nodes are created; root is the only user that can
> normally write to / and thereby create udev's rules and rulesets.
>
> In what valid way does access to /usr become something that udev may be
> required to support?

It is a matter of what else do you end having in /bin and /lib.
Remember that udev rules can execute arbitrary code. Do all that code
needs to be moved to /bin and /lib also?

I keep telling: it is a difficult problem.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 11:32:27 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> On Thu, Sep 8, 2011 at 6:11 AM, Alan McKinnon
> <alan.mckinnon@gmail.com> wrote:
> > An initramfs is optional becuase i can disable it in the kernel. I
> > would like to keep that optional.
> 
> udev at some point was optional, then it wasn't. Right now initramfs
> is optional primarily because of embedded systems.

initramfs exists primarily because there's no other sane way to boot a
binary distro without one and still have something approximating the
goal of only having what's required. The other case is booting off a
volume that the boot loader does not support (such as / on LVM)

I have yet to see a valid reason for requiring an initramfs for normal
use. Every reason advanced so far to my mind reduces to "because I
changed the code to work that way".

Do note that I'm not asking the devs to engage is some peculiar code
writing or support something unusual. I'm asking them to leave the
existing code in place because it works just fine without security
risks or difficult technical issues (as far as I can see)
 
> Change happens, I repeat.
> 
> > FHS says I can have /usr on a separate partition and I would like to
> > keep that because it is a good idea.
> 
> FHS is dead: for years we didn't hear from it, and it was until a few
> months that some activity was registered from it. For practical
> reasons, it's dead, and nobody follows it completely (where in FHS is
> /usr/libexec? do you use /srv like FHS says?)

Just because FHS is not being tweaked often does not mean it is no
longer valid. It contains many excellent ideas, such as the possibility
of a minimal / with only 4 required directories containing everything
needed to boot and repair.

FHS never said anywhere that one could not have extra directories
like /usr/libexec, one is free to make and use such if one wishes.

Yes, I do use /srv. I do like to keep my externally shared data (ftp,
web etc) in a different place to my internally shared data (portage
caches etc). I like and use this idea because I think it's a good idea.
But I'm not required to and I understand that. /srv is merely a good
idea that can assist with maintenance, it is nowhere near being the
same class as the minimal contents of /, this is a core design feature
that determines how the entire system as a whole works. Not really
comparable

> 
> But, if you think /usr in a separate partition is a good idea, then by
> all means write the code for it.

Why not let me keep the code I already have, which is not broken?

I have yet to see a valid reason why this sane default is no longer
valid.


> > FHS says I can mount /usr read-only if I choose, which is also a
> > good idea. On a shared jumphost with 570 concurrent users it's
> > actually a VERY GOOD ODEA and I'd rather not lose that facility
> > thankyouverymuch.
> 
> Then don't loose it. Just use an initramfs.
> 
> > I do not need, want nor can I find a valid reason to *require* an
> > initramfs. Systems boot just fine without them.
> 
> Then either restrain yourself to security updates (which may be a good
> idea if you support a server for 570 concurrent users), or write the
> code to support a separated /usr without initramfs.

I'd like this aspect to remain just as it is currently. I see no defect
with it and you are not advancing one.
>
> > FHS says I can have the minimal software and tools to effect a
> > system repair on / and put then entirety of user-space on /usr.
> > Everything involved in this thread runs early in the boot process
> > and I fail to find a single convincing reason why /usr is involved
> > at all. Anything required at this point can simply be put
> > into /bin, /sbin and /lib{,64} which one will note is exactly how
> > we have been doing it all along.
> 
> If it is so easy, then write the code to do it.

I do not need to write this code. It already exists and is shipped with
the standard tarballs I build Gentoo systems from right now.

> > This whole mess has every indication of a singular maintainer who
> > cannot be bothered taking other people's needs into account and
> > foisting off his own personal preferences onto an entire ecosystem.
> 
> And he magically convinces his distribution (and ours) to follow
> through? Man, he must be really powerful. I don't think that it is
> even possible that *maybe* some of his reasoning actually makes sense.

It may make sense - I will of course concede that, but I have yet to
see a valid explanation of the reasoning. Once again, it may exist, it
simply hasn't become visible to me.

> > I think such people should take note of how Torvalds works and
> > emulate him as opposed to emulating say Drepper as a role-model for
> > good project mantainership practice.
> 
> The people that writes the code, gets shit done. Code talks.

And sometimes the guy that writes code fucks it up enormously.

The onus for proving my ideas are good rests with me, I accept this.
Similarly, the onus is on the author of a new feature to demonstrate
that his idea is not dangerous, and if it is to offer valid reasons as
to why it is preferable to proceed. I don't see either of these
arguments from the udev maintainer.
> 
> We always have the option to write the
>code ourselves, and get shit
> done the way we want it. Don't want to do that? Then accept that you
> will need to follow what the writers of the code decide.

Why is it so hard for these same maintainers to leave things as they
are? It is not broken currently

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 3206 bytes --]

On Thu, 8 Sep 2011 15:13:55 -0400, Canek Peláez Valdés wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Thu, Sep 8, 2011 at 2:05 PM, David W Noon <dwnoon@ntlworld.com>
> wrote:
[snip]
> > I don't know if the kernel offers any particular blessing to any
> > hotplug handler.
> 
> udev is the device manager for the Linux kernel. It replaced devfs.

One can use mdev just as readily as udev.

> It's related, but doesn't (necessarily) need to be the same that the
> user space part.
> 
> Yeah, udev is mandatory in the kernel, unless you use a traditional
> /dev directory.

But udev isn't actually part of the kernel.  Only hotplug support is
actually in the kernel.  The udev daemon is started during the sysinit
run-level and it connects itself to hotplug support.

[snip]
> >> Dracut automatizes this. Is a non-problem.
> >
> > If dracut actually worked ...
> 
> What doesn't work for you?

Since dracut is not yet stable, I don't have any problems with it
because I don't use it.  But it does have quite a few open bugs in
Gentoo's Bugzilla, and I suspect many more in other distro's bug
trackers.

> > During the "do stuff" phase, /usr is also writeable, which is
> > undesirable on production systems.  That's the *original* problem
> > with merging a read-only /usr with /. [We seem to be going in
> > circles with this one.]
> 
> It's the same when you upgrade the system. If you don't allow rw in
> /user *ever*, then you are not allowed to upgrade. Which I was chewed
> up because I said it was an alternative.

Production systems have strictly scheduled change-control windows,
usually only once or twice a year.  Having to schedule database changes
to match application change-control would not be workable.  That is
why /etc cannot be mounted read-only and still have /usr secured as
read-only.  This brings us back to a requirement that / and /usr be
physically separate filesystems.

[snip]
> > I have about 6 or 7 backup jobs that run during the night and
> > parse /etc/mtab to see if they need to place a copy of the backup
> > onto an external medium.  These examine the mount options and don't
> > understand the non-standard options offered by Linux
> > in /proc/mounts.
> 
> Really? You cannot grep -v those options to another file and make the
> jobs read this other file?

I would use gawk rather than grep.  But since I have code that already
works, why should I need to develop a new script?

> In my experience that sounds like a problem with the jobs.

They work currently.

Moreover, my rootfs is not read-only.  It is not desirable to have the
rootfs mounted read-only because of this problem and the other
problems it causes.  But for production systems it is desirable for /usr
to be mounted read-only and only made writeable during a change-control
window.

[snip]
> > They already don't do that.
> 
> Well, then you already know what to do.

Indeed I do.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:24 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
[ big fat snip ]
> Why is it so hard for these same maintainers to leave things as they
> are? It is not broken currently

I really need to get some work done, so I will keep it shortly,
because this is a dup of the CUPS discussion:

With code, you cannot have your cake and eat it. If you want the new
features, you need to follow the developers direction. Keeping
existing functionality has a cost, which in some cases exceeds the
cost of a new version of the code that forces us to do things
differently.

I htink almost everyone understand this.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 16:21:11 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> > I do not have an initramfs, do not
> > need one, see no need to have one and have not yet seen a valid
> > technical reason for why having one is ideal.  
> 
> It's not "ideal" (I don't think anybody has said that). Almost nothing
> is "ideal" in computer science.
> 
> Maybe it's not enough for you, but I repeat: we need dynamic /dev
> trees, udev giveus that, the udev code lives in user space, we need an
> early user space => initramfs.

I didn't say I don't use udev, I do. I too have cameras, USB gadgets
and a huge array of possible hotplug objects in the shops I can buy at
any time. udev makes that all work well.

I don't agree with the assertion that "user space => initramfs".

You obviously must start udev as soon as possible in the boot process.
For it to work at all, one of the minimum requirements is something
mounted at / containing udev rules. This can be an initramfs or a
physical disk or anything else that can possibly behave as a block
device. I know of nothing in the kernel that *requires* it to be an
initramfs. The code should be generic enough that I can mount whatever
I want, then do whatever I need to do within limits and finally pivot
mount the real /

I don't see a reasonable argument as to why things cannot continue to
behave just like this.

> 
> > My gentoo systems do not
> > run binary distros, I have no need for a generic mechanism designed
> > to cope with any hardware Fedora might happen to find itself
> > booting on, hardware that the devs have no idea of when they
> > compile their distros.  
> 
> Hey, I compile all my modules inside my kernels. That has nothing to
> do with udev, because you can connect via USB or eSATA *any* hardware
> into your computer, and the /dev tree needs to update dynamically.
> 
> Maybe *you* don't want that, and that's fine: but the majority of
> users do want that. Your use-case is not the most important one in the
> whole world.

I never said it is, I never said we don't need udev. I am saying in
this thread that I do not understand the new requirements for /usr -
everything there can be mandated to be in / instead where it
is guaranteed to be accessible to udev  
-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 22:05:36 schrieb Alan McKinnon:
> On Thu, 08 Sep 2011 19:11:04 +0200
> 
> Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > > Then design the correct solution and implement it. If it's
> > > technically sound, it will prevail. I think it's a rather
> > > complicated problem with a non trivial solution, but the code is
> > > there if you feel like give it a try.
> > 
> > Where did I write, that I am in the position to write such a beast?
> > I only take the freedom to name this a design flaw in udev.
> > It needs things from userspace, which are not yet available at the
> > point it requests them. An initramsfs is a workaround for this, not a
> > proper fix.
> 
> If that is the argument from the udev devs you just quoted, then I do
> not understand it at all.

It's my understanding, that this is their point.

> Why can there not be a restriction that udev may only run code in the
> traditional / space (i.e. it will not attempt to run code in the /usr
> or /home spaces)?

Yes. I really wonder, why we have /bin, /sbin and /lib
 
> Device nodes are a root function; root is the only user that should
> dictate how device nodes are created; root is the only user that can
> normally write to / and thereby create udev's rules and rulesets.
> 
> In what valid way does access to /usr become something that udev may be
> required to support?

As udev is able to run arbitrary scripts, there *might* be some code, that 
requires something from /usr/*. So they want this beast be mounted, before 
udev starts doing it's job.
 
> Not arguing with *you* here Michael, just wondering about the validity
> of the position you quoted

Understood :)

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:25 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> On Thu, 8 Sep 2011 15:13:55 -0400, Canek Peláez Valdés wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> On Thu, Sep 8, 2011 at 2:05 PM, David W Noon <dwnoon@ntlworld.com>
>> wrote:
> [snip]
>> > I don't know if the kernel offers any particular blessing to any
>> > hotplug handler.
>>
>> udev is the device manager for the Linux kernel. It replaced devfs.
>
> One can use mdev just as readily as udev.
>
>> It's related, but doesn't (necessarily) need to be the same that the
>> user space part.
>>
>> Yeah, udev is mandatory in the kernel, unless you use a traditional
>> /dev directory.
>
> But udev isn't actually part of the kernel.  Only hotplug support is
> actually in the kernel.  The udev daemon is started during the sysinit
> run-level and it connects itself to hotplug support.

And what do you think the udev daemon speaks to?

> [snip]
>> >> Dracut automatizes this. Is a non-problem.
>> >
>> > If dracut actually worked ...
>>
>> What doesn't work for you?
>
> Since dracut is not yet stable, I don't have any problems with it
> because I don't use it.  But it does have quite a few open bugs in
> Gentoo's Bugzilla, and I suspect many more in other distro's bug
> trackers.

Well, dracut's job is not rocket science.

>> > During the "do stuff" phase, /usr is also writeable, which is
>> > undesirable on production systems.  That's the *original* problem
>> > with merging a read-only /usr with /. [We seem to be going in
>> > circles with this one.]
>>
>> It's the same when you upgrade the system. If you don't allow rw in
>> /user *ever*, then you are not allowed to upgrade. Which I was chewed
>> up because I said it was an alternative.
>
> Production systems have strictly scheduled change-control windows,
> usually only once or twice a year.  Having to schedule database changes
> to match application change-control would not be workable.  That is
> why /etc cannot be mounted read-only and still have /usr secured as
> read-only.  This brings us back to a requirement that / and /usr be
> physically separate filesystems.
>
> [snip]
>> > I have about 6 or 7 backup jobs that run during the night and
>> > parse /etc/mtab to see if they need to place a copy of the backup
>> > onto an external medium.  These examine the mount options and don't
>> > understand the non-standard options offered by Linux
>> > in /proc/mounts.
>>
>> Really? You cannot grep -v those options to another file and make the
>> jobs read this other file?
>
> I would use gawk rather than grep.  But since I have code that already
> works, why should I need to develop a new script?
>
>> In my experience that sounds like a problem with the jobs.
>
> They work currently.

So you want all the new functionality, but without needing to do anything.

I want pink ponies too. Just not gonna happen.

> Moreover, my rootfs is not read-only.  It is not desirable to have the
> rootfs mounted read-only because of this problem and the other
> problems it causes.  But for production systems it is desirable for /usr
> to be mounted read-only and only made writeable during a change-control
> window.

Then use an initramfs and get /usr separated. You can do it on one of
your twice a year down times.

> [snip]
>> > They already don't do that.
>>
>> Well, then you already know what to do.
>
> Indeed I do.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez Valdés:
> > In what valid way does access to /usr become something that udev may be
> > required to support?
> 
> It is a matter of what else do you end having in /bin and /lib.
> Remember that udev rules can execute arbitrary code. Do all that code
> needs to be moved to /bin and /lib also?

Of course. That's what /bin, /sbin and /lib are for.

> I keep telling: it is a difficult problem.

No. Just move or copy the binaries and libs *you* use for *your* udev-scripts 
to /bin, /sbin and /lib

> Regards.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 21:25:31 +0100
David W Noon <dwnoon@ntlworld.com> wrote:

> > Well, then you already know what to do.  
> 
> Indeed I do.

If your databases are not ASE or Oracle, then migrating to FreeBSD is a
fine option. It's what we did at work. The benefits were immediate:

- we didn't lose the functioning desktop we didn't use
- I/O throughput improved
- maintenance simplified
- KISS, we now have traditional Unix systems that work how we want
  them to
- we need real Unix sysadmins so all the one-day wonders who attended
  Red Hat seminars went back to the Windows division
- you get all the good parts of a binary distro (a maintained
  ready-built base) plus all the good parts of portage (from ports)

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:38 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 8 Sep 2011 16:21:11 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> > I do not have an initramfs, do not
>> > need one, see no need to have one and have not yet seen a valid
>> > technical reason for why having one is ideal.
>>
>> It's not "ideal" (I don't think anybody has said that). Almost nothing
>> is "ideal" in computer science.
>>
>> Maybe it's not enough for you, but I repeat: we need dynamic /dev
>> trees, udev giveus that, the udev code lives in user space, we need an
>> early user space => initramfs.
>
> I didn't say I don't use udev, I do. I too have cameras, USB gadgets
> and a huge array of possible hotplug objects in the shops I can buy at
> any time. udev makes that all work well.
>
> I don't agree with the assertion that "user space => initramfs".
>
> You obviously must start udev as soon as possible in the boot process.
> For it to work at all, one of the minimum requirements is something
> mounted at / containing udev rules. This can be an initramfs or a
> physical disk or anything else that can possibly behave as a block
> device. I know of nothing in the kernel that *requires* it to be an
> initramfs. The code should be generic enough that I can mount whatever
> I want, then do whatever I need to do within limits and finally pivot
> mount the real /

The only simple answer I have is this one: if it is so simple, then it
would work this way.

It isn't. If I am wrong (and, of course, that is possible), someone
will come out and do it, and you guys will be able to keep your
separated /usr and no initramfs.

But I really don't think so. But, please, prove me wrong. I would
certainly be interesting to see another solution.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer <grimlog@gmx.de> wrote:
> Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez Valdés:
>> > In what valid way does access to /usr become something that udev may be
>> > required to support?
>>
>> It is a matter of what else do you end having in /bin and /lib.
>> Remember that udev rules can execute arbitrary code. Do all that code
>> needs to be moved to /bin and /lib also?
>
> Of course. That's what /bin, /sbin and /lib are for.
>
>> I keep telling: it is a difficult problem.
>
> No. Just move or copy the binaries and libs *you* use for *your* udev-scripts
> to /bin, /sbin and /lib

I *really* don't think bluetoothd belongs to /sbin. But, hey, that's me.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 08 Sep 2011 22:40:07 +0200
Michael Schreckenbauer <grimlog@gmx.de> wrote:

> Am Donnerstag, 8. September 2011, 22:05:36 schrieb Alan McKinnon:
> > On Thu, 08 Sep 2011 19:11:04 +0200
> > 
> > Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > > > Then design the correct solution and implement it. If it's
> > > > technically sound, it will prevail. I think it's a rather
> > > > complicated problem with a non trivial solution, but the code is
> > > > there if you feel like give it a try.
> > > 
> > > Where did I write, that I am in the position to write such a
> > > beast? I only take the freedom to name this a design flaw in udev.
> > > It needs things from userspace, which are not yet available at the
> > > point it requests them. An initramsfs is a workaround for this,
> > > not a proper fix.
> > 
> > If that is the argument from the udev devs you just quoted, then I
> > do not understand it at all.
> 
> It's my understanding, that this is their point.
> 
> > Why can there not be a restriction that udev may only run code in
> > the traditional / space (i.e. it will not attempt to run code in
> > the /usr or /home spaces)?
> 
> Yes. I really wonder, why we have /bin, /sbin and /lib

The / partition may contain, at a absolute minimum, only that software
required to boot and start userspace. So you find mount, fsck and sh in
there.

Everything else that non-root users use routinely (and not required to
boot and repair) is in /usr.

The software on / splits into three logical categories:

- code that only works when run as root in /sbin. Example: fsck
- code that is required on /, but regular users may use it under
  normal circumstances, it cannot go in /sbin so it is in /bin.
  Example: mount
- libraries, which go in /lib

The distinction between bin and sbin is traditional and not really
required. It comes from the days when resources were scarce and /bin
and /sbin were a little too big combined to work efficiently. A nice
side-effect is that /sbin could go in root's PATH so the code in it
would not clutter up the user's space with tab-completion and which.
This is not a security feature, just a convenience (same thinking as
dot files).

The same three directories exist under /usr, where the system package
manager will maintain the contents. And the same three directories
exist under /usr/local which is where the local sysadmin installs stuff
he compiles himself and where the package manager won't touch it.

Pre-compiled binaries such as Skype, Opera and database products go
in /opt, completely contained in one directory tree (much like Windows
and MacOS). This too is good as you can remove an entire product's
files by deleting one directory tree without having to worry about all
the places the files might be (the package manager is usually of no
help in this).


>  
> > Device nodes are a root function; root is the only user that should
> > dictate how device nodes are created; root is the only user that can
> > normally write to / and thereby create udev's rules and rulesets.
> > 
> > In what valid way does access to /usr become something that udev
> > may be required to support?
> 
> As udev is able to run arbitrary scripts, there *might* be some code,
> that requires something from /usr/*. So they want this beast be
> mounted, before udev starts doing it's job.
>  
> > Not arguing with *you* here Michael, just wondering about the
> > validity of the position you quoted
> 
> Understood :)
> 
> Regards,
> Michael
> 
> 



-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Thu, Sep 8, 2011 at 4:03 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 3:37 PM, Michael Mol <mikemol@gmail.com> wrote:
>> On Thu, Sep 8, 2011 at 3:00 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>> Sorry you read that way; I keep trying to explain why I don't find a
>>> separated /usr a necessity and why I don't think an initramfs is such
>>> a big problem.
>>
>> The rhythmic baseline of your explanation is "and you can use an
>> initramfs for that." I haven't seen much in argument for why initramfs
>> isn't a problem apart from your expectation that the *kernel* will
>> eventually require it, and dracut theoretically solve the problem of
>> updating initrd. (Which sounds nice, but it's yet another moving part
>> in the system, and another point of failure.)
>
> No, I think you haven't been reading carefully enough. Again:
>
> 1. In 2011, we need a dynamic /dev tree. I'm not going to argue why.

I'll go ahead and conflate "dynamic /dev tree" with things like
alternate names for network interfaces, and then I can agree with you
there. (I don't put up with this ethN, sitN, tunN BS. Give me an
interface name that says what it's for. I need udev for that, though.)

> 2. udev, successor of devfs, which was successor of the classical /dev
> tree, after years of design and development iterations, solves the
> problem. It's not perfect, but I think that is as close as it could
> be, for the problem it tries to solve, and with the feature set it
> has.

I'll give a pass on this one; I don't know enough about how it
operates to be very informed.

> 3. udev needs either an initramfs, because it needs an early user
> space, or a /usr inside /.

First, it sounds like we're conflating /usr with "user space". User
space is any executing space in the system, but outside the kernel.
/usr is just one of several traditional mountpoints in UNIX-like
systems.

Second, why can't we say, "anything that needs to be operated by udev
needs its binaries in /bin or /sbin"? As I said, I thought that was
the whole *point* of having those distinct from paths under /usr;
anything you needed in order to perform online system recovery would
be in one of those two folders. Anything needed to perform online
system recovery ought to be nearly sufficient to bootstrap the system.
(The only mountpoint exception I can think of is /etc)

Yes, this may mean modifying many more packages, but that's what
having a clean system architecture is about. Unless I'm
misunderstanding sysadmin history, that would seem to be the Correct
solution, and would be the strongest broad solution.

> From this 3 points, I make my conclusion: keep up with the changes, or
> code an alternative (that includes using something like mdev).
>> I thought the primary reason we had
>> /{,s}bin and /usr/{,s}bin was to differentiate between system-vital
>> programs and other programs.
>
> That enters perflectly in my other solution: code an alternative. That
> includes changing (or trying to) udev.

"Changing udev" wasn't on the table earlier.

>
>>> That people don't like the answer doesn't mean is not true.
>>
>> There's also no solid evidence that it's true, either.
>
> Prove me wrong, or find the code that proves me wrong. If it's that
> important to you, do it. My whole point is that it's easier (and in
> the long term more beneficial) to roll with the changes.

What's important to me is clarity of the situation and understanding
the Whys and trying to understand the logic (and illogical aspects) of
the scenario. I'm just trying to be clear on everything here. As I
said, I can deal, it's just a pain. As has been mentioned, there are
other architectures which will have a harder time of it. That's their
problem.

Still, if the Kernel is deciding that udev is the be-all hotplug
manager, though, and udev is hanging edge architectures like MIPS out
to dry, then we'll be lucky if a common, familiar and compatible
environment will continue to be available on MIPS. Maybe they'll fall
back to a static /dev. Maybe they'll come up with something good to
replace udev. Or maybe manufacturers using MIPS will use something
other than Linux, such as QNX, and MIPS will be a much less hackable
platform than it once was.

>> I remember when
>> sysfs came about. Linux Journal's diff -u column described it as a
>> herculean effort accomplishing something the majority of kernel devs
>> thought impossible or not worth the time. I rather like sysfs, but at
>> one time, it was the thing very few people thought was in the possible
>> set of solutions.
>
> I don't see the connection. If someone actually goes and writes the
> code for an alternative to udev (or /usr inside /, or an initramfs),
> then it enters my second alternative.

See my note above about modifying udev not being clearly on the table
earlier. In an environment where a single dev is responsible for
something, foreign patches aren't often welcome.

>> People are beginning to consider alternatives, but you've shot them
>> down without offering improvements, suggesting adjustments or even
>> pointing out specific flaws. As a result, you come across as something
>> akin to a fanboy with your mind already made up, which just seems
>> *bizarre*.
>
> I don't really care if people mind me a fanboi: I know I'm not. I'm
> old enoguh to be a fanold, though.

Heh. Fair enough.

>
> Seriously, maybe I'm not communicating my point clearly. The only
> alternative mentioned (I think: please correct me if I'm wrong) is
> mdev... which I didn't shoot down, I just mentioned that I don't think
> will solve the problem. And if you have followed the development of
> Linux as I have, then you know it is a *really* hard problem, and that
> udev is the result of many man-hours of thinking, designing and
> implementing. That's why I don't give to much hope to a project I have
> never heard about.

I'd be _very_ careful about that mentality. If you've been following
Linux as long as I have (and I honestly think you've been following it
longer; I only started in 1998), you'll remember that Linux was once
an upstart nobody'd heard of. I'm not trying to draw a direct analogy
between Linux and mdev, I'm just trying to illustrate that it's
problematic to ignore small projects because it's assume the big ones
have already thought something through. I've personally tended to find
that big projects and organizations tend to evolve a set of
assumptions (we'll call it groupthink) and don't challenge it often
enough.

> That doesn't mean I'm not mistaken, of course.

Sure. And I'm not trying to say you're necessarily *wrong*, I'm just
trying to break down and poke the assumptions and arguments I'm
hearing.

>>> Sometimes the devs do stupid things; but most of the times they really
>>> think and come to a solution. And the affected users usually just see
>>> how that solution affects them in the short term, instead of trying to
>>> see the big picture and how affects the whole distribution, the
>>> community, and the technological path that Linux is following.
>>
>> For being irritated that people aren't seeing it, you haven't been
>> clarifying it much.
>
> When did I say I was being irritated? I'm only trying to express my POV.

Hm. I may have read too much emotion into the text. My bad.

>> That much is appreciated, but you didn't say *why* you doubt that'll
>> be the solution. Downvotes aren't debate, nor are they discussion.
>> They're expressions of dissatisfaction.
>
> Why should I be dissatisfied? The development of Linux goes exactly in
> the direction *I* want, I don't complain about any of the required
> changes discussed.
>
> Why I doubt it I answered some paragraphs above.

Understood, points taken, some addressed.

>
>>>> As for me, this will be a royal inconvenience, and may require the
>>>> rebuilding of my primary machine. Still, I can deal. It'll mean
>>>> learning how to build initramfs*, how to make sure it contains the
>>>> needed tools, and probably a half-dozen other things I didn't even see
>>>> coming when I set up this box last fall.
>>>
>>> I compile my own kernels (no genkernel for me). I don't use modules
>>> (except for the stupid scsi_wait_scan), and I didn't used an initramfs
>>> until I started using systemd. The arguments for using it convinced
>>> me, and I made the switch in all my machines.
>>>
>>> I don't see why it would require to "rebuild" your primary machine,
>>> but I don't know your configuration. I know that any sane
>>> configuration would only require to install dracut and modify a line
>>> in grub. Maybe a kernel rebuild.
>>
>> My / isn't large enough to hold /usr.
>
> I don't understand. You said you were to try building an initramfs,
> and if that's the case, then you can keep /usr separated.
>
> If you will put /usr in / (and hence rebuild your system), then why
> you said you would learn how to built an initramfs?

I fully expected building an initramfs with a bunch of userland
binaries to be a PITA. If drago is as simple as you describe it, then
perhaps not. At the original point when I indicated I'd need to
rebuild my machine, that was under the assumption I wouldn't be able
to figure out the initramfs.

>>> Apparently. But is not "grossly" unnecessary: udev need it, and udev
>>> solves a kinda complicated problem. Maybe mdev can solve it in a
>>> simpler way.
>>>
>>> But again, I doubt it.
>>
>> One question: Why? Are there specific technical reasons to your
>> doubts, or is it simply confidence that the devs are more likely to
>> have made a good choice than a bad choice?
>
> I answered that already (actually, in that paragraph). But again: udev
> is not trivial, and it solves a (far from) trivial problem. If some
> developers think they can outsmart the kernel devs, please, lets try
> it. Maybe they will.

I was largely looking for an elucidation, which you provided farther up.


Anyway, the discussion has been interesting and, in places,
enlightening. I need to stop getting sucked back into it. :)
-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Canek Peláez Valdés wrote:
> I htink almost everyone understand this. Regards. 

I think you are one of *very* few that understands this.

This reminds me of a old joke.  One in four people have a mental issue.  
Check three friends and if they are OK, you are it.  Again, it is a joke 
but my point is, very few people are liking this.  That alone should say 
a lot.  This is a very few people forcing a change that no one wants.  
You seem to fail to understand that.  If this "new way" of doing things 
causes someones server to be hacked, I would be looking for that dev 
that started this mess.  I don't run some large server but some on here 
do and this is important as it gets.

Personally, if I'm going to have to start running my Gentoo box like a 
binary based distro, I may as well use a binary based distro.  If others 
feel like I do, then Gentoo may start losing users.  I got away from 
Mandrake for reasons such as this.  A init* is just one more thing to 
break.  If you been on this list long enough, you know my record for 
finding things that are really crappy.  One that comes to mind is hal.  
I can assure you I can find other examples.  People complained about hal 
and the dev didn't seem to listen until it really hit the fan.  I think 
the replacement was made by the same dev but maybe after listening a bit 
he found where he could improve things.  I wish the person behind this 
could do the same before he breaks a lot of stuff.  By the way, as Alan 
and others can point out, I never got hal to work on my system.  It was 
nothing fancy either.  At the time it was a Abit NF7 mobo with IDE 
drives and a PS/2 mouse and keyboard.  If a package can't work right on 
something as basic as that, it has little hope of anything fancy for sure.

I'm going back to my garden.  You have fun promoting this mess that is 
being created.  You seem to enjoy it a lot.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 16:48:45 schrieb Canek Peláez Valdés:
> On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer <grimlog@gmx.de> 
wrote:
> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez Valdés:
> >> > In what valid way does access to /usr become something that udev
> >> > may be required to support?
> >> 
> >> It is a matter of what else do you end having in /bin and /lib.
> >> Remember that udev rules can execute arbitrary code. Do all that code
> >> needs to be moved to /bin and /lib also?
> > 
> > Of course. That's what /bin, /sbin and /lib are for.
> > 
> >> I keep telling: it is a difficult problem.
> > 
> > No. Just move or copy the binaries and libs *you* use for *your*
> > udev-scripts to /bin, /sbin and /lib
> 
> I *really* don't think bluetoothd belongs to /sbin. But, hey, that's me.

Well, I don't use bluethoothd, so I don't see a valid reason not to copy 
binaries from /usr/* to /*

> Regards.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 16:23:36 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> It is a matter of what else do you end having in /bin and /lib.
> Remember that udev rules can execute arbitrary code. Do all that code
> needs to be moved to /bin and /lib also?

Why should it not move to /bin and /lib?

Is there a valid case where udev can and should execute arbitrary code
that is completely under the control of a *non-root* user?

We are discussing device nodes. That to me is a root-only function.
Root can access /. The package manager runs as root. Simply put all
code that udev runs into /bin, /sbin and /lib.

Why not restrict the location of said code to the one place where root
is guaranteed to always be able to get to it? It's a reasonable
restriction - the code must be guaranteed to exist on the same
partition as /, where it is guaranteed to be accessible.

> I keep telling: it is a difficult problem.

And I maintain it is the same problem. It is code that is required for
a minimal system to run.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 22:56:07 schrieb Alan McKinnon:
> On Thu, 08 Sep 2011 22:40:07 +0200
> 
> Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > Am Donnerstag, 8. September 2011, 22:05:36 schrieb Alan McKinnon:
> > > On Thu, 08 Sep 2011 19:11:04 +0200
> > > 
> > > Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > > > > Then design the correct solution and implement it. If it's
> > > > > technically sound, it will prevail. I think it's a rather
> > > > > complicated problem with a non trivial solution, but the
> > > > > code is
> > > > > there if you feel like give it a try.
> > > > 
> > > > Where did I write, that I am in the position to write such a
> > > > beast? I only take the freedom to name this a design flaw in
> > > > udev.
> > > > It needs things from userspace, which are not yet available at
> > > > the
> > > > point it requests them. An initramsfs is a workaround for this,
> > > > not a proper fix.
> > > 
> > > If that is the argument from the udev devs you just quoted, then I
> > > do not understand it at all.
> > 
> > It's my understanding, that this is their point.
> > 
> > > Why can there not be a restriction that udev may only run code in
> > > the traditional / space (i.e. it will not attempt to run code in
> > > the /usr or /home spaces)?
> > 
> > Yes. I really wonder, why we have /bin, /sbin and /lib
> 
> The / partition may contain, at a absolute minimum, only that software
> required to boot and start userspace. So you find mount, fsck and sh in
> there.
> <snip>

Thanks, Alan
http://en.wikipedia.org/wiki/Rhetorical_question

;)

Best,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 16:48:45 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer
> <grimlog@gmx.de> wrote:
> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez
> > Valdés:
> >> > In what valid way does access to /usr become something that udev
> >> > may be required to support?
> >>
> >> It is a matter of what else do you end having in /bin and /lib.
> >> Remember that udev rules can execute arbitrary code. Do all that
> >> code needs to be moved to /bin and /lib also?
> >
> > Of course. That's what /bin, /sbin and /lib are for.
> >
> >> I keep telling: it is a difficult problem.
> >
> > No. Just move or copy the binaries and libs *you* use for *your*
> > udev-scripts to /bin, /sbin and /lib
> 
> I *really* don't think bluetoothd belongs to /sbin. But, hey, that's
> me.

Then do what all sane code does when the scripts it uses fails or cannot
be found - throw an error and continue.

Direct the distros to design their systems such that code for essential
functionality is guaranteed to always be available, / is a fine way to
do this.

Everything else and especially functions that the system can tolerate
not having, goes wherever the distro feels like having it.

The distro (or user) can then decide what to do and decide what is and
isn't essential, without having to rig things so that the system's
entire codebase is always on-line. It certainly isn't the udev
maintainer's fault if the distro puts code required to use SATA in a
place that cannot be found. That's a distro bug.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 16:46:25 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> The only simple answer I have is this one: if it is so simple, then it
> would work this way.
> 
> It isn't. If I am wrong (and, of course, that is possible), someone
> will come out and do it, and you guys will be able to keep your
> separated /usr and no initramfs.
> 
> But I really don't think so. But, please, prove me wrong. I would
> certainly be interesting to see another solution.

It already works that way.

I'd be interested in reading the same literature you appear to have
read that convinced you on how this stuff will work in the future.
Nothing I have read has convinced me so far. I conclude you have read
something I haven't. Got a link? 

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan Mackenzie]

Hi, all.

Forgive me butting in at a random place in this rather heated thread,
but ....

On Thu, Sep 08, 2011 at 10:43:29PM +0200, Michael Schreckenbauer wrote:
> Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez Valdés:
> > > In what valid way does access to /usr become something that udev may be
> > > required to support?

> > It is a matter of what else do you end having in /bin and /lib.
> > Remember that udev rules can execute arbitrary code. Do all that code
> > needs to be moved to /bin and /lib also?

> Of course. That's what /bin, /sbin and /lib are for.

> > I keep telling: it is a difficult problem.

> No. Just move or copy the binaries and libs *you* use for *your* udev-scripts 
> to /bin, /sbin and /lib

Would it not be possible to have a minimal /usr tree in the root
partition for udev's use at boot time, and to later mount a more robust /usr
partition over this?  What am I missing here?

> > Regards.

> Regards,
> Michael

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 08 Sep 2011 23:06:43 +0200
Michael Schreckenbauer <grimlog@gmx.de> wrote:

> Thanks, Alan
> http://en.wikipedia.org/wiki/Rhetorical_question
> 
> ;)

Oops :-)

I need to pay more attention to everything that everyone posts. I
didn't know enough (or couldn't remember enough) to judge how many or
few years you'd been hacking Linux, so gave a full literal answer.

In retrospect, that was probably a dumb assumption, the answer was
right there in front of me. Hind-sight truly is an exact science.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 21:29:40 +0000
Alan Mackenzie <acm@muc.de> wrote:

> Hi, all.
> 
> Forgive me butting in at a random place in this rather heated thread,
> but ....
> 
> On Thu, Sep 08, 2011 at 10:43:29PM +0200, Michael Schreckenbauer
> wrote:
> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez
> > Valdés:
> > > > In what valid way does access to /usr become something that
> > > > udev may be required to support?
> 
> > > It is a matter of what else do you end having in /bin and /lib.
> > > Remember that udev rules can execute arbitrary code. Do all that
> > > code needs to be moved to /bin and /lib also?
> 
> > Of course. That's what /bin, /sbin and /lib are for.
> 
> > > I keep telling: it is a difficult problem.
> 
> > No. Just move or copy the binaries and libs *you* use for *your*
> > udev-scripts to /bin, /sbin and /lib
> 
> Would it not be possible to have a minimal /usr tree in the root
> partition for udev's use at boot time, and to later mount a more
> robust /usr partition over this?  What am I missing here?

A big problem will be that the package manager cannot easily maintain
that "phase 1" code as it's under another mount point. Doing so would
require the package manager to bind-mount / somewhere and
copy updated binaries of essential packages there as well as into the
real /usr. Not an insurmountable problem, it just requires changes to
all affected packages, and well within the capabilities of distros.

As a workaround, it's certainly a fine example. But I suspect it will
annoy a lot of users and support people due to this "hidden" code being
on the filesystem. If I were a package maintainer, I know I'd feel a
little annoyed with having to track yet another trait in my packages.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 23:44:41 schrieb Alan McKinnon:
> On Thu, 8 Sep 2011 21:29:40 +0000
> 
> Alan Mackenzie <acm@muc.de> wrote:
> > Would it not be possible to have a minimal /usr tree in the root
> > partition for udev's use at boot time, and to later mount a more
> > robust /usr partition over this?  What am I missing here?
> 
> A big problem will be that the package manager cannot easily maintain
> that "phase 1" code as it's under another mount point. Doing so would
> require the package manager to bind-mount / somewhere and
> copy updated binaries of essential packages there as well as into the
> real /usr. Not an insurmountable problem, it just requires changes to
> all affected packages, and well within the capabilities of distros.

Couldn't whatever mounts /usr bind-mount this "hidden" /usr somewhere (where, 
I think, could be a good question here) before mounting the real one?
Then it would be visible even after the real /usr is mounted.

> As a workaround, it's certainly a fine example. But I suspect it will
> annoy a lot of users and support people due to this "hidden" code being
> on the filesystem. If I were a package maintainer, I know I'd feel a
> little annoyed with having to track yet another trait in my packages.

Agreed.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Donnerstag, 8. September 2011, 23:38:01 schrieb Alan McKinnon:
> On Thu, 08 Sep 2011 23:06:43 +0200
> 
> Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > Thanks, Alan
> > http://en.wikipedia.org/wiki/Rhetorical_question
> > 
> > ;)
> 
> Oops :-)
> 
> I need to pay more attention to everything that everyone posts. I
> didn't know enough (or couldn't remember enough) to judge how many or
> few years you'd been hacking Linux, so gave a full literal answer.

I'm not the most active person here on the list, so no surprise you didn't 
remember :)
For your records: I'm using gentoo since short time after it started, that's 
~9 years now. Before that, I used Suse (you are allowed to beat me). I have a 
FreeBSD-Box as a spare for some years now and a very oldish machine (~1993/4) 
running HP-UX. That aside, I'm still learning things every day.

Best,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Alan Mackenzie]

On Thu, Sep 08, 2011 at 11:44:41PM +0200, Alan McKinnon wrote:

> > Would it not be possible to have a minimal /usr tree in the root
> > partition for udev's use at boot time, and to later mount a more
> > robust /usr partition over this?  What am I missing here?

> A big problem will be that the package manager cannot easily maintain
> that "phase 1" code as it's under another mount point. Doing so would
> require the package manager to bind-mount / somewhere and
> copy updated binaries of essential packages there as well as into the
> real /usr. Not an insurmountable problem, it just requires changes to
> all affected packages, and well within the capabilities of distros.

> As a workaround, it's certainly a fine example. But I suspect it will
> annoy a lot of users and support people due to this "hidden" code being
> on the filesystem. If I were a package maintainer, I know I'd feel a
> little annoyed with having to track yet another trait in my packages.

I'm trying to think of some solution that won't annoy lots of people.
What on earth were the developers thinking when they swept away the
fundamentals assumptions of booting?

> -- 
> Alan McKinnnon
> alan.mckinnon@gmail.com

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1191 bytes --]

Unless I misunderstood this and referenced threads, all this agro is being 
generated because udev devs decided to give primacy not to the linux fs and 
prevailing FHS conventions, but their udev code and what may have been an easy 
workaround for them?

Given that I do not understand the ins and outs of udev, or the way gentoo and 
upstream manage such proposals and ultimately accept changes, why don't gentoo 
devs raise alternative options with the Fedora dev or who ever had this idea 
upstream that udev code effort is more precious than all the workarounds 
(initramfs, repartitioning, etc.) that some of us have to go through?

The alternatives I've read so far that advocate the avoidance of the 
imposition of an initramfs or merging /usr into / for the sake of a udev 
design choice, seem more 'intelligent' to me - in a gentoo principle sort of 
way.

On the other hand, for a binary distro the udev dev approach would of course 
seem less disruptive and therefore our small gentoo user base may need to 
shout really loud to be heard.

Do we get to vote on this?  Can we make a difference other than venting here 
and in the forums?
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 5:11 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 8 Sep 2011 16:48:45 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer
>> <grimlog@gmx.de> wrote:
>> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez
>> > Valdés:
>> >> > In what valid way does access to /usr become something that udev
>> >> > may be required to support?
>> >>
>> >> It is a matter of what else do you end having in /bin and /lib.
>> >> Remember that udev rules can execute arbitrary code. Do all that
>> >> code needs to be moved to /bin and /lib also?
>> >
>> > Of course. That's what /bin, /sbin and /lib are for.
>> >
>> >> I keep telling: it is a difficult problem.
>> >
>> > No. Just move or copy the binaries and libs *you* use for *your*
>> > udev-scripts to /bin, /sbin and /lib
>>
>> I *really* don't think bluetoothd belongs to /sbin. But, hey, that's
>> me.
>
> Then do what all sane code does when the scripts it uses fails or cannot
> be found - throw an error and continue.

This is the problem with people not seeing the big picture: Imagine a
modern system with a bluetooth keyboard. I know, we are geeks, we use
real-men keyboards, not connected with bluetooth. But believe me,
there are people doing that, attaching bluetooth keyboards to their
systems.

And then, the system does a fsck, and it fails and the user needs to
enter her password to go into maintenance mode. But guess what, if the
bluetoothd daemon is not running, she is going to be a very very very
sad user, because "it will throw an error and continue", leaving her
unable to fix it.

We are no longer in 1982: we actually have bluetooth keyboards. It is
a very valid and very actual (probably more in the future) use case.
To solve this case the devs had to choose between putting everything
and the kitchen sink into /sbin or /bin (it is not bounded: anything
can be executed from udev), or to ask to put /usr into / or using an
initramfs.

Sorry if I concur with the later option.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 6:33 PM, Mick <michaelkintzios@gmail.com> wrote:
> Unless I misunderstood this and referenced threads, all this agro is being
> generated because udev devs decided to give primacy not to the linux fs and
> prevailing FHS conventions, but their udev code and what may have been an easy
> workaround for them?
>
> Given that I do not understand the ins and outs of udev, or the way gentoo and
> upstream manage such proposals and ultimately accept changes, why don't gentoo
> devs raise alternative options with the Fedora dev or who ever had this idea
> upstream that udev code effort is more precious than all the workarounds
> (initramfs, repartitioning, etc.) that some of us have to go through?
>
> The alternatives I've read so far that advocate the avoidance of the
> imposition of an initramfs or merging /usr into / for the sake of a udev
> design choice, seem more 'intelligent' to me - in a gentoo principle sort of
> way.
>
> On the other hand, for a binary distro the udev dev approach would of course
> seem less disruptive and therefore our small gentoo user base may need to
> shout really loud to be heard.
>
> Do we get to vote on this?

Not really: you can vote with your feet and use another
distro/operating system. But the choice is theirs.

>  Can we make a difference other than venting here
> and in the forums?

Yes: design and write a different system.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 689 bytes --]

On Thu, 8 Sep 2011 16:42:04 -0400, Canek Peláez Valdés wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Thu, Sep 8, 2011 at 4:25 PM, David W Noon <dwnoon@ntlworld.com>
> wrote:
[snip]
> > They work currently.
> 
> So you want all the new functionality, but without needing to do
> anything.

No.

I just want what currently works to keep working.  Any external change
that is both unnecessary to me and breaks existing code is unwelcome.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 5:03 PM, Dale <rdalek1967@gmail.com> wrote:
> Canek Peláez Valdés wrote:
>>
>> I htink almost everyone understand this. Regards.
>
> I think you are one of *very* few that understands this.
>
> This reminds me of a old joke.  One in four people have a mental issue.
>  Check three friends and if they are OK, you are it.  Again, it is a joke
> but my point is, very few people are liking this.  That alone should say a
> lot.

I know, but Open Source has never been a democracy. It is a
meritocracy. No matter how many get upset by a change, the opinions
that matter are from those writing the code.

> This is a very few people forcing a change that no one wants.

That's a contradiction, isn't it? The "few people" forcing the change
want it, I hope.

> You seem to fail to understand that.

I don't agree with the "few people" and the "no one wants" parts. I
understand that this change is upseting some people, but I don't think
you (nor I) can say for sure if it's even a majority of Gentoo users,
and even if it were, again, Open Source is not a democracy.


> If this "new way" of doing things causes
> someones server to be hacked, I would be looking for that dev that started
> this mess.  I don't run some large server but some on here do and this is
> important as it gets.

If you don't trust this change, you can always change distro/OS (Alan
even recommended it).

> Personally, if I'm going to have to start running my Gentoo box like a
> binary based distro, I may as well use a binary based distro.  If others
> feel like I do, then Gentoo may start losing users.  I got away from
> Mandrake for reasons such as this.

That's your prerrogative. And that's why I'm saying my word in the
list: I'm pretty sure many users in the list (which are not all the
Gentoo users) are not really upset with this change. The other POV has
to be heard.

> A init* is just one more thing to break.
>  If you been on this list long enough, you know my record for finding things
> that are really crappy.  One that comes to mind is hal.  I can assure you I
> can find other examples.  People complained about hal and the dev didn't
> seem to listen until it really hit the fan.  I think the replacement was
> made by the same dev but maybe after listening a bit he found where he could
> improve things.  I wish the person behind this could do the same before he
> breaks a lot of stuff.  By the way, as Alan and others can point out, I
> never got hal to work on my system.  It was nothing fancy either.  At the
> time it was a Abit NF7 mobo with IDE drives and a PS/2 mouse and keyboard.
>  If a package can't work right on something as basic as that, it has little
> hope of anything fancy for sure.

I agree with HAL being a failed experiment: but I think we had to try
it before discarding the idea. Maybe the crap will also hit the fan
with this: I don't know (lost my crystal ball, sorry). But I really
don't believe it, and I have some experience with Linux and Unix and
this kind of stuff. Maybe I'm wrong of course.

> I'm going back to my garden.  You have fun promoting this mess that is being
> created.  You seem to enjoy it a lot.

I'm not promoting anything. Just want to get into the record that some
users don't mind this change, and some of us even welcome it.

The discussion I think has been interesting and civil. I do enjoy it.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 18:39:21 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> On Thu, Sep 8, 2011 at 6:33 PM, Mick <michaelkintzios@gmail.com>
> wrote:
> > Unless I misunderstood this and referenced threads, all this agro
> > is being generated because udev devs decided to give primacy not to
> > the linux fs and prevailing FHS conventions, but their udev code
> > and what may have been an easy workaround for them?
> >
> > Given that I do not understand the ins and outs of udev, or the way
> > gentoo and upstream manage such proposals and ultimately accept
> > changes, why don't gentoo devs raise alternative options with the
> > Fedora dev or who ever had this idea upstream that udev code effort
> > is more precious than all the workarounds (initramfs,
> > repartitioning, etc.) that some of us have to go through?
> >
> > The alternatives I've read so far that advocate the avoidance of the
> > imposition of an initramfs or merging /usr into / for the sake of a
> > udev design choice, seem more 'intelligent' to me - in a gentoo
> > principle sort of way.
> >
> > On the other hand, for a binary distro the udev dev approach would
> > of course seem less disruptive and therefore our small gentoo user
> > base may need to shout really loud to be heard.
> >
> > Do we get to vote on this?
> 
> Not really: you can vote with your feet and use another
> distro/operating system. But the choice is theirs.
> 
> >  Can we make a difference other than venting here
> > and in the forums?
> 
> Yes: design and write a different system.

That's a really poor answer. You are offering two distasteful options
at either end of the spectrum when the real solution is plainly obvious
right in the middle:

Communicate to whichever devs are making the calls, explain the issue
caused by the proposed changes, open and entertain dialogue, let all
voices be heard and let sanity prevail.

You have consistently offered only two realistic options: their way or
the highway. This presumes that the devs involved are impervious to the
concept of dialogue at all, and cannot be contacted or swayed.

You see, none of that is true. There is *always* a third way and it is
almost always the best possible route to follow. 


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri, 09 Sep 2011 00:28:24 +0200
Michael Schreckenbauer <grimlog@gmx.de> wrote:

> Am Donnerstag, 8. September 2011, 23:38:01 schrieb Alan McKinnon:
> > On Thu, 08 Sep 2011 23:06:43 +0200
> > 
> > Michael Schreckenbauer <grimlog@gmx.de> wrote:
> > > Thanks, Alan
> > > http://en.wikipedia.org/wiki/Rhetorical_question
> > > 
> > > ;)
> > 
> > Oops :-)
> > 
> > I need to pay more attention to everything that everyone posts. I
> > didn't know enough (or couldn't remember enough) to judge how many
> > or few years you'd been hacking Linux, so gave a full literal
> > answer.
> 
> I'm not the most active person here on the list, so no surprise you
> didn't remember :)
> For your records: I'm using gentoo since short time after it started,
> that's ~9 years now. Before that, I used Suse (you are allowed to
> beat me). I have a FreeBSD-Box as a spare for some years now and a
> very oldish machine (~1993/4) running HP-UX. That aside, I'm still
> learning things every day.

You are forgiven, completely and without conditions.

I too, once used SuSE.
I too, no longer do so :-)

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 22:31:15 +0000
Alan Mackenzie <acm@muc.de> wrote:

> On Thu, Sep 08, 2011 at 11:44:41PM +0200, Alan McKinnon wrote:
> 
> > > Would it not be possible to have a minimal /usr tree in the root
> > > partition for udev's use at boot time, and to later mount a more
> > > robust /usr partition over this?  What am I missing here?
> 
> > A big problem will be that the package manager cannot easily
> > maintain that "phase 1" code as it's under another mount point.
> > Doing so would require the package manager to bind-mount /
> > somewhere and copy updated binaries of essential packages there as
> > well as into the real /usr. Not an insurmountable problem, it just
> > requires changes to all affected packages, and well within the
> > capabilities of distros.
> 
> > As a workaround, it's certainly a fine example. But I suspect it
> > will annoy a lot of users and support people due to this "hidden"
> > code being on the filesystem. If I were a package maintainer, I
> > know I'd feel a little annoyed with having to track yet another
> > trait in my packages.
> 
> I'm trying to think of some solution that won't annoy lots of people.
> What on earth were the developers thinking when they swept away the
> fundamentals assumptions of booting?

FWIW, I definitely expect at least some distros to consider your idea
seriously (or something similar to it).

It's hacky, no denying it, but no less hacky than the entire concept of
an initramfs itself (think it through for a few minutes, the parallels
are so exact it's amazing).


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 18:36:56 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> On Thu, Sep 8, 2011 at 5:11 PM, Alan McKinnon
> <alan.mckinnon@gmail.com> wrote:
> > On Thu, 8 Sep 2011 16:48:45 -0400
> > Canek Peláez Valdés <caneko@gmail.com> wrote:
> >
> >> On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer
> >> <grimlog@gmx.de> wrote:
> >> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez
> >> > Valdés:
> >> >> > In what valid way does access to /usr become something that
> >> >> > udev may be required to support?
> >> >>
> >> >> It is a matter of what else do you end having in /bin and /lib.
> >> >> Remember that udev rules can execute arbitrary code. Do all that
> >> >> code needs to be moved to /bin and /lib also?
> >> >
> >> > Of course. That's what /bin, /sbin and /lib are for.
> >> >
> >> >> I keep telling: it is a difficult problem.
> >> >
> >> > No. Just move or copy the binaries and libs *you* use for *your*
> >> > udev-scripts to /bin, /sbin and /lib
> >>
> >> I *really* don't think bluetoothd belongs to /sbin. But, hey,
> >> that's me.
> >
> > Then do what all sane code does when the scripts it uses fails or
> > cannot be found - throw an error and continue.
> 
> This is the problem with people not seeing the big picture: Imagine a
> modern system with a bluetooth keyboard. I know, we are geeks, we use
> real-men keyboards, not connected with bluetooth. But believe me,
> there are people doing that, attaching bluetooth keyboards to their
> systems.
> 
> And then, the system does a fsck, and it fails and the user needs to
> enter her password to go into maintenance mode. But guess what, if the
> bluetoothd daemon is not running, she is going to be a very very very
> sad user, because "it will throw an error and continue", leaving her
> unable to fix it.
> 
> We are no longer in 1982: we actually have bluetooth keyboards. It is
> a very valid and very actual (probably more in the future) use case.
> To solve this case the devs had to choose between putting everything
> and the kitchen sink into /sbin or /bin (it is not bounded: anything
> can be executed from udev), or to ask to put /usr into / or using an
> initramfs.

You don't need every possible thing that udev could ever run to be
avialable on /, just the things that are essential. That is quite a
small list subset of the full list of all possible devices:

All HID devices
All console devices
All code to access and read file systems
Everything that can be used in place of a physical keyboard (serial,
console over ethernet)

That looks like it might be a large amount of disk space, but
in fact it isn't. This very mail is being typed on a binary distro
(Ubuntu):

The bluez package is 1.6M.
/lib alone is 331M, I use a fraction of it but it is still there.
/lib/modules contains two kernel versions of 136M each.

Ubuntu has no choice but to ship every imaginable device driver they
support, so / is *already* rather large, and that's just device drivers.

I say leave it up to the distros what to put where. Ubuntu and Fedora
will ship support for bluetooth in the scenario you describe.
JoeBlowLinux might not. If Joe's users are left up the creek, that's
Joe's problem to deal with, but all the infinite variety of possible
Joes still have the same freedom they always had - to fix or break
their shipped stuff at will.

Incidentally, your bluetooth keyboard example is a tad disingenuous.
That fictitious user is going to have an almighty problem at grub time
selecting what system to boot. A bluetooth keyboard is the ONLY
possible keyboard input device? What kind of general purpose computing
hardware were you picturing?


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 7:00 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 8 Sep 2011 18:39:21 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> On Thu, Sep 8, 2011 at 6:33 PM, Mick <michaelkintzios@gmail.com>
>> wrote:
>> > Unless I misunderstood this and referenced threads, all this agro
>> > is being generated because udev devs decided to give primacy not to
>> > the linux fs and prevailing FHS conventions, but their udev code
>> > and what may have been an easy workaround for them?
>> >
>> > Given that I do not understand the ins and outs of udev, or the way
>> > gentoo and upstream manage such proposals and ultimately accept
>> > changes, why don't gentoo devs raise alternative options with the
>> > Fedora dev or who ever had this idea upstream that udev code effort
>> > is more precious than all the workarounds (initramfs,
>> > repartitioning, etc.) that some of us have to go through?
>> >
>> > The alternatives I've read so far that advocate the avoidance of the
>> > imposition of an initramfs or merging /usr into / for the sake of a
>> > udev design choice, seem more 'intelligent' to me - in a gentoo
>> > principle sort of way.
>> >
>> > On the other hand, for a binary distro the udev dev approach would
>> > of course seem less disruptive and therefore our small gentoo user
>> > base may need to shout really loud to be heard.
>> >
>> > Do we get to vote on this?
>>
>> Not really: you can vote with your feet and use another
>> distro/operating system. But the choice is theirs.
>>
>> >  Can we make a difference other than venting here
>> > and in the forums?
>>
>> Yes: design and write a different system.
>
> That's a really poor answer. You are offering two distasteful options
> at either end of the spectrum when the real solution is plainly obvious
> right in the middle:
>
> Communicate to whichever devs are making the calls, explain the issue
> caused by the proposed changes, open and entertain dialogue, let all
> voices be heard and let sanity prevail.
>
> You have consistently offered only two realistic options: their way or
> the highway. This presumes that the devs involved are impervious to the
> concept of dialogue at all, and cannot be contacted or swayed.
>
> You see, none of that is true. There is *always* a third way and it is
> almost always the best possible route to follow.

In the case of Gentoo, the dialog is having place in the dev list, at
this very moment. In the case of Fedora (and, I think, OpenSuse), the
dialog is actually over. The Gentoo devs are just going with the flow.

(This is how I see things, I could have some facts wrong).

It is not an arbitrary decision, and it is not from one developer
(this kind of things never are). The dialog happend (or is happening)
among those who construct the stack or the distributions. We have a
say, of course (we always do), but I don't really think that it should
be that important. I really, truly believe that the decision is (and
should be) in the hands of the people actually writing the code.

I think this is how Linux rose to be what it is today, and how it will
keep going on strong. Sometimes mistakes will be made, and some users
will be burned by them.

I (personally, IMHO, etc., etc.) don't think this is one of those
times. And that is way I'm expressing myself in this thread.

That is all. I know what I say a lot of people don't like, but I think
it should be said, clear and loud.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 999 bytes --]

On Thu, 8 Sep 2011 22:45:21 +0200, Alan McKinnon wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Thu, 8 Sep 2011 21:25:31 +0100
> David W Noon <dwnoon@ntlworld.com> wrote:
> 
> > > Well, then you already know what to do.  
> > 
> > Indeed I do.
> 
> If your databases are not ASE or Oracle, then migrating to FreeBSD is
> a fine option.

They are mostly PostgreSQL, with a few tiddlers under SQLite.

I have been contemplating a FreeBSD install for a couple of years now.
I will have a spare box in a week or so, if I sort out the hardware
issues, so I could test FreeBSD on that.  I am still contemplating
whether to use the Gentoo FreeBSD and build everything from source, or
download an install DVD and see what the "vanilla" option offers.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 7:23 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> On Thu, 8 Sep 2011 18:36:56 -0400
> Canek Peláez Valdés <caneko@gmail.com> wrote:
>
>> On Thu, Sep 8, 2011 at 5:11 PM, Alan McKinnon
>> <alan.mckinnon@gmail.com> wrote:
>> > On Thu, 8 Sep 2011 16:48:45 -0400
>> > Canek Peláez Valdés <caneko@gmail.com> wrote:
>> >
>> >> On Thu, Sep 8, 2011 at 4:43 PM, Michael Schreckenbauer
>> >> <grimlog@gmx.de> wrote:
>> >> > Am Donnerstag, 8. September 2011, 16:23:36 schrieb Canek Peláez
>> >> > Valdés:
>> >> >> > In what valid way does access to /usr become something that
>> >> >> > udev may be required to support?
>> >> >>
>> >> >> It is a matter of what else do you end having in /bin and /lib.
>> >> >> Remember that udev rules can execute arbitrary code. Do all that
>> >> >> code needs to be moved to /bin and /lib also?
>> >> >
>> >> > Of course. That's what /bin, /sbin and /lib are for.
>> >> >
>> >> >> I keep telling: it is a difficult problem.
>> >> >
>> >> > No. Just move or copy the binaries and libs *you* use for *your*
>> >> > udev-scripts to /bin, /sbin and /lib
>> >>
>> >> I *really* don't think bluetoothd belongs to /sbin. But, hey,
>> >> that's me.
>> >
>> > Then do what all sane code does when the scripts it uses fails or
>> > cannot be found - throw an error and continue.
>>
>> This is the problem with people not seeing the big picture: Imagine a
>> modern system with a bluetooth keyboard. I know, we are geeks, we use
>> real-men keyboards, not connected with bluetooth. But believe me,
>> there are people doing that, attaching bluetooth keyboards to their
>> systems.
>>
>> And then, the system does a fsck, and it fails and the user needs to
>> enter her password to go into maintenance mode. But guess what, if the
>> bluetoothd daemon is not running, she is going to be a very very very
>> sad user, because "it will throw an error and continue", leaving her
>> unable to fix it.
>>
>> We are no longer in 1982: we actually have bluetooth keyboards. It is
>> a very valid and very actual (probably more in the future) use case.
>> To solve this case the devs had to choose between putting everything
>> and the kitchen sink into /sbin or /bin (it is not bounded: anything
>> can be executed from udev), or to ask to put /usr into / or using an
>> initramfs.
>
> You don't need every possible thing that udev could ever run to be
> avialable on /, just the things that are essential. That is quite a
> small list subset of the full list of all possible devices:
>
> All HID devices
> All console devices
> All code to access and read file systems
> Everything that can be used in place of a physical keyboard (serial,
> console over ethernet)
>
> That looks like it might be a large amount of disk space, but
> in fact it isn't. This very mail is being typed on a binary distro
> (Ubuntu):
>
> The bluez package is 1.6M.
> /lib alone is 331M, I use a fraction of it but it is still there.
> /lib/modules contains two kernel versions of 136M each.

Again, it is not bounded. Today is bluez, tomorrow we don't know.
That's the point of udev, really.

> Ubuntu has no choice but to ship every imaginable device driver they
> support, so / is *already* rather large, and that's just device drivers.
>
> I say leave it up to the distros what to put where. Ubuntu and Fedora
> will ship support for bluetooth in the scenario you describe.
> JoeBlowLinux might not. If Joe's users are left up the creek, that's
> Joe's problem to deal with, but all the infinite variety of possible
> Joes still have the same freedom they always had - to fix or break
> their shipped stuff at will.

Oh, so we came to same conclusion, just from different sides of the
spectrum. Nice.

It just seems weird to me that most distros (including Gentoo,
apparentely) seem to be going for the initramfs way. Maybe there is
some technical merit to that option that I can't see. O, wait, I do
see it.

> Incidentally, your bluetooth keyboard example is a tad disingenuous.
> That fictitious user is going to have an almighty problem at grub time
> selecting what system to boot. A bluetooth keyboard is the ONLY
> possible keyboard input device? What kind of general purpose computing
> hardware were you picturing?

A PC preloaded with only Linux?

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Canek Peláez Valdés wrote:
> On Thu, Sep 8, 2011 at 5:03 PM, Dale<rdalek1967@gmail.com>  wrote:
>> Canek Peláez Valdés wrote:
>>> I htink almost everyone understand this. Regards.
>> I think you are one of *very* few that understands this.
>>
>> This reminds me of a old joke.  One in four people have a mental issue.
>>   Check three friends and if they are OK, you are it.  Again, it is a joke
>> but my point is, very few people are liking this.  That alone should say a
>> lot.
> I know, but Open Source has never been a democracy. It is a
> meritocracy. No matter how many get upset by a change, the opinions
> that matter are from those writing the code.
>
>>   This is a very few people forcing a change that no one wants.
> That's a contradiction, isn't it? The "few people" forcing the change
> want it, I hope.

It's not.  So far, one dev made the decision to do this and a few have 
agreed.  There are lots of people, as noted in this thread, that 
disagree.  Some of those people have been using Linux for a very long 
time.  I don't know how long you have been using Linux but I'm pushing 
ten years myself.  I suspect that Neil and Alan, and maybe others, have 
been using Linux a LOT longer than that.  Maybe more than both of us put 
together.  When I see a post by Alan or Neil, I read it carefully.  
There are Linux idiots in this world but they are not one of them.  On 
some subjects, I fall into the ignorance category.  I don't claim to 
know it all but some things I do know well.


>
>>   You seem to fail to understand that.
> I don't agree with the "few people" and the "no one wants" parts. I
> understand that this change is upseting some people, but I don't think
> you (nor I) can say for sure if it's even a majority of Gentoo users,
> and even if it were, again, Open Source is not a democracy.

I don't think you quite understood my wording.  I think you mentioned 
English is not your first language so this happens a lot.  I hope the 
above helped.


>
>>   If this "new way" of doing things causes
>> someones server to be hacked, I would be looking for that dev that started
>> this mess.  I don't run some large server but some on here do and this is
>> important as it gets.
> If you don't trust this change, you can always change distro/OS (Alan
> even recommended it).

I'm putting that in the consideration bin.  It could be a possibility.  
I like to stick with things but if I'm going to be told to bend over and 
take it, they could at least bring some Vaseline.  It seems some of the 
things I left Mandrake over are coming to Gentoo.  Almost makes me 
wonder if I should have left.  Well, I have had some good years so far.  
Plus I like helping folks on this list too.

>
>> Personally, if I'm going to have to start running my Gentoo box like a
>> binary based distro, I may as well use a binary based distro.  If others
>> feel like I do, then Gentoo may start losing users.  I got away from
>> Mandrake for reasons such as this.
> That's your prerrogative. And that's why I'm saying my word in the
> list: I'm pretty sure many users in the list (which are not all the
> Gentoo users) are not really upset with this change. The other POV has
> to be heard.
>
>> A init* is just one more thing to break.
>>   If you been on this list long enough, you know my record for finding things
>> that are really crappy.  One that comes to mind is hal.  I can assure you I
>> can find other examples.  People complained about hal and the dev didn't
>> seem to listen until it really hit the fan.  I think the replacement was
>> made by the same dev but maybe after listening a bit he found where he could
>> improve things.  I wish the person behind this could do the same before he
>> breaks a lot of stuff.  By the way, as Alan and others can point out, I
>> never got hal to work on my system.  It was nothing fancy either.  At the
>> time it was a Abit NF7 mobo with IDE drives and a PS/2 mouse and keyboard.
>>   If a package can't work right on something as basic as that, it has little
>> hope of anything fancy for sure.
> I agree with HAL being a failed experiment: but I think we had to try
> it before discarding the idea. Maybe the crap will also hit the fan
> with this: I don't know (lost my crystal ball, sorry). But I really
> don't believe it, and I have some experience with Linux and Unix and
> this kind of stuff. Maybe I'm wrong of course.

It was more than failed.  It was miserable.
>
>> I'm going back to my garden.  You have fun promoting this mess that is being
>> created.  You seem to enjoy it a lot.
> I'm not promoting anything. Just want to get into the record that some
> users don't mind this change, and some of us even welcome it.
>
> The discussion I think has been interesting and civil. I do enjoy it.
>
> Regards.

Some don't but my point still stands.  If this becomes a security issue 
for someone with a rig that can't adapt, I'd have some really choice 
words for a dev if I had a server that I depended on.  I think Alan has 
already pointed out some of those exceptions.  Alan is more able to 
explain that than me for sure.  He sits at some of those servers and I 
don't.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-08 20:40, Canek Peláez Valdés wrote:

> The world is never black or white. Under several definitions,
> *nothing* has stopped working.

Hm. Using the same setup as I always have, nothing will stop working
with the new "paradigm"? Great, then what are we arguing about? If
things will work the way they always have without me having to _add_
something completely unnecessary just to be able to boot then I'm all
for it.

>You just need to do extra stuff to keep
> it working.

So functionality _is_ removed then? By definition, I would say that's
something that has stopped working. No, I don't want to install more
crap or change my partitioning that doesn't add any value to _me_...

But as you say, if the rest of the _binary distro_ world decides to do
it there's not much we can do about it than arguing. I don't expect
Gentoo devs to support/maintain something that upstream doesn't support.
I'm just really unhappy about it.

> (Of course, again, is not black and white: the kernel devs, I trust
> almost withouth doubt. GNOME devs I trust less. An unknown hacker with
> a new project I trust almost nothing.)

Hm... I think the kernel devs, Gentoo maintainers are, for the most part
extremely competent and I trust their judgements for the most part (even
though I agree with Linus that the kernel has become bloated).
To put it bluntly: GNOME devs seems like blathering idiots to me,
spewing out crap, trying to mimic MS desktop/functionality as best as
they can which usually ends up a very poor (and buggy) version of it.
Unfortunately it's hard to escape their influence completely but I'm
trying my best (and Gentoo is one of the very few ways of having some
control of this). Concerning new projects I judge it by functionality,
dependencies, "buggyness"...

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-08 20:41, Canek Peláez Valdés wrote:

> Yeah, but not the second part that you conveniently omitted: the
> freedom to modify the code.

What does the freedom to modify the code has to do with this discussion?
I thought we were discussing removed functionality (or changed as you
like to see it)...

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 5193 bytes --]

On Friday 09 Sep 2011 00:26:33 Canek Peláez Valdés wrote:
> On Thu, Sep 8, 2011 at 7:00 PM, Alan McKinnon <alan.mckinnon@gmail.com> 
wrote:
> > On Thu, 8 Sep 2011 18:39:21 -0400
> > 
> > Canek Peláez Valdés <caneko@gmail.com> wrote:
> >> On Thu, Sep 8, 2011 at 6:33 PM, Mick <michaelkintzios@gmail.com>
> >> 
> >> wrote:
> >> > Unless I misunderstood this and referenced threads, all this agro
> >> > is being generated because udev devs decided to give primacy not to
> >> > the linux fs and prevailing FHS conventions, but their udev code
> >> > and what may have been an easy workaround for them?
> >> > 
> >> > Given that I do not understand the ins and outs of udev, or the way
> >> > gentoo and upstream manage such proposals and ultimately accept
> >> > changes, why don't gentoo devs raise alternative options with the
> >> > Fedora dev or who ever had this idea upstream that udev code effort
> >> > is more precious than all the workarounds (initramfs,
> >> > repartitioning, etc.) that some of us have to go through?
> >> > 
> >> > The alternatives I've read so far that advocate the avoidance of the
> >> > imposition of an initramfs or merging /usr into / for the sake of a
> >> > udev design choice, seem more 'intelligent' to me - in a gentoo
> >> > principle sort of way.
> >> > 
> >> > On the other hand, for a binary distro the udev dev approach would
> >> > of course seem less disruptive and therefore our small gentoo user
> >> > base may need to shout really loud to be heard.
> >> > 
> >> > Do we get to vote on this?
> >> 
> >> Not really: you can vote with your feet and use another
> >> distro/operating system. But the choice is theirs.
> >> 
> >> >  Can we make a difference other than venting here
> >> > and in the forums?
> >> 
> >> Yes: design and write a different system.
> > 
> > That's a really poor answer. You are offering two distasteful options
> > at either end of the spectrum when the real solution is plainly obvious
> > right in the middle:
> > 
> > Communicate to whichever devs are making the calls, explain the issue
> > caused by the proposed changes, open and entertain dialogue, let all
> > voices be heard and let sanity prevail.
> > 
> > You have consistently offered only two realistic options: their way or
> > the highway. This presumes that the devs involved are impervious to the
> > concept of dialogue at all, and cannot be contacted or swayed.
> > 
> > You see, none of that is true. There is *always* a third way and it is
> > almost always the best possible route to follow.
> 
> In the case of Gentoo, the dialog is having place in the dev list, at
> this very moment. In the case of Fedora (and, I think, OpenSuse), the
> dialog is actually over. The Gentoo devs are just going with the flow.
> 
> (This is how I see things, I could have some facts wrong).

Aha!  This is I think where it went wrong.

The Gentoo devs should *not* have gone with the flow.  Giving the udev code 
primacy over the conventional FHS way, rather than spending some more time to 
sort out the genuine cause of the problem (udev) is something that in this 
case affects the Gentoo principle of doing it the 'Gentoo way' - more than 
binary distros who are already using initramfs.

So this is a Gentoo user/use case argument more than upstream devs may care to 
examine.


> It is not an arbitrary decision, and it is not from one developer
> (this kind of things never are). The dialog happend (or is happening)
> among those who construct the stack or the distributions. We have a
> say, of course (we always do), but I don't really think that it should
> be that important. I really, truly believe that the decision is (and
> should be) in the hands of the people actually writing the code.

You have made this point clear enough, but the way this has been decided 
clearly cuts across the choice of freedom that Gentoo users had until now.

People are getting upset and using an initramfs, repartitioning, or becoming 
Linux developers overnight to write their own udev code is not a particularly 
attractive option for most Gentoo users.


> I think this is how Linux rose to be what it is today, and how it will
> keep going on strong. Sometimes mistakes will be made, and some users
> will be burned by them.
> 
> I (personally, IMHO, etc., etc.) don't think this is one of those
> times. And that is way I'm expressing myself in this thread.

Fair enough.  It is evident that there are quite a few of us that disagree 
with your view on this matter.

I think that in this case some devs followed what is convenient or expedient, 
rather than choosing a more purist/elegant approach that fixes what's broken 
(udev) without affecting adversely the wider ecosystem.


> That is all. I know what I say a lot of people don't like, but I think
> it should be said, clear and loud.

I believe that you have repeated your position enough times that we all get 
it.  Your position though advocates a design solution which cuts across the 
Gentoo way of doing things.  This makes Gentoo less valuable to some of us.

-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Mick wrote:
> On Friday 09 Sep 2011 00:26:33 Canek Peláez Valdés wrote:
> In the case of Gentoo, the dialog is having place in the dev list, at
> this very moment. In the case of Fedora (and, I think, OpenSuse), the
> dialog is actually over. The Gentoo devs are just going with the flow.
>
> (This is how I see things, I could have some facts wrong).
> Aha!  This is I think where it went wrong.
>
> The Gentoo devs should *not* have gone with the flow.  Giving the udev code
> primacy over the conventional FHS way, rather than spending some more time to
> sort out the genuine cause of the problem (udev) is something that in this
> case affects the Gentoo principle of doing it the 'Gentoo way' - more than
> binary distros who are already using initramfs.
>
> So this is a Gentoo user/use case argument more than upstream devs may care to
> examine.

This is my understanding and what I can recall reading on -dev.  
Basically someone, a dev, at Fedora decided to do it this way.  That was 
where the discussion ended.  I read somewhere that the dev in question 
won't even reply, maybe not even read, complaints to what he/she is 
doing.   Basically, he/she is saying what has been said in this thread.  
It is my way or the highway.  Along with the loss of options, having 
this big a change with the person inflicting it not listening is 
disturbing.  What's next, /home will be need on / as well?   I really 
think having one or even just a few people that can cause a change like 
this needs to be revisited.

I would like to know what Linus thinks about this mess.  Does he know?  
Is he thinking this is OK?

Dale

:-)  :-)

[gentoo-user] Re: /dev/sda* missing at boot

[Nicolas Sebrecht]

The 09/09/11, Michael Schreckenbauer wrote:
> Am Donnerstag, 8. September 2011, 23:44:41 schrieb Alan McKinnon:
> > On Thu, 8 Sep 2011 21:29:40 +0000
> > 
> > Alan Mackenzie <acm@muc.de> wrote:
> > > Would it not be possible to have a minimal /usr tree in the root
> > > partition for udev's use at boot time, and to later mount a more
> > > robust /usr partition over this?  What am I missing here?
> > 
> > A big problem will be that the package manager cannot easily maintain
> > that "phase 1" code as it's under another mount point. Doing so would
> > require the package manager to bind-mount / somewhere and
> > copy updated binaries of essential packages there as well as into the
> > real /usr. Not an insurmountable problem, it just requires changes to
> > all affected packages, and well within the capabilities of distros.
> 
> Couldn't whatever mounts /usr bind-mount this "hidden" /usr somewhere (where, 
> I think, could be a good question here) before mounting the real one?
> Then it would be visible even after the real /usr is mounted.

So, you're asking if it's smart to use yet another path (hidden once
finished to properly boot) to store what is currently stored in /bin and
/sbin...

Remember: the only reason why /bin and /sbin exist is to have tools
available during boot time to mount /usr.

-- 
Nicolas Sebrecht

Re: [gentoo-user] /dev/sda* missing at boot

[Paul Colquhoun]

On Thu, 8 Sep 2011 04:03:53 PM Canek Peláez Valdés wrote:

> No, I think you haven't been reading carefully enough. Again:
> 
> 1. In 2011, we need a dynamic /dev tree. I'm not going to argue why.
> 2. udev, successor of devfs, which was successor of the classical /dev
> tree, after years of design and development iterations, solves the
> problem. It's not perfect, but I think that is as close as it could
> be, for the problem it tries to solve, and with the feature set it
> has.
> 3. udev needs either an initramfs, because it needs an early user
> space, or a /usr inside /.
> 
> From this 3 points, I make my conclusion: keep up with the changes, or
> code an alternative (that includes using something like mdev).


From my point of view, as an old Solaris admin, point 3) is the problem.

If what-ever-it-is is needed during boot, it should be in /sbin or /bin or 
/lib

If it is curently in /usr/* then it is in the wrong place, and that package 
should be modified.

Later in the thread you mentioned a bluetooth keyboard. This obviously 
requires either a driver module, or a bluetooth server process, or similar, 
which  belong in /lib{32,64}/modules or /sbin

Having udev able to execute arbitrary code during boot looks like yet another 
large security hole opening up. At least keep the code it can execute tied 
down to the directories that were set up for this purpose.


-- 
Reverend Paul Colquhoun, ULC.    http://andor.dropbear.id.au/~paulcol
 Before you criticize someone, you should walk a mile in their shoes.
Then, when you do, you'll be a mile away, and you'll have their shoes.

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Paul Colquhoun wrote:
> >From my point of view, as an old Solaris admin, point 3) is the 
> problem. If what-ever-it-is is needed during boot, it should be in 
> /sbin or /bin or /lib If it is curently in /usr/* then it is in the 
> wrong place, and that package should be modified. Later in the thread 
> you mentioned a bluetooth keyboard. This obviously requires either a 
> driver module, or a bluetooth server process, or similar, which belong 
> in /lib{32,64}/modules or /sbin Having udev able to execute arbitrary 
> code during boot looks like yet another large security hole opening 
> up. At least keep the code it can execute tied down to the directories 
> that were set up for this purpose. 

Picking a random post to reply to.

I been using Linux for a while.  Let me see if I understand this 
correctly.  As I understand it, when a system boots it needs /bin, 
/sbin, /lib*, and /etc and nothing else other than /boot for grub to 
load the kernel.  Those directories are for booting the system and for 
"system" operations.  That is my understanding of how it has been since 
further back than I care to explore.  Things that are used after a 
system boots, such as things in the default runlevel or KDE, goes into 
/usr somewhere.  This is the reason that /usr and /var can be on 
separate partitions.  I have always understood that /usr and /var can be 
put on separate partitions for security reasons or to put some larger 
partitions on separate drives.  If I recall correctly, websites files 
are under /var.  Those can get pretty large quick I would guess.

So, now someone has decided to change this and it seems a few think this 
is nothing users should worry about.  I don't run a large server or 
anything but this still worries me.  I don't like the fact that the 
changes I had planned will now require me to also install one more thing 
to break.  My system is simple and I like to keep it that way.  The 
fanciest thing I have is a camera and a printer that I use once in a 
blue moon.  I want to put /usr on a spare partition because it is 
growing fairly quickly with the KDE4 updates and others too.  Now, it 
looks like I have to do a whole redo of everything.  Something that was 
simple just got complicated.

My choices are:

1: move from Gentoo to something else.  I'm seriously considering this 
one.  If I can learn Gentoo, I can learn any distro!  LFS may be 
excluded tho.
2: Stick with Gentoo and hope this is corrected like hal was dealt with.
2b:  Go with LVM for everything and have a init* to boot.
2c:  Move /usr and use init* with no LVM.
2d:  Just redo my whole system with a larger / partition.

I liked my original plan better.

1:  Go to boot runlevel.
2:  Mount what will be new /usr partition to some mount point.
3:  Copy /usr to the new partition
4:  rm the old /usr data.
5:  Mount the new /usr partition and add it to fstab
6:  Switch back to default runlevel and life goes on.

Can I slap whoever started this?  The more I think on this, the worse it 
sounds.  I can't even imagine someone who runs some large server.  Any 
hair left?  lol

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Thursday, September 08, 2011 06:55:32 PM Canek Peláez Valdés wrote:
> On Thu, Sep 8, 2011 at 5:03 PM, Dale <rdalek1967@gmail.com> wrote:
> > Canek Peláez Valdés wrote:
> >> I htink almost everyone understand this. Regards.
> > 
> > I think you are one of *very* few that understands this.
> > 
> > This reminds me of a old joke.  One in four people have a mental issue.
> >  Check three friends and if they are OK, you are it.  Again, it is a
> > joke
> > but my point is, very few people are liking this.  That alone should say
> > a lot.
> 
> I know, but Open Source has never been a democracy. It is a
> meritocracy. No matter how many get upset by a change, the opinions
> that matter are from those writing the code.

I don't agree. There are people with opinions that matter even though they 
don't write the code. There are plenty of Open Source projects where the 
opinions and comments from users also matter. And if those users actually put 
time and effort into the documentation/support side they get listened to more 
often.

> > This is a very few people forcing a change that no one wants.
> 
> That's a contradiction, isn't it? The "few people" forcing the change
> want it, I hope.

Ok, lets do it by numbers.
People forcing it: 5 (maybe? not that many more)
People liking it (including the above 5): 10 (maybe?)
total number of users: 1,000,000 (pulled out of my head)

Percentage of users liking it of all the users: 10 / 1,000,000 = 0.0001 %.
That's a very low number that in most cases would be rounded to 0. Eg. noone.

> > You seem to fail to understand that.
> 
> I don't agree with the "few people" and the "no one wants" parts. I
> understand that this change is upseting some people, but I don't think
> you (nor I) can say for sure if it's even a majority of Gentoo users,

I think the majority of Gentoo users will happily continue the way they have 
been working with their systems. Then, when this change gets forced upon them, 
they will all start complaining loudly because all their systems no longer 
boot.

> and even if it were, again, Open Source is not a democracy.

Actually, it is. People tend to vote with their feet (ok, downloads) and if 
they don't like something, they walk away.

> > Personally, if I'm going to have to start running my Gentoo box like a
> > binary based distro, I may as well use a binary based distro.  If others
> > feel like I do, then Gentoo may start losing users.  I got away from
> > Mandrake for reasons such as this.
> 
> That's your prerrogative. And that's why I'm saying my word in the
> list: I'm pretty sure many users in the list (which are not all the
> Gentoo users) are not really upset with this change. The other POV has
> to be heard.

I haven't gone through the whole thread, but it seems to me there are several 
people against this change and only one who is for.

I kept quiet as my arguments were already being raised and I dislike "+1" 
postings. But in this case, I feel an exception is needed.

> > I'm going back to my garden.  You have fun promoting this mess that is
> > being created.  You seem to enjoy it a lot.
> 
> I'm not promoting anything. Just want to get into the record that some
> users don't mind this change, and some of us even welcome it.

Why would anyone welcome a change where an initramfs (or whatever it's called 
these days) is necessary just to boot your system?
This also needs to keep getting updated whenever a needed piece of software is 
updated. I tend to update the software more often then the kernel. Now, I'll 
have to rebuild my kernel more regularly. Even though, from my point of view, 
nothing will have changed.

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri, 09 Sep 2011 03:53:26 -0500
Dale <rdalek1967@gmail.com> wrote:

> Paul Colquhoun wrote:
> > >From my point of view, as an old Solaris admin, point 3) is the 
> > problem. If what-ever-it-is is needed during boot, it should be in 
> > /sbin or /bin or /lib If it is curently in /usr/* then it is in the 
> > wrong place, and that package should be modified. Later in the
> > thread you mentioned a bluetooth keyboard. This obviously requires
> > either a driver module, or a bluetooth server process, or similar,
> > which belong in /lib{32,64}/modules or /sbin Having udev able to
> > execute arbitrary code during boot looks like yet another large
> > security hole opening up. At least keep the code it can execute
> > tied down to the directories that were set up for this purpose. 
> 
> Picking a random post to reply to.
> 
> I been using Linux for a while.  Let me see if I understand this 
> correctly.  As I understand it, when a system boots it needs /bin, 
> /sbin, /lib*, and /etc and nothing else other than /boot for grub to 
> load the kernel.  Those directories are for booting the system and
> for "system" operations.  That is my understanding of how it has been
> since further back than I care to explore.  

Correct.

/ is often set up with only the minimal packages needed to guarantee
that single user mode will work correctly if the only thing mounted
is / itself.

> Things that are used
> after a system boots, such as things in the default runlevel or KDE,
> goes into /usr somewhere.  This is the reason that /usr and /var can
> be on separate partitions.  I have always understood that /usr
> and /var can be put on separate partitions for security reasons or to
> put some larger partitions on separate drives.  If I recall
> correctly, websites files are under /var.  Those can get pretty large
> quick I would guess.

Correct again.

/var is for "variable data", usually persistent data like log files,
databases, web sites, caches. It is writeable by root and system data
goes there (as opposed to user data).

> 
> So, now someone has decided to change this and it seems a few think
> this is nothing users should worry about.  I don't run a large server
> or anything but this still worries me.  I don't like the fact that
> the changes I had planned will now require me to also install one
> more thing to break.  My system is simple and I like to keep it that
> way.  The fanciest thing I have is a camera and a printer that I use
> once in a blue moon.  I want to put /usr on a spare partition because
> it is growing fairly quickly with the KDE4 updates and others too.
> Now, it looks like I have to do a whole redo of everything.
> Something that was simple just got complicated.

The truth is that with these changes your system will continue to work
just fine. Just like my laptops work just fine (I have one big
partition with another for /home on laptops).

My laptops don't need a separate /usr, but my servers do.

So it really looks like someone is forcing a change that makes udev's
life easier and potentially wreaks everything else in doing so.

> My choices are:
> 
> 1: move from Gentoo to something else.  I'm seriously considering
> this one.  If I can learn Gentoo, I can learn any distro!  LFS may be 
> excluded tho.

It's not a Gentoo change, it's a udev change. So you'll be stuck with
this new stuff regardless of which distro you go with.

> 2: Stick with Gentoo and hope this is corrected like hal was dealt
> with. 2b:  Go with LVM for everything and have a init* to boot.
> 2c:  Move /usr and use init* with no LVM.
> 2d:  Just redo my whole system with a larger / partition.

2e. Migrate to Windows where you too can have one partition on / and
have it fully supported by Microsoft!! OK, my sarcasm is showing.

> I liked my original plan better.
> 
> 1:  Go to boot runlevel.
> 2:  Mount what will be new /usr partition to some mount point.
> 3:  Copy /usr to the new partition
> 4:  rm the old /usr data.
> 5:  Mount the new /usr partition and add it to fstab
> 6:  Switch back to default runlevel and life goes on.
> 
> Can I slap whoever started this?  The more I think on this, the worse
> it sounds.  I can't even imagine someone who runs some large server.
> Any hair left?  lol

I'm lucky, I can vote with my feet. Out of 140, I have two servers that
*require* Linux. One runs Sybase ASE, the other runs Oracle. Everything
else works like a bomb on FreeBSD. 

kthankxbyeudev, thanksfornotplayingnicely

Not everyone else is so fortunate though.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Thursday, September 08, 2011 03:01:10 PM Canek Peláez Valdés wrote:
> On Thu, Sep 8, 2011 at 1:47 PM, Michael Mol <mikemol@gmail.com> wrote:
> > On Thu, Sep 8, 2011 at 1:35 PM, pk <peterk2@coolmail.se> wrote:
> >> On 2011-09-08 16:51, Canek Peláez Valdés wrote:
> >>> But the freedom is still there. The freedom to either keep your
> >>> system
> >>> as it is (don't upgrade)
> >> 
> >>            ^^^^^^^^^^^^^
> >> You do realise that this is quite valid for Windows (and all other
> >> OS's
> >> in existence)? At least so far...
> > 
> > Don't get *me* started. My _day job_ is C++/MFC on Windows. _Please_
> > upgrade, you'll make my life much easier.
> > 
> > Outdated operating systems make baby coder cry.
> 
> I already mentioned that you update security flaws.

Update the security flaws is all nice and well, but won't hold up for very 
long.
Security updates for older versions will stop within a short period. And not 
sufficient information will be available to keep patching the software 
individually.

> And again, that's only if you resist the change.

This sounds like "We are borg, resistance is futile...." :)

--
Joost

Re: [gentoo-user] Re: /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Freitag, 9. September 2011, 10:06:21 schrieb Nicolas Sebrecht:
> The 09/09/11, Michael Schreckenbauer wrote:
> > Am Donnerstag, 8. September 2011, 23:44:41 schrieb Alan McKinnon:
> > > On Thu, 8 Sep 2011 21:29:40 +0000
> > > 
> > > Alan Mackenzie <acm@muc.de> wrote:
> > > > Would it not be possible to have a minimal /usr tree in the root
> > > > partition for udev's use at boot time, and to later mount a more
> > > > robust /usr partition over this?  What am I missing here?
> > > 
> > > A big problem will be that the package manager cannot easily
> > > maintain
> > > that "phase 1" code as it's under another mount point. Doing so
> > > would
> > > require the package manager to bind-mount / somewhere and
> > > copy updated binaries of essential packages there as well as into
> > > the
> > > real /usr. Not an insurmountable problem, it just requires changes
> > > to
> > > all affected packages, and well within the capabilities of distros.
> > 
> > Couldn't whatever mounts /usr bind-mount this "hidden" /usr somewhere
> > (where, I think, could be a good question here) before mounting the
> > real one? Then it would be visible even after the real /usr is mounted.
> 
> So, you're asking if it's smart to use yet another path (hidden once
> finished to properly boot) to store what is currently stored in /bin and
> /sbin...
> Remember: the only reason why /bin and /sbin exist is to have tools
> available during boot time to mount /usr.

The question arose, when Canek mentioned bluetoothd, that udev seems to need 
in some cases. If bluetoothd doesn't quite fit to /bin or /sbin (I tend to 
agree here), but is needed before /usr is mounted, then it has to be put 
*somewhere*. I don't say, that this is the way to go. Only searching for 
alternatives to a forced initramfs.

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Thu, 8 Sep 2011 19:34:56 -0400
Canek Peláez Valdés <caneko@gmail.com> wrote:

> > You don't need every possible thing that udev could ever run to be
> > avialable on /, just the things that are essential. That is quite a
> > small list subset of the full list of all possible devices:
> >
> > All HID devices
> > All console devices
> > All code to access and read file systems
> > Everything that can be used in place of a physical keyboard (serial,
> > console over ethernet)
> >
> > That looks like it might be a large amount of disk space, but
> > in fact it isn't. This very mail is being typed on a binary distro
> > (Ubuntu):
> >
> > The bluez package is 1.6M.
> > /lib alone is 331M, I use a fraction of it but it is still there.
> > /lib/modules contains two kernel versions of 136M each.  
> 
> Again, it is not bounded. Today is bluez, tomorrow we don't know.
> That's the point of udev, really.

You're still not getting it.

Just because it appears convenient to make udev unbounded does not mean
that all possible code on the machine has to be accessible to udev.
Or that udev will potentially run any arbitrary code you might have.
Or put another way, udev might be able to run anything, like say
lauching KDE, but the simple truth is that it won't in any reasonable
scenario. Therefore you do not need to support or entertain that
possibility. The truth is that a very small portion of the total code
on the machine needs to be accessible to udev and all of it (including
all foreseeable code) fits into a traditional / quite nicely.

There is no upper limit on the size of /, you simply make it as large as
you need and put everything supported in there.

Once again, and this is very important, the only things that are
absolutely required to be in / is all the code that must run
before /usr is mounted. That list of things is very small, and if the
user or the distro happens to cock it up, then the user or distro must
fix it. 

Why is this apparently so hard to understand? The solution seems
blindingly obvious:

Any code launched by udev must be available on the same partition as /.
However the system is rigged, that one condition must be satisfied. And
consider who is setting this up:

- root, who presumably knows what they are doing
- distro devs, who also know what they are doing

Or are the udev devs seriously contemplating allowing udev hooks so
that any arbitrary user can launch any arbitrary code that might
arbitrarily be anywhere?

I still maintain this "fix" is for a problem that does not exist.


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Dale writes:

Wow, what a big thread. While I also do not really like udev
requiring /usr at boot time, I also understand that there are some
arguments pro doing so.
But then, I wonder what the big deal is. If an initramfs is now required
for people using a separate /usr, then let's all use an initramfs, if we
can't change how udev is going. It's annoying, we may feel it is wrong,
but to me it seems that for most of us it is not a really big problem.
What I fear much more is when good old grub is no longer supported and I
have to use grub2, which I tried to understand, but failed.

> My choices are:
> 
> 1: move from Gentoo to something else.  I'm seriously considering this 
> one.  If I can learn Gentoo, I can learn any distro!  LFS may be 
> excluded tho.

So, because you want to avoid to change your Gentoo installation to use
an initramfs, you switch to another distribution, which most likely uses
an initramfs anayway?

> 2: Stick with Gentoo and hope this is corrected like hal was dealt with.
> 2b:  Go with LVM for everything and have a init* to boot.

LVM is great and I suggest everyone using it, but it's not necessary here.

> 2c:  Move /usr and use init* with no LVM.

If you can extend you root partition, yes, just copy /usr there, and all
will be fine.

> 2d:  Just redo my whole system with a larger / partition.

Which would be a lot of work.

Personally I do not care much about this, as I already am using an
initramfs :) That's because all my partitions are encrypted LVM volumes.
Except for /boot, which is on on USB stick.

When I switched to using an initramfs, it was not very complicated. I
simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my 
/usr/src/linux/.config and does not change the kernel options. Then comes genkernel --install --lvm -luks all, and I have kernel and initramfs
in /boot. I manually add them to my grub.conf. emerge @module-rebuild,
and I'm done. I guess for most of us this would work. I don't know what
Michael has to do in order to keep nvidia-drivers instead of nouveau, but
I assume some howto or new item will come up to solve this. Whenever
Gentoo had us to do major changes, there was a good explanation of what
to do, and it worked fine. Migration to openrc was more complicated I
think. And hey, I was satisfied with the way it's been before.

> I liked my original plan better.
> 
> 1:  Go to boot runlevel.
> 2:  Mount what will be new /usr partition to some mount point.
> 3:  Copy /usr to the new partition
> 4:  rm the old /usr data.
> 5:  Mount the new /usr partition and add it to fstab
> 6:  Switch back to default runlevel and life goes on.

I don't get this one. Why do you want to copy an existing /usr partition
to another one?

> Can I slap whoever started this?  The more I think on this, the worse
> it sounds.  I can't even imagine someone who runs some large server.
> Any hair left?  lol

Yes, I also feel sorry for guys like Alan. But for us desktop users
I think's it's not such a big deal.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

David W Noon writes:

> The more I think about this merge of / and /usr, the dumber I think the
> idea is.  As I wrote in an earlier message on this list, the initramfs
> will be many times larger than the kernel itself.  Indeed, my /boot
> partition is only 32 MiB, and that will be too small to contain all the
> extra libraries and programs to run the initramfs script.

Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5 M
for the initramfs.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alex Schuster wrote:
> David W Noon writes:
>
>> The more I think about this merge of / and /usr, the dumber I think the
>> idea is.  As I wrote in an earlier message on this list, the initramfs
>> will be many times larger than the kernel itself.  Indeed, my /boot
>> partition is only 32 MiB, and that will be too small to contain all the
>> extra libraries and programs to run the initramfs script.
> Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5 M
> for the initramfs.
>
> 	Wonko
>

Well, that may not be the case for everyone else.

root@fireball / # du -shc /boot/
84M     /boot/
84M     total
root@fireball / #

Of course, while I am redoing my partitions, I guess I can make /boot 
bigger as well.  Heck, may have to change something else before to 
long.  I'm sure someone will find some side corner case where something 
might happen and decide to fix what isn't broke.  Yep, sounds about 
right to me.  It's not the first time for this sort of thing to happen.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 4072 bytes --]

On Friday 09 Sep 2011 12:35:47 Alex Schuster wrote:
> Dale writes:
> 
> Wow, what a big thread. While I also do not really like udev
> requiring /usr at boot time, I also understand that there are some
> arguments pro doing so.
> But then, I wonder what the big deal is. If an initramfs is now required
> for people using a separate /usr, then let's all use an initramfs, if we
> can't change how udev is going. It's annoying, we may feel it is wrong,
> but to me it seems that for most of us it is not a really big problem.
> What I fear much more is when good old grub is no longer supported and I
> have to use grub2, which I tried to understand, but failed.
> 
> > My choices are:
> > 
> > 1: move from Gentoo to something else.  I'm seriously considering this
> > one.  If I can learn Gentoo, I can learn any distro!  LFS may be
> > excluded tho.
> 
> So, because you want to avoid to change your Gentoo installation to use
> an initramfs, you switch to another distribution, which most likely uses
> an initramfs anayway?
> 
> > 2: Stick with Gentoo and hope this is corrected like hal was dealt with.
> > 2b:  Go with LVM for everything and have a init* to boot.
> 
> LVM is great and I suggest everyone using it, but it's not necessary here.
> 
> > 2c:  Move /usr and use init* with no LVM.
> 
> If you can extend you root partition, yes, just copy /usr there, and all
> will be fine.
> 
> > 2d:  Just redo my whole system with a larger / partition.
> 
> Which would be a lot of work.
> 
> Personally I do not care much about this, as I already am using an
> initramfs :) That's because all my partitions are encrypted LVM volumes.
> Except for /boot, which is on on USB stick.
> 
> When I switched to using an initramfs, it was not very complicated. I
> simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my
> /usr/src/linux/.config and does not change the kernel options. Then comes
> genkernel --install --lvm -luks all, and I have kernel and initramfs in
> /boot. I manually add them to my grub.conf. emerge @module-rebuild, and
> I'm done. I guess for most of us this would work. I don't know what
> Michael has to do in order to keep nvidia-drivers instead of nouveau, but
> I assume some howto or new item will come up to solve this. Whenever
> Gentoo had us to do major changes, there was a good explanation of what to
> do, and it worked fine. Migration to openrc was more complicated I think.
> And hey, I was satisfied with the way it's been before.
> 
> > I liked my original plan better.
> > 
> > 1:  Go to boot runlevel.
> > 2:  Mount what will be new /usr partition to some mount point.
> > 3:  Copy /usr to the new partition
> > 4:  rm the old /usr data.
> > 5:  Mount the new /usr partition and add it to fstab
> > 6:  Switch back to default runlevel and life goes on.
> 
> I don't get this one. Why do you want to copy an existing /usr partition
> to another one?
> 
> > Can I slap whoever started this?  The more I think on this, the worse
> > it sounds.  I can't even imagine someone who runs some large server.
> > Any hair left?  lol
> 
> Yes, I also feel sorry for guys like Alan. But for us desktop users
> I think's it's not such a big deal.

It's not a catastrophically big deal, but it is an imposed workaround that 
goes against the freedom of choice that we gentoo-ers have enjoyed hitherto.

It also seems counter-intuitive that udev devs' convenience should take 
primacy over the FHS convention and the prevailing minimal booting process.

It will only affect one out of three boxen of mine and I could surely fix 
that, but I am against restricting unquestioningly what I can do with gentoo, 
just because a udev coder didn't think it through enough to come up with a 
smarter solution; and then the Gentoo devs did not put up a fight in 
representing their user base.

It's a point of principle and on this basis I'd like to object to it, not for 
a poxy little box which I can reconfigure one day if I must.
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Dale writes:

> Alex Schuster wrote:
> > David W Noon writes:
> >
> >> The more I think about this merge of / and /usr, the dumber I think
> >> the idea is.  As I wrote in an earlier message on this list, the
> >> initramfs will be many times larger than the kernel itself.  Indeed,
> >> my /boot partition is only 32 MiB, and that will be too small to
> >> contain all the extra libraries and programs to run the initramfs
> >> script.
> > Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5
> > M for the initramfs.

> Well, that may not be the case for everyone else.

Sure, but how much bigger are your kernels actually?

> root@fireball / # du -shc /boot/
> 84M     /boot/
> 84M     total
> root@fireball / #

I get 82M, but I have ten kernels in there. What stuff do you have
in /boot?

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Canek Peláez Valdés]

On Thu, Sep 8, 2011 at 10:55 PM, Dale <rdalek1967@gmail.com> wrote:
> Canek Peláez Valdés wrote:
>>
>> On Thu, Sep 8, 2011 at 5:03 PM, Dale<rdalek1967@gmail.com>  wrote:
>>>
>>> Canek Peláez Valdés wrote:
>>>>
>>>> I htink almost everyone understand this. Regards.
>>>
>>> I think you are one of *very* few that understands this.
>>>
>>> This reminds me of a old joke.  One in four people have a mental issue.
>>>  Check three friends and if they are OK, you are it.  Again, it is a joke
>>> but my point is, very few people are liking this.  That alone should say
>>> a
>>> lot.
>>
>> I know, but Open Source has never been a democracy. It is a
>> meritocracy. No matter how many get upset by a change, the opinions
>> that matter are from those writing the code.
>>
>>>  This is a very few people forcing a change that no one wants.
>>
>> That's a contradiction, isn't it? The "few people" forcing the change
>> want it, I hope.
>
> It's not.  So far, one dev made the decision to do this and a few have
> agreed.  There are lots of people, as noted in this thread, that disagree.
>  Some of those people have been using Linux for a very long time.  I don't
> know how long you have been using Linux but I'm pushing ten years myself.  I
> suspect that Neil and Alan, and maybe others, have been using Linux a LOT
> longer than that.  Maybe more than both of us put together.  When I see a
> post by Alan or Neil, I read it carefully.  There are Linux idiots in this
> world but they are not one of them.  On some subjects, I fall into the
> ignorance category.  I don't claim to know it all but some things I do know
> well.

The "contradiction" part was a joke. A bad one, it seems.

I started using Linux in 1996, when I started college (Computer
Science, if you must know). I used RedHat, then Mandrake, then Gentoo,
around 2003. After college I worked in several companies, doing mostly
programming, but also a lot of system administration. I have worked
with Solaris, HP-UX, SCO, and a tiny little bit of AIX, but the bulk
of my curriculum is in Linux.

In 2005 I got bored of being like Dilbert, and went back to school to
get my masters in 2008 (Computer Science, again), and after getting
back to work less than six months, I returned to Academia to get my
PhD (Computer Science, what the hell), which I hope to get next year.

That is not going to happen if instead of finishing writing my papers,
I keep posting to threads in gentoo-user.

I have some experience with Linux and Unix. I have followed the
development of Linux, GNOME and everything in beetween in the stack
like some people follow soap operas or football games. I think I kinda
know what I'm talking about.

But of course, I could be wrong in this issue. I just don't think so.

I said my points and listened to very different and interesting ones.
From my POV (and I say this with all the respect possible), I see a
lot of people afraid of change or too worried about their pet
configurations, but not a really Earth-shattering technical strong
point that makes me believe this change is "unnecessary",
"irrational", or "lazy". It is incovenient? Sure, but in the long run
I think it would make Linux better.

This I haven't said, I think: I care about Linux, and basically Linux
only. I want it to be on all my electronics, from my cell phone to my
refrigerator and of course in my desktop. That is already happening,
and the direction it is heading.

But to do that, Linux cannot be a "classical Unix". It needs to be so
much more. It needs to do thinks *DIFFERENTLY*. So, even if Linux will
be always able to do anything any other Unix could do, it will do it
in a fundamentally different way. So if you care for a Unix boxen that
only does Unix-boxen things, in the classical, 1970-way, then probably
Linux is not the best option for you.

And for sure *I* don't want progress stopped only so Linux is able to
do the things already does in the same way, with the only argument
being "my script/setup/partition works now, why should I changed it?"

Change happens.

I appreciate the discussion, and I think it was enlightening and
entertaining, but I will not participate anymore. I need to get my PhD
one of this days.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-09 13:35, Alex Schuster wrote:

> When I switched to using an initramfs, it was not very complicated. I
> simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my 
> /usr/src/linux/.config and does not change the kernel options. Then comes genkernel --install --lvm -luks all, and I have kernel and initramfs

And for those that like to do without genkernel? Again, adding another
layer for things to go wrong.

> I don't get this one. Why do you want to copy an existing /usr partition
> to another one?

He said he wishes to move his /usr to a spare partition (the part about
KDE4)... I assume his /usr currently resides on / (or maybe a smaller
partition that he cannot easily expand).

> Yes, I also feel sorry for guys like Alan. But for us desktop users
> I think's it's not such a big deal.

I'm a desktop and a (personal server) user and I think it's quite a big
deal. I want simplicity; adding layers increases complexity. I think
it's the same for Dale and most other people objecting to this. To me
it's a very big deal (this is a deal breaker, or close to it). I've been
using Linux continously since around 1998 (well, I did my first install
on my amiga 4000 in 1995 using 9 floppy disks, don't remember the
distro) and I've been using (not much administration though) Solaris,
AIX and HP-UX since around that time as well (at school & at work). It
seems some developers are hell bent on inventing Windows all over again
(this goes not only for udev but also for Gnome and their supporting
libraries)...

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

pk writes:

> On 2011-09-09 13:35, Alex Schuster wrote:
> 
> > When I switched to using an initramfs, it was not very complicated. I
> > simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my 
> > /usr/src/linux/.config and does not change the kernel options. Then
> > comes genkernel --install --lvm -luks all, and I have kernel and
> > initramfs
> 
> And for those that like to do without genkernel? Again, adding another
> layer for things to go wrong.

I just wanted to say that it _can_ be easy. When I installed my system, I
knew I would need an initramfs, and while I knew what that is, I did
not know how to set it up. But then I thought about trying genkernel,
which I never used before, and it worked very well. I did not have to
care about the details. Instead of make bzImage modules modules_install
and copying the results to /boot, I use the genkernel command, and that's
it.

> > I don't get this one. Why do you want to copy an existing /usr
> > partition to another one?
> 
> He said he wishes to move his /usr to a spare partition (the part about
> KDE4)... I assume his /usr currently resides on / (or maybe a smaller
> partition that he cannot easily expand).

Right, I somehow overlooked this, thanks for pointing that out. Dale, if
you want to avoid the initramfs, what about moving large stuff
like /usr/src to another location and symlinking it? That's a hack, but a
small one compared to what an initramfs is :)

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Fri, Sep 9, 2011 at 1:04 PM, Alex Schuster <wonko@wonkology.org> wrote:
> pk writes:
>
>> On 2011-09-09 13:35, Alex Schuster wrote:
>>
>> > When I switched to using an initramfs, it was not very complicated. I
>> > simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my
>> > /usr/src/linux/.config and does not change the kernel options. Then
>> > comes genkernel --install --lvm -luks all, and I have kernel and
>> > initramfs
>>
>> And for those that like to do without genkernel? Again, adding another
>> layer for things to go wrong.
>
> I just wanted to say that it _can_ be easy. When I installed my system, I
> knew I would need an initramfs, and while I knew what that is, I did
> not know how to set it up. But then I thought about trying genkernel,
> which I never used before, and it worked very well. I did not have to
> care about the details. Instead of make bzImage modules modules_install
> and copying the results to /boot, I use the genkernel command, and that's
> it.
>
>> > I don't get this one. Why do you want to copy an existing /usr
>> > partition to another one?
>>
>> He said he wishes to move his /usr to a spare partition (the part about
>> KDE4)... I assume his /usr currently resides on / (or maybe a smaller
>> partition that he cannot easily expand).
>
> Right, I somehow overlooked this, thanks for pointing that out. Dale, if
> you want to avoid the initramfs, what about moving large stuff
> like /usr/src to another location and symlinking it? That's a hack, but a
> small one compared to what an initramfs is :)

Why symlink? Why not make it its own mountpoint? :)


-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-09 10:53, Dale wrote:

> Can I slap whoever started this?  The more I think on this, the worse it

Yes Dale, you have my permission! And while you're at it, slap him from
me too! ;-)

It _may_ be this guy that's responsible for this crap:
http://linuxplumbersconf.org/ocw/users/58

Also:
http://comments.gmane.org/gmane.linux.hotplug.devel/16994

PS. If things go "tits up" you may want to have a look at FreeBSD (or
some other BSD). I'm quite sure they wouldn't put up with crap like
this... I know I will investigate my options at least.

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Schreckenbauer]

Am Freitag, 9. September 2011, 19:24:06 schrieb pk:
> On 2011-09-09 10:53, Dale wrote:
> > Can I slap whoever started this?  The more I think on this, the worse it
> 
> Yes Dale, you have my permission! And while you're at it, slap him from
> me too! ;-)
> 
> It _may_ be this guy that's responsible for this crap:
> http://linuxplumbersconf.org/ocw/users/58
> 
> Also:
> http://comments.gmane.org/gmane.linux.hotplug.devel/16994

OMG!
What a mess. udev treats all exit-codes except 0 the same.
That's so bad, I have no words for it.
Definitely not a developer I trust to do things the right way.

> PS. If things go "tits up" you may want to have a look at FreeBSD (or
> some other BSD). I'm quite sure they wouldn't put up with crap like
> this... I know I will investigate my options at least.

Agree, FreeBSD is really a fine OS.

> Best regards
> Peter K

Thanks for the links,
regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 1999 bytes --]

On Fri, 9 Sep 2011 13:41:07 +0200, Alex Schuster wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> David W Noon writes:
> 
> > The more I think about this merge of / and /usr, the dumber I think
> > the idea is.  As I wrote in an earlier message on this list, the
> > initramfs will be many times larger than the kernel itself.
> > Indeed, my /boot partition is only 32 MiB, and that will be too
> > small to contain all the extra libraries and programs to run the
> > initramfs script.
> 
> Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5
> M for the initramfs.

My kernels are even smaller than yours: around 1.8MiB; and I have no
initramfs at all -- currently.

The problem is the initramfs will bloat out significantly once large
run-time libraries are required for early housekeeping, such as fsck
for various types of filesystem.  In particular, the old e2fsck.static
program has been dropped from e2fspprogs (about 3 years ago) and we now
have the following:

dwn@karnak ~ % ldd /sbin/e2fsck
	linux-gate.so.1 =>  (0xb7832000)
	libext2fs.so.2 => /lib/libext2fs.so.2 (0xb77c1000)
	libcom_err.so.2 => /lib/libcom_err.so.2 (0xb77bd000)
	libblkid.so.1 => /lib/libblkid.so.1 (0xb7798000)
	libuuid.so.1 => /lib/libuuid.so.1 (0xb7793000)
	libe2p.so.2 => /lib/libe2p.so.2 (0xb778b000)
	libc.so.6 => /lib/libc.so.6 (0xb7604000)
	libpthread.so.0 => /lib/libpthread.so.0 (0xb75ea000)
	/lib/ld-linux.so.2 (0xb7833000)

As you can see, the fsck utility for ext2/3/4 filesystems requires
glibc and libpthread, as well as its smaller custom libraries.  Putting
all the run-time libraries into the initramfs will make it both large
and a maintenance chore.

What kind of libraries do you have inside your initramfs?
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

David W Noon writes:

> On Fri, 9 Sep 2011 13:41:07 +0200, Alex Schuster wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
> 
> > David W Noon writes:
> > 
> > > The more I think about this merge of / and /usr, the dumber I think
> > > the idea is.  As I wrote in an earlier message on this list, the
> > > initramfs will be many times larger than the kernel itself.
> > > Indeed, my /boot partition is only 32 MiB, and that will be too
> > > small to contain all the extra libraries and programs to run the
> > > initramfs script.
> > 
> > Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5
> > M for the initramfs.
> 
> My kernels are even smaller than yours: around 1.8MiB; and I have no
> initramfs at all -- currently.
> 
> The problem is the initramfs will bloat out significantly once large
> run-time libraries are required for early housekeeping, such as fsck
> for various types of filesystem.  In particular, the old e2fsck.static
> program has been dropped from e2fspprogs (about 3 years ago) and we now
> have the following:
> 
> dwn@karnak ~ % ldd /sbin/e2fsck
> 	linux-gate.so.1 =>  (0xb7832000)
> 	libext2fs.so.2 => /lib/libext2fs.so.2 (0xb77c1000)
> 	libcom_err.so.2 => /lib/libcom_err.so.2 (0xb77bd000)
> 	libblkid.so.1 => /lib/libblkid.so.1 (0xb7798000)
> 	libuuid.so.1 => /lib/libuuid.so.1 (0xb7793000)
> 	libe2p.so.2 => /lib/libe2p.so.2 (0xb778b000)
> 	libc.so.6 => /lib/libc.so.6 (0xb7604000)
> 	libpthread.so.0 => /lib/libpthread.so.0 (0xb75ea000)
> 	/lib/ld-linux.so.2 (0xb7833000)
> 
> As you can see, the fsck utility for ext2/3/4 filesystems requires
> glibc and libpthread, as well as its smaller custom libraries.  Putting
> all the run-time libraries into the initramfs will make it both large
> and a maintenance chore.

Okay, it seems I very much underestimated the problems. In my case, I
only need the initramfs in order to scan for logical volumes and to open
the luks-encrypted root partition. Other partitions are mounted _after_
the initramfs was left.
With the UDEV change, /usr needs to be mounted from _inside_ the
initramfs. So you're right, much more stuff is being needed. The
above libraries and the e2fsck binary total to 2.3 M here. The initramfs
is gzipped, so we have 1 M. Still not _that_ much, but I don't know what
else might be needed.
And something must put it into the initramfs... I assume genkernel will
get this feature? Surely the Gentoo devs won't expect us users to do this
all by ourselves?

> What kind of libraries do you have inside your initramfs?

I have no idea... but I can have a look. Ah - none at all. /lib contains a
directory with all sorts of keymaps, an empty luks directory, and some 56
kernel modules.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

pk wrote:
> On 2011-09-09 13:35, Alex Schuster wrote:
>
>> When I switched to using an initramfs, it was not very complicated. I
>> simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my
>> /usr/src/linux/.config and does not change the kernel options. Then comes genkernel --install --lvm -luks all, and I have kernel and initramfs
> And for those that like to do without genkernel? Again, adding another
> layer for things to go wrong.

I tried genkernel.  All I got was a kernel that wouldn't boot.  Heck, it 
barely even started to boot.  The kernel wouldn't even finish loading.  
After several tries, I put genkernel in the trash.  It worked a LOT 
better there for me.  It was out of sight and mind.  ;-)


>> I don't get this one. Why do you want to copy an existing /usr partition
>> to another one?
> He said he wishes to move his /usr to a spare partition (the part about
> KDE4)... I assume his /usr currently resides on / (or maybe a smaller
> partition that he cannot easily expand).
>

You hit it, for some reason I put /usr on the root partition without 
thinking.  This is where I am now:

rootfs                19534436  10693048   8841388  55% /

Over half full.  When I have a critical partition get over 60%, I start 
looking for expansion.  Moving /usr was my plan but someone stole that 
from me I guess.  Now I got to figure out what I want to do next.


>> Yes, I also feel sorry for guys like Alan. But for us desktop users
>> I think's it's not such a big deal.
> I'm a desktop and a (personal server) user and I think it's quite a big
> deal. I want simplicity; adding layers increases complexity. I think
> it's the same for Dale and most other people objecting to this. To me
> it's a very big deal (this is a deal breaker, or close to it). I've been
> using Linux continously since around 1998 (well, I did my first install
> on my amiga 4000 in 1995 using 9 floppy disks, don't remember the
> distro) and I've been using (not much administration though) Solaris,
> AIX and HP-UX since around that time as well (at school&  at work). It
> seems some developers are hell bent on inventing Windows all over again
> (this goes not only for udev but also for Gnome and their supporting
> libraries)...
>
> Best regards
>
> Peter K
>

I'm a desktop user to and I'm not liking this one bit.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alex Schuster wrote:
> Right, I somehow overlooked this, thanks for pointing that out. Dale, 
> if you want to avoid the initramfs, what about moving large stuff like 
> /usr/src to another location and symlinking it? That's a hack, but a 
> small one compared to what an initramfs is :) Wonko 

I already have portage on a separate partition and I clean out my kernel 
sources once I get a really good stable kernel.  I actually cleaned out 
/boot and /usr/src last night.  The kernel I am running now has let me 
have weeks of uptimes so I guess it is stable, at least everything works 
and no random crashes or anything.  Well, kpat locks up on me sometimes 
but that is nothing new.  As soon as I see a way to win, it locks up 
tight.  Pisses me off when it does that.  lol

I got a spare drive in here.  I may just do a install there and use it 
to play with init crap and maybe LVM.  Sort of see what I want to do.  
Still thinking about just picking something else tho.  I'm just not 
seeing the need to continue if options are going to be removed.  
Eventually Gentoo will be like Mandrake where you just install and say a 
prayer it works.  The things that are going away are the reasons I chose 
Gentoo to begin with.

< sighs >

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Michael Schreckenbauer wrote:
> Am Freitag, 9. September 2011, 19:24:06 schrieb pk:
>> On 2011-09-09 10:53, Dale wrote:
>>> Can I slap whoever started this?  The more I think on this, the worse it
>> Yes Dale, you have my permission! And while you're at it, slap him from
>> me too! ;-)
>>
>> It _may_ be this guy that's responsible for this crap:
>> http://linuxplumbersconf.org/ocw/users/58
>>
>> Also:
>> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
> OMG!
> What a mess. udev treats all exit-codes except 0 the same.
> That's so bad, I have no words for it.
> Definitely not a developer I trust to do things the right way.
>
>> PS. If things go "tits up" you may want to have a look at FreeBSD (or
>> some other BSD). I'm quite sure they wouldn't put up with crap like
>> this... I know I will investigate my options at least.
> Agree, FreeBSD is really a fine OS.
>
>> Best regards
>> Peter K
> Thanks for the links,
> regards,
> Michael
>

I know one thing, BSD is secure as heck.  I installed it once on a old 
rig and typed the password in wrong during setup.  I never could get 
into that thing again.  I had to start over.  lol  That is why I chose 
Linux in general.  I want something that is secure enough that I don't 
have to worry about some script kiddie messing with me.

BSD is one option I will be looking into if I move from Gentoo.  After 
all, they are fairly close maybe even a step up.  Especially now.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alex Schuster wrote:
> Dale writes:
>
>> Alex Schuster wrote:
>>> David W Noon writes:
>>>
>>>> The more I think about this merge of / and /usr, the dumber I think
>>>> the idea is.  As I wrote in an earlier message on this list, the
>>>> initramfs will be many times larger than the kernel itself.  Indeed,
>>>> my /boot partition is only 32 MiB, and that will be too small to
>>>> contain all the extra libraries and programs to run the initramfs
>>>> script.
>>> Here, I only need 2.2 M for the kernel, 1.7 M for System.map, and 3.5
>>> M for the initramfs.
>> Well, that may not be the case for everyone else.
> Sure, but how much bigger are your kernels actually?
>
>> root@fireball / # du -shc /boot/
>> 84M     /boot/
>> 84M     total
>> root@fireball / #
> I get 82M, but I have ten kernels in there. What stuff do you have
> in /boot?
>
> 	Wonko
>
>

Well, I *had* several old kernels in there.  I save stable kernels as I 
upgrade until I have a really good one then I remove the older ones.  I 
always keep at least two kernels tho.  If one fails, I got a fall back.  
I have had to use those fall backs before so I won't be changing that 
policy here any time soon.

I think I had about a dozen or so in there until my cleaning out party 
last night.  I also save back up configs to just in case a kernel goes 
bad or I need to go back.  I version my kernels too.  Long story.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Fri, Sep 9, 2011 at 9:15 PM, Dale <rdalek1967@gmail.com> wrote:
> Michael Schreckenbauer wrote:
>>
>> Am Freitag, 9. September 2011, 19:24:06 schrieb pk:
>>>
>>> On 2011-09-09 10:53, Dale wrote:
>>>>
>>>> Can I slap whoever started this?  The more I think on this, the worse it
>>>
>>> Yes Dale, you have my permission! And while you're at it, slap him from
>>> me too! ;-)
>>>
>>> It _may_ be this guy that's responsible for this crap:
>>> http://linuxplumbersconf.org/ocw/users/58
>>>
>>> Also:
>>> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
>>
>> OMG!
>> What a mess. udev treats all exit-codes except 0 the same.
>> That's so bad, I have no words for it.
>> Definitely not a developer I trust to do things the right way.
>>
>>> PS. If things go "tits up" you may want to have a look at FreeBSD (or
>>> some other BSD). I'm quite sure they wouldn't put up with crap like
>>> this... I know I will investigate my options at least.
>>
>> Agree, FreeBSD is really a fine OS.
>>
>>> Best regards
>>> Peter K
>>
>> Thanks for the links,
>> regards,
>> Michael
>>
>
> I know one thing, BSD is secure as heck.  I installed it once on a old rig
> and typed the password in wrong during setup.  I never could get into that
> thing again.  I had to start over.  lol  That is why I chose Linux in
> general.  I want something that is secure enough that I don't have to worry
> about some script kiddie messing with me.
>
> BSD is one option I will be looking into if I move from Gentoo.  After all,
> they are fairly close maybe even a step up.  Especially now.

Doesn't Gentoo have a BSD target? The problem here is with udev, which
doesn't apply to BSD, AFAIK. Gentoo/BSD might be a good direction to
go.

Also, where does FreeBSD's kernel stand, with respect to device drivers?

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> I'm lucky, I can vote with my feet. Out of 140, I have two servers 
> that *require* Linux. One runs Sybase ASE, the other runs Oracle. 
> Everything else works like a bomb on FreeBSD. kthankxbyeudev, 
> thanksfornotplayingnicely Not everyone else is so fortunate though. 

I guess I understood more than I thought then.  Shocking.  I understand 
that but the udev guru doesn't.  ;-)

I may go the BSD route too if I leave Gentoo.  So, my feet works too.  I 
wonder if I would even be missed here?  :/

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Fri, Sep 9, 2011 at 9:25 PM, Dale <rdalek1967@gmail.com> wrote:
> I may go the BSD route too if I leave Gentoo.  So, my feet works too.  I
> wonder if I would even be missed here?  :/

I'd hate it if you left. In the short time I've been on this list,
your usage habits and history are the ones I've identified most with.
:)

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Michael Mol wrote:
> Doesn't Gentoo have a BSD target? The problem here is with udev, which
> doesn't apply to BSD, AFAIK. Gentoo/BSD might be a good direction to
> go.
>
> Also, where does FreeBSD's kernel stand, with respect to device drivers?
>

If I recall correctly, Gentoo is sort of based on BSD.  I don't think 
using their target would solve the problem with udev tho.

I have no idea on device drivers but I suspect Alan might.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Michael Mol wrote:
> On Fri, Sep 9, 2011 at 9:25 PM, Dale<rdalek1967@gmail.com>  wrote:
>> I may go the BSD route too if I leave Gentoo.  So, my feet works too.  I
>> wonder if I would even be missed here?  :/
> I'd hate it if you left. In the short time I've been on this list,
> your usage habits and history are the ones I've identified most with.
> :)
>

The bad thing is, I like helping people and enjoy this list.  I think me 
and Alan are the top posters here so I guess me and him like helping 
folks.  Alan has a lot of server type experience and I have a bit of 
desktop experience.  We may have some overlap there tho.  Me, I'm a 
desktop user and I like to run a distro that I'm proud of.  In the past, 
it seemed Gentoo sort of lead on some things.  Now, it seems to follow 
instead.  If I want a distro that just follows, I could have stayed with 
Mandrake/Mandriva.  It follows Redhat if I recall correctly.  It also 
uses the init* stuff too.  Which as I pointed out before is one reason I 
left that.  If I got to use one with Gentoo, that just takes one reason 
for using Gentoo and all the compiling stuff away.  Gentoo has some good 
points but lately, they seem to be getting lost on the point scale.

Well, I got divorced once.  I just hope reason will pop up and I don't 
have to shift something important to me again.  This would be as bad as 
me divorcing my ex.  Heck, maybe worse.  I been using Gentoo looooong 
before I met my ex.

< sighs >

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri, 09 Sep 2011 20:25:22 -0500
Dale <rdalek1967@gmail.com> wrote:

> Alan McKinnon wrote:
> > I'm lucky, I can vote with my feet. Out of 140, I have two servers 
> > that *require* Linux. One runs Sybase ASE, the other runs Oracle. 
> > Everything else works like a bomb on FreeBSD. kthankxbyeudev, 
> > thanksfornotplayingnicely Not everyone else is so fortunate though. 
> 
> I guess I understood more than I thought then.  Shocking.  I
> understand that but the udev guru doesn't.  ;-)
> 
> I may go the BSD route too if I leave Gentoo.  So, my feet works
> too.  I wonder if I would even be missed here?  :/
> 
> Dale


Noooooooo Dale you can't leeeeeeeeaaaaaaaaavvvvvveee!

Seriously, you're an institution around here, you would be sorely
missed.


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-10 03:49, Dale wrote:

> If I recall correctly, Gentoo is sort of based on BSD.  I don't think
> using their target would solve the problem with udev tho.

FreeBSD uses "Ports" which Portage is based on, AIUI. The FreeBSD kernel
doesn't use udev. They do have a similar thing though called devd but I
don't think it will be quite the same mess as (future?) udev.

> I have no idea on device drivers but I suspect Alan might.

AFAIK, device driver situation is worse than Linux. How much worse I
don't know but I assume they're trailing behind (i.e. what ever is
supported under Linux will eventually come to FreeBSD as well, assuming
there's enough interest). So if you're patient and not running state of
the art hardware... :-)

Hm... One more thing to try with Linux would be to do without udev (i.e.
static /dev). That is still supported by Gentoo (AFAIK) and could be a
way to solve this mess (yes, you would miss the dynamic creation of
device nodes but...).

As someone else mentioned there's also mdev from busybox (it's mainly
aimed at embedded systems AIUI) to try:
https://wildanm.wordpress.com/2007/08/21/mdev-mini-udev-in-busybox/

Not sure how well supported mdev under Gentoo is though...

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri, 09 Sep 2011 20:58:23 -0500
Dale <rdalek1967@gmail.com> wrote:

> Michael Mol wrote:
> > On Fri, Sep 9, 2011 at 9:25 PM, Dale<rdalek1967@gmail.com>  wrote:
> >> I may go the BSD route too if I leave Gentoo.  So, my feet works
> >> too.  I wonder if I would even be missed here?  :/
> > I'd hate it if you left. In the short time I've been on this list,
> > your usage habits and history are the ones I've identified most
> > with. :)
> >
> 
> The bad thing is, I like helping people and enjoy this list.  I think
> me and Alan are the top posters here so I guess me and him like
> helping folks.  Alan has a lot of server type experience and I have a
> bit of desktop experience.  We may have some overlap there tho.

You give me too much credit :-)

There's also Neil, Wonko, Volker, Stroller, Grant, meino.cramer, Mick,
Paul, Harry, Albert, Alex, Walter, Alan Mackenzie (awesome name!),
James, kashani, Pandu and about a 1000 more whose names I can't exactly
recall right now.

This here mailing-list has got the most varied and highest skills of
any technical list I've ever subscribed to. We have regular desktop
users, folks who work in server rooms, devs, owners of software
companies, regular sysadmins, fellows who ship embedded devices, and at
least one of everything in between.

I don't mean to go all fuzzy feel-good here, but it's an honour to be
able to communicate and interact with so many skilled people for so many
years.


-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Fri, 9 Sep 2011 21:23:42 -0400
Michael Mol <mikemol@gmail.com> wrote:

> On Fri, Sep 9, 2011 at 9:15 PM, Dale <rdalek1967@gmail.com> wrote:
> > Michael Schreckenbauer wrote:
> >>
> >> Am Freitag, 9. September 2011, 19:24:06 schrieb pk:
> >>>
> >>> On 2011-09-09 10:53, Dale wrote:
> >>>>
> >>>> Can I slap whoever started this?  The more I think on this, the
> >>>> worse it
> >>>
> >>> Yes Dale, you have my permission! And while you're at it, slap
> >>> him from me too! ;-)
> >>>
> >>> It _may_ be this guy that's responsible for this crap:
> >>> http://linuxplumbersconf.org/ocw/users/58
> >>>
> >>> Also:
> >>> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
> >>
> >> OMG!
> >> What a mess. udev treats all exit-codes except 0 the same.
> >> That's so bad, I have no words for it.
> >> Definitely not a developer I trust to do things the right way.
> >>
> >>> PS. If things go "tits up" you may want to have a look at FreeBSD
> >>> (or some other BSD). I'm quite sure they wouldn't put up with
> >>> crap like this... I know I will investigate my options at least.
> >>
> >> Agree, FreeBSD is really a fine OS.
> >>
> >>> Best regards
> >>> Peter K
> >>
> >> Thanks for the links,
> >> regards,
> >> Michael
> >>
> >
> > I know one thing, BSD is secure as heck.  I installed it once on a
> > old rig and typed the password in wrong during setup.  I never
> > could get into that thing again.  I had to start over.  lol  That
> > is why I chose Linux in general.  I want something that is secure
> > enough that I don't have to worry about some script kiddie messing
> > with me.
> >
> > BSD is one option I will be looking into if I move from Gentoo.
> >  After all, they are fairly close maybe even a step up.  Especially
> > now.
> 
> Doesn't Gentoo have a BSD target? The problem here is with udev, which
> doesn't apply to BSD, AFAIK. Gentoo/BSD might be a good direction to
> go.
> 
> Also, where does FreeBSD's kernel stand, with respect to device
> drivers?
 
There is a BSD target on Gentoo, but I don't know too much about it (I
just use regular parts on my installs). It seems to suffer from lack of
manpower, starting up and dying down occasionally. Which is a shame as
it has huge potential but Linux grabs most of the limelight.

Driver support is excellent. At least on server grade hardware
everything I've ever bought just works, but I can't comment on desktop
hardware.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> You give me too much credit :-)
>
> There's also Neil, Wonko, Volker, Stroller, Grant, meino.cramer, Mick,
> Paul, Harry, Albert, Alex, Walter, Alan Mackenzie (awesome name!),
> James, kashani, Pandu and about a 1000 more whose names I can't exactly
> recall right now.
>
> This here mailing-list has got the most varied and highest skills of
> any technical list I've ever subscribed to. We have regular desktop
> users, folks who work in server rooms, devs, owners of software
> companies, regular sysadmins, fellows who ship embedded devices, and at
> least one of everything in between.
>
> I don't mean to go all fuzzy feel-good here, but it's an honour to be
> able to communicate and interact with so many skilled people for so many
> years.
>
>

That is true.  There are lots who post a lot here.  I just recall seeing 
some stats somewhere and me and you were the top two.  That was about a 
year ago so it may have changed.  Just had to go find that link again.  
Here it is:

http://archives.gentoo.org/stats/gentoo-user-per-year.xml

We have a new comer.  lol

I think the mailing lists, and forums, are one of the key features of 
Gentoo.  The docs seemed to have slumped some but I think it was down to 
one or two people for a while.  I think someone jumped in the fire a few 
weeks ago tho.  Maybe they will catch up.  I'm sure it is hard to keep 
up with all the changes that are going on tho.  Gentoo has a LOT of 
stuff to document.

If we are so skilled, why is the Fedora dev not listening you reckon?

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> On Fri, 09 Sep 2011 20:25:22 -0500
> Dale<rdalek1967@gmail.com>  wrote:
>
>> Alan McKinnon wrote:
>>> I'm lucky, I can vote with my feet. Out of 140, I have two servers
>>> that *require* Linux. One runs Sybase ASE, the other runs Oracle.
>>> Everything else works like a bomb on FreeBSD. kthankxbyeudev,
>>> thanksfornotplayingnicely Not everyone else is so fortunate though.
>> I guess I understood more than I thought then.  Shocking.  I
>> understand that but the udev guru doesn't.  ;-)
>>
>> I may go the BSD route too if I leave Gentoo.  So, my feet works
>> too.  I wonder if I would even be missed here?  :/
>>
>> Dale
>
> Noooooooo Dale you can't leeeeeeeeaaaaaaaaavvvvvveee!
>
> Seriously, you're an institution around here, you would be sorely
> missed.
>
>

I sometimes think people get tired of the chatter box.  lol   I wonder 
if I am on somebody's blacklist?  :/

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat, 10 Sep 2011 02:54:58 -0500
Dale <rdalek1967@gmail.com> wrote:

> That is true.  There are lots who post a lot here.  I just recall
> seeing some stats somewhere and me and you were the top two.  That
> was about a year ago so it may have changed.  Just had to go find
> that link again. Here it is:
> 
> http://archives.gentoo.org/stats/gentoo-user-per-year.xml
> 
> We have a new comer.  lol

I had absolutely no idea I sent *that* much mail to gentoo-user :-)

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 3544 bytes --]

On Saturday 10 Sep 2011 08:36:59 Alan McKinnon wrote:
> On Fri, 9 Sep 2011 21:23:42 -0400
> 
> Michael Mol <mikemol@gmail.com> wrote:
> > On Fri, Sep 9, 2011 at 9:15 PM, Dale <rdalek1967@gmail.com> wrote:
> > > Michael Schreckenbauer wrote:
> > >> Am Freitag, 9. September 2011, 19:24:06 schrieb pk:
> > >>> On 2011-09-09 10:53, Dale wrote:
> > >>>> Can I slap whoever started this?  The more I think on this, the
> > >>>> worse it
> > >>> 
> > >>> Yes Dale, you have my permission! And while you're at it, slap
> > >>> him from me too! ;-)
> > >>> 
> > >>> It _may_ be this guy that's responsible for this crap:
> > >>> http://linuxplumbersconf.org/ocw/users/58
> > >>> 
> > >>> Also:
> > >>> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
> > >> 
> > >> OMG!
> > >> What a mess. udev treats all exit-codes except 0 the same.
> > >> That's so bad, I have no words for it.
> > >> Definitely not a developer I trust to do things the right way.
> > >> 
> > >>> PS. If things go "tits up" you may want to have a look at FreeBSD
> > >>> (or some other BSD). I'm quite sure they wouldn't put up with
> > >>> crap like this... I know I will investigate my options at least.
> > >> 
> > >> Agree, FreeBSD is really a fine OS.
> > >> 
> > >>> Best regards
> > >>> Peter K
> > >> 
> > >> Thanks for the links,
> > >> regards,
> > >> Michael
> > > 
> > > I know one thing, BSD is secure as heck.  I installed it once on a
> > > old rig and typed the password in wrong during setup.  I never
> > > could get into that thing again.  I had to start over.  lol  That
> > > is why I chose Linux in general.  I want something that is secure
> > > enough that I don't have to worry about some script kiddie messing
> > > with me.
> > > 
> > > BSD is one option I will be looking into if I move from Gentoo.
> > >  After all, they are fairly close maybe even a step up.  Especially
> > > now.
> > 
> > Doesn't Gentoo have a BSD target? The problem here is with udev, which
> > doesn't apply to BSD, AFAIK. Gentoo/BSD might be a good direction to
> > go.
> > 
> > Also, where does FreeBSD's kernel stand, with respect to device
> > drivers?
> 
> There is a BSD target on Gentoo, but I don't know too much about it (I
> just use regular parts on my installs). It seems to suffer from lack of
> manpower, starting up and dying down occasionally. Which is a shame as
> it has huge potential but Linux grabs most of the limelight.
> 
> Driver support is excellent. At least on server grade hardware
> everything I've ever bought just works, but I can't comment on desktop
> hardware.

I've flirted with Slackware before I came over to Gentoo and the reason I 
chose Gentoo is because it gave me more freedom to built and configure an OS 
exactly as I wanted it.  I was at the time thinking of trying BSD with 
portage, but when I was keeping an eye on it there was this start/stop 
development as Alan mentions and loads of packages were unstable for yonks or 
missing completely.  This made me decide to stay on Linux.

I don't think we should give up completely yet.  Perhaps we need to lobby a 
bit more effectively (Can we email directly the dev(s)? Where do they live? 
Ha, ha!)

PS.  Dale please don't leave!  There'll be no mammoth threads without you and 
who are we going to rely to trash the credibility of packages - HAL springs to 
mind!  Come to think of it, given HAL's demise can you also please have a go 
at udev?  You never know ...  ;)
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Dale writes:

> I know one thing, BSD is secure as heck.  I installed it once on a old 
> rig and typed the password in wrong during setup.  I never could get 
> into that thing again.  I had to start over.

That's what you thought :)  Normally, all you have to do is to boot in
single user mode, this gives you a root shell without asking for a
password. Unless you have changed 

    console  none  unknown  off  secure

in /etc/ttys to:

    console  none  unknown  off  insecure

It will then prompt for a password, but even this will not help much. As
long as you have physical access to a machine, you can simply boot it from
a CD or via USB, mount the partitions and remove the password
in /etc/passwd, or simply chroot and do whatever you want. To make it
really secure, you have to encrypt the whole system. Which is fairly easy
BTW.

> lol  That is why I chose 
> Linux in general.  I want something that is secure enough that I don't 
> have to worry about some script kiddie messing with me.

Just make sure to block or disable flash content when surfing the web.

> BSD is one option I will be looking into if I move from Gentoo.  After 
> all, they are fairly close maybe even a step up.  Especially now.

BSD is elegant, simpler, and has some nice features like a file
system that can be checked in the background while being in use already.
With the drawback of being quite slow compared to others.

But I would miss many things. I think portage is much superior these
days. Builds that continue when a package fails, or even parallel builds
are not possible AFAIK. The driver situation is worse I believe, when it
comes to graphics hardware. And I just read [*] that some KDE guys are
rethinking whether they will support other operating systems than Linux
for the plasma desktop, because it may not be worth the effort.

	Wonko

[*] http://blog.martin-graesslin.com/blog/2011/08/thoughts-about-kde-plasma-on-non-linux-systems/

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Dale writes:

> pk wrote:
> > On 2011-09-09 13:35, Alex Schuster wrote:
> >
> >> When I switched to using an initramfs, it was not very complicated. I
> >> simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my
> >> /usr/src/linux/.config and does not change the kernel options. Then
> >> comes genkernel --install --lvm -luks all, and I have kernel and
> >> initramfs
> > And for those that like to do without genkernel? Again, adding another
> > layer for things to go wrong.

That's just what I do and what I know about. I'm reading about dracut
here, but I don't know how easy that is. Sure, I also wouldn't like an
additional layer, but what if there are no good alternatives? 

> I tried genkernel.  All I got was a kernel that wouldn't boot.  Heck,
> it barely even started to boot.  The kernel wouldn't even finish
> loading. After several tries, I put genkernel in the trash.  It worked
> a LOT better there for me.  It was out of sight and mind.  ;-)

Yes, I remember the discussion. But I think you used genkernel as it was
designed, to generate a new kernel .config from scratch. This is not
necessary, as I wrote above you can also make it use your
working .config. That's what I do, and it also gives me the initramfs I
need, without having to think about how it does that. I suggest you just
try it, and I'd say there is a really good chance it just works. When you
use 'genkernel --install kernel', you should get the same kernel as when
you build it manually, just with a different name. With 'genkernel
--install all', you also get the initramfs.
I can't guarantee this, though, and especially you seem to have a history
of being bitten by bugs. But then, that's what people say about me, too,
and I'm using genkernel just fine.

> >> I don't get this one. Why do you want to copy an existing /usr
> >> partition to another one?
> > He said he wishes to move his /usr to a spare partition (the part
> > about KDE4)... I assume his /usr currently resides on / (or maybe a
> > smaller partition that he cannot easily expand).
> 
> You hit it, for some reason I put /usr on the root partition without 
> thinking.  This is where I am now:
> 
> rootfs                19534436  10693048   8841388  55% /
> 
> Over half full.  When I have a critical partition get over 60%, I start 
> looking for expansion.  Moving /usr was my plan but someone stole that 
> from me I guess.  Now I got to figure out what I want to do next.

Uh. So you think about leaving Gentopo, because your root partition is
barely over half full, and moving /usr somewhere else might involve an
initramfs soon? I'd just wait until it starts getting to 80-90%, and
think about it again.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> On Sat, 10 Sep 2011 02:54:58 -0500
> Dale<rdalek1967@gmail.com>  wrote:
>
>> That is true.  There are lots who post a lot here.  I just recall
>> seeing some stats somewhere and me and you were the top two.  That
>> was about a year ago so it may have changed.  Just had to go find
>> that link again. Here it is:
>>
>> http://archives.gentoo.org/stats/gentoo-user-per-year.xml
>>
>> We have a new comer.  lol
> I had absolutely no idea I sent *that* much mail to gentoo-user :-)
>

Me either.  That's when I had to accept that I was a true chatter box.  
O_O  I wonder if Neil knows this?  He may not realize how many he sends 
either.  He comes in third several times.  Does that qualify as a 
chatter box too?  lol

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Pandu Poluan]

[-- Attachment #1: Type: text/plain, Size: 1034 bytes --]

On Sep 10, 2011 10:06 PM, "Dale" <rdalek1967@gmail.com> wrote:
>
> Alan McKinnon wrote:
>>
>> On Sat, 10 Sep 2011 02:54:58 -0500
>> Dale<rdalek1967@gmail.com>  wrote:
>>
>>> That is true.  There are lots who post a lot here.  I just recall
>>> seeing some stats somewhere and me and you were the top two.  That
>>> was about a year ago so it may have changed.  Just had to go find
>>> that link again. Here it is:
>>>
>>> http://archives.gentoo.org/stats/gentoo-user-per-year.xml
>>>
>>> We have a new comer.  lol
>>
>> I had absolutely no idea I sent *that* much mail to gentoo-user :-)
>>
>
> Me either.  That's when I had to accept that I was a true chatter box.
 O_O  I wonder if Neil knows this?  He may not realize how many he sends
either.  He comes in third several times.  Does that qualify as a chatter
box too?  lol

Whoa, I'm number eight on the list?! o_O

That's strange... especially considering I'm quiet through lengthy threads
(e.g., this one)

Dubious achievement... not sure if I should be proud or not -.-

Rgds,

[-- Attachment #2: Type: text/html, Size: 1506 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alex Schuster wrote:
> Dale writes:
>
>> pk wrote:
>>> On 2011-09-09 13:35, Alex Schuster wrote:
>>>
>>>> When I switched to using an initramfs, it was not very complicated. I
>>>> simply use genkernel. With CLEAN="no" and MRPROPER="no", it uses my
>>>> /usr/src/linux/.config and does not change the kernel options. Then
>>>> comes genkernel --install --lvm -luks all, and I have kernel and
>>>> initramfs
>>> And for those that like to do without genkernel? Again, adding another
>>> layer for things to go wrong.
> That's just what I do and what I know about. I'm reading about dracut
> here, but I don't know how easy that is. Sure, I also wouldn't like an
> additional layer, but what if there are no good alternatives?

There is alternative, get a proper fix for udev.  Since udev is needed 
to boot, put those files in /bin or /sbin where they should be.  It's 
really that simple.  If they are truly opposed to that idea, have a /run 
directory or something like that.  Then let it be documented, FHS maybe, 
that that directory has to be on / just like /bin and /sbin.

>
>> I tried genkernel.  All I got was a kernel that wouldn't boot.  Heck,
>> it barely even started to boot.  The kernel wouldn't even finish
>> loading. After several tries, I put genkernel in the trash.  It worked
>> a LOT better there for me.  It was out of sight and mind.  ;-)
> Yes, I remember the discussion. But I think you used genkernel as it was
> designed, to generate a new kernel .config from scratch. This is not
> necessary, as I wrote above you can also make it use your
> working .config. That's what I do, and it also gives me the initramfs I
> need, without having to think about how it does that. I suggest you just
> try it, and I'd say there is a really good chance it just works. When you
> use 'genkernel --install kernel', you should get the same kernel as when
> you build it manually, just with a different name. With 'genkernel
> --install all', you also get the initramfs.
> I can't guarantee this, though, and especially you seem to have a history
> of being bitten by bugs. But then, that's what people say about me, too,
> and I'm using genkernel just fine.

The point is, its one more thing to break and as Alan explains, it 
breaks the tradition of what is required for booting up.  I run into 
enough problems already.  I don't want to add yet one more, to booting 
at that.  I'm sure I'm not the only one that doesn't like init* stuff.  
That is part two of not liking this idea.


>
>>>> I don't get this one. Why do you want to copy an existing /usr
>>>> partition to another one?
>>> He said he wishes to move his /usr to a spare partition (the part
>>> about KDE4)... I assume his /usr currently resides on / (or maybe a
>>> smaller partition that he cannot easily expand).
>> You hit it, for some reason I put /usr on the root partition without
>> thinking.  This is where I am now:
>>
>> rootfs                19534436  10693048   8841388  55% /
>>
>> Over half full.  When I have a critical partition get over 60%, I start
>> looking for expansion.  Moving /usr was my plan but someone stole that
>> from me I guess.  Now I got to figure out what I want to do next.
> Uh. So you think about leaving Gentopo, because your root partition is
> barely over half full, and moving /usr somewhere else might involve an
> initramfs soon? I'd just wait until it starts getting to 80-90%, and
> think about it again.
>
> 	Wonko
>

No, I'm thinking about this because one of the reasons I left binary 
based distros was crap like this.  I didn't like init* stuff because 
they caused me grief when booting.  I wasn't as skilled as I am now so I 
didn't know how to fix them.  Heck, I still don't and really don't want 
to be forced to learn either.  It's the same reason I don't use LVM.  It 
just adds one more layer to cause problems.  I *may* use LVM for data or 
something but not for anything required for booting.

I'm sort of like this.  I want to be able to boot to single user at a 
minimum as simple as possible.  If I can do that, I can fix whatever is 
broke.  If I install a init* thingy and I can't boot because it got 
screwed up somehow, I'm not going to be happy.  I don't get mad often 
but this could be much worse than hal.  Given my medical situation, I 
need to avoid that at all costs.  One of those costs may be me picking 
something else for my OS.  I like Gentoo but the reasons I started using 
Gentoo are slipping away.  I'm running out of reasons to have this 
installed.  It's like having a car.  If you have a old car that breaks a 
lot, at some point you have to decide whether the reason you have the 
car is the same as when you bought the car.  If you can't depend on it 
to get you where you are going, it's time for a new car.  Cars are to 
get a person from point A to point B for most people.  I want a OS that 
is simple enough to get it running but I also like the control Gentoo 
offers or used to.  Yea, I know, Gentoo isn't that simple but it is 
simple in how it works once set up.  Load grub, load kernel, start the 
init process.  Simple.  The init* process is going to add yet one more 
thing to this that shouldn't be needed.  Others with more knowledge than 
me have already posted better ways to fix this.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Mick wrote:
> I've flirted with Slackware before I came over to Gentoo and the 
> reason I chose Gentoo is because it gave me more freedom to built and 
> configure an OS exactly as I wanted it. I was at the time thinking of 
> trying BSD with portage, but when I was keeping an eye on it there was 
> this start/stop development as Alan mentions and loads of packages 
> were unstable for yonks or missing completely. This made me decide to 
> stay on Linux. I don't think we should give up completely yet. Perhaps 
> we need to lobby a bit more effectively (Can we email directly the 
> dev(s)? Where do they live? Ha, ha!) PS. Dale please don't leave! 
> There'll be no mammoth threads without you and who are we going to 
> rely to trash the credibility of packages - HAL springs to mind! Come 
> to think of it, given HAL's demise can you also please have a go at 
> udev? You never know ... ;) 

Well, this is me.  I like things simple.  I have built my own 
computers.  I have rebuilt car engines, small mower engines, worked on 
about every machine, except heavy equipment such as a crane, and done 
well.  What I don't do is add unneeded junk to something.  This init* 
stuff is one of them.  If a person wants to use it, then fine, by all 
means use it.  Thing is, this is going to affect a lot of people in a 
negative way.  When I first installed Mandrake and didn't know a lot 
about Linux, I had /usr on a separate partition.  As ignorant as I was, 
I knew it was a good thing to do.  I clearly wasn't thinking when I put 
/usr on / when I did this install.  Maybe the excitement of my new rig 
got the better of me.  lol

I have to admit, hal got on my bad nerve.  Thing is, I could at least 
boot up to fix it.  Console was still working.  If this init* thing 
breaks, I can't even do that.  Trust me, I'm going to be super duper 
pissed if this init thing fails.  I won't care why.  The point is, I 
don't want the thing to begin with and it shouldn't even be needed.  
There are better ways to do this.

 From my understanding, the dev is not listening.  That is another thing 
that bothers me.  When devs stop listening to users, that causes a 
problem.  Remember hal?  How many people complained early on about the 
config files?  Lots.  I also don't like that a very few people or just 
one person can make a decision like this that will have a negative 
affect on a LOT and I mean a LOT of users.  That is something that needs 
to be dealt with.  What I would like to see is this, a good stable 
alternative that works well with a proper fix and for that to push udev 
out and render it null.  I think that would serve the dev right.  Listen 
to the people that use it or people will use something else.  The mdev 
package comes to mind here.  Maybe this will push it to take udevs 
place.  It seems there is enough people that opposes this.  If a few 
commercial and paying people can help, it may just be the next better 
thing.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Sat, Sep 10, 2011 at 12:09 PM, Dale <rdalek1967@gmail.com> wrote:
> Mick wrote:
> From my understanding, the dev is not listening.  That is another thing that
> bothers me.  When devs stop listening to users, that causes a problem.
>  Remember hal?  How many people complained early on about the config files?
>  Lots.  I also don't like that a very few people or just one person can make
> a decision like this that will have a negative affect on a LOT and I mean a
> LOT of users.  That is something that needs to be dealt with.  What I would
> like to see is this, a good stable alternative that works well with a proper
> fix and for that to push udev out and render it null.  I think that would
> serve the dev right.  Listen to the people that use it or people will use
> something else.  The mdev package comes to mind here.  Maybe this will push
> it to take udevs place.  It seems there is enough people that opposes this.
>  If a few commercial and paying people can help, it may just be the next
> better thing.

As I understand it, nothing of udev itself is in /usr, but instead
packages and scripts which plug themselves into udev to be triggered
by various events.

Perhaps the real solution is to circumvent udev and get those other
packages and scripts to not put hotplug-active files under /usr.


-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Dale writes:

> Alan McKinnon wrote:
> > On Sat, 10 Sep 2011 02:54:58 -0500
> > Dale<rdalek1967@gmail.com>  wrote:
> >
> >> That is true.  There are lots who post a lot here.  I just recall
> >> seeing some stats somewhere and me and you were the top two.  That
> >> was about a year ago so it may have changed.  Just had to go find
> >> that link again. Here it is:
> >>
> >> http://archives.gentoo.org/stats/gentoo-user-per-year.xml
> >>
> >> We have a new comer.  lol
> > I had absolutely no idea I sent *that* much mail to gentoo-user :-)
> 
> Me either.  That's when I had to accept that I was a true chatter box.  
> O_O  I wonder if Neil knows this?  He may not realize how many he sends 
> either.

Since I am on this list, I tend to confuse Alan and Neil. Is this only me?

> He comes in third several times.  Does that qualify as a  chatter box
> too?  lol

Or as yet another very helpful and competent person.

I very much agree with Alan, this is a great list indeed.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Pandu Poluan]

[-- Attachment #1: Type: text/plain, Size: 748 bytes --]

On Sep 10, 2011 11:22 PM, "Michael Mol" <mikemol@gmail.com> wrote:
>
> As I understand it, nothing of udev itself is in /usr, but instead
> packages and scripts which plug themselves into udev to be triggered
> by various events.
>
> Perhaps the real solution is to circumvent udev and get those other
> packages and scripts to not put hotplug-active files under /usr.
>

Agree with both your points.

udev *should* refuse accessing anything under /usr if it's still in the
sysinit phase. After all, that's what the FHS assumed (/usr does not contain
anything required during boot)

IIRC, one of the most vocal designer of the FHS is Red Hat; if the dev of
udev is also Red Hat-related, shouldn't he at least try to follow FHS's
philosophy?

Rgds,

[-- Attachment #2: Type: text/html, Size: 916 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Michael Mol wrote:
> On Sat, Sep 10, 2011 at 12:09 PM, Dale<rdalek1967@gmail.com>  wrote:
>> Mick wrote:
>>  From my understanding, the dev is not listening.  That is another thing that
>> bothers me.  When devs stop listening to users, that causes a problem.
>>   Remember hal?  How many people complained early on about the config files?
>>   Lots.  I also don't like that a very few people or just one person can make
>> a decision like this that will have a negative affect on a LOT and I mean a
>> LOT of users.  That is something that needs to be dealt with.  What I would
>> like to see is this, a good stable alternative that works well with a proper
>> fix and for that to push udev out and render it null.  I think that would
>> serve the dev right.  Listen to the people that use it or people will use
>> something else.  The mdev package comes to mind here.  Maybe this will push
>> it to take udevs place.  It seems there is enough people that opposes this.
>>   If a few commercial and paying people can help, it may just be the next
>> better thing.
> As I understand it, nothing of udev itself is in /usr, but instead
> packages and scripts which plug themselves into udev to be triggered
> by various events.
>
> Perhaps the real solution is to circumvent udev and get those other
> packages and scripts to not put hotplug-active files under /usr.
>
>

Either way, udev is what is the root of it.  Udev is calling for 
something that is not there.  Either udev needs to change or it needs to 
require the files to be somewhere other than /usr.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Sat, Sep 10, 2011 at 12:36 PM, Pandu Poluan <pandu@poluan.info> wrote:
>
> On Sep 10, 2011 11:22 PM, "Michael Mol" <mikemol@gmail.com> wrote:
>>
>> As I understand it, nothing of udev itself is in /usr, but instead
>> packages and scripts which plug themselves into udev to be triggered
>> by various events.
>>
>> Perhaps the real solution is to circumvent udev and get those other
>> packages and scripts to not put hotplug-active files under /usr.
>>
>
> Agree with both your points.
>
> udev *should* refuse accessing anything under /usr if it's still in the
> sysinit phase. After all, that's what the FHS assumed (/usr does not contain
> anything required during boot)
>
> IIRC, one of the most vocal designer of the FHS is Red Hat; if the dev of
> udev is also Red Hat-related, shouldn't he at least try to follow FHS's
> philosophy?

In Portage, we have all kinds of warnings and QA notices pop up when
building things. Could we patch udev to throw warnings when files
under /usr are accessed? (such as by hooking in via strace and
grepping for fopen() calls)

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[William Kenworthy]

On Sat, 2011-09-10 at 02:54 -0500, Dale wrote:
> Alan McKinnon wrote:
> > You give me too much credit :-)
> >
> > There's also Neil, Wonko, Volker, Stroller, Grant, meino.cramer, Mick,
> > Paul, Harry, Albert, Alex, Walter, Alan Mackenzie (awesome name!),
> > James, kashani, Pandu and about a 1000 more whose names I can't exactly
> > recall right now.
> >
> > This here mailing-list has got the most varied and highest skills of
> > any technical list I've ever subscribed to. We have regular desktop
> > users, folks who work in server rooms, devs, owners of software
> > companies, regular sysadmins, fellows who ship embedded devices, and at
> > least one of everything in between.
> >
> > I don't mean to go all fuzzy feel-good here, but it's an honour to be
> > able to communicate and interact with so many skilled people for so many
> > years.
> >
> >
> 
> That is true.  There are lots who post a lot here.  I just recall seeing 
> some stats somewhere and me and you were the top two.  That was about a 
> year ago so it may have changed.  Just had to go find that link again.  
> Here it is:
> 
> http://archives.gentoo.org/stats/gentoo-user-per-year.xml
> 
> We have a new comer.  lol
> 
> I think the mailing lists, and forums, are one of the key features of 
> Gentoo.  The docs seemed to have slumped some but I think it was down to 
> one or two people for a while.  I think someone jumped in the fire a few 
> weeks ago tho.  Maybe they will catch up.  I'm sure it is hard to keep 
> up with all the changes that are going on tho.  Gentoo has a LOT of 
> stuff to document.
> 
> If we are so skilled, why is the Fedora dev not listening you reckon?
> 
> Dale
> 
> :-)  :-)
> 

Ha, I got on the list in 2005!  Unfortunately it doesnt go back into the
90's when I started with Gentoo and the learning curve was REALLY
steep :)

BillK

Re: [gentoo-user] /dev/sda* missing at boot

[William Kenworthy]

On Sun, 2011-09-11 at 01:33 +0800, William Kenworthy wrote:
> On Sat, 2011-09-10 at 02:54 -0500, Dale wrote:
> > Alan McKinnon wrote:
> > > You give me too much credit :-)
> > >
> > > There's also Neil, Wonko, Volker, Stroller, Grant, meino.cramer, Mick,
> > > Paul, Harry, Albert, Alex, Walter, Alan Mackenzie (awesome name!),
> > > James, kashani, Pandu and about a 1000 more whose names I can't exactly
> > > recall right now.
> > >
> > > This here mailing-list has got the most varied and highest skills of
> > > any technical list I've ever subscribed to. We have regular desktop
> > > users, folks who work in server rooms, devs, owners of software
> > > companies, regular sysadmins, fellows who ship embedded devices, and at
> > > least one of everything in between.
> > >
> > > I don't mean to go all fuzzy feel-good here, but it's an honour to be
> > > able to communicate and interact with so many skilled people for so many
> > > years.
> > >
> > >
> > 
> > That is true.  There are lots who post a lot here.  I just recall seeing 
> > some stats somewhere and me and you were the top two.  That was about a 
> > year ago so it may have changed.  Just had to go find that link again.  
> > Here it is:
> > 
> > http://archives.gentoo.org/stats/gentoo-user-per-year.xml
> > 
> > We have a new comer.  lol
> > 
> > I think the mailing lists, and forums, are one of the key features of 
> > Gentoo.  The docs seemed to have slumped some but I think it was down to 
> > one or two people for a while.  I think someone jumped in the fire a few 
> > weeks ago tho.  Maybe they will catch up.  I'm sure it is hard to keep 
> > up with all the changes that are going on tho.  Gentoo has a LOT of 
> > stuff to document.
> > 
> > If we are so skilled, why is the Fedora dev not listening you reckon?
> > 
> > Dale
> > 
> > :-)  :-)
> > 
> 
> Ha, I got on the list in 2005!  Unfortunately it doesnt go back into the
> 90's when I started with Gentoo and the learning curve was REALLY
> steep :)
> 
> BillK
> 
 Actually, thats a bit optimistic - 2002

moriah ~ # ls -alth /var/backups/rattus/20110710/tree/etc/

....

drwxr-xr-x   3 root     root       104 Sep  6  2003 hsf
drwxr-xr-x   2 root     root        80 May 13  2003 sysconfig
drwxr-xr-x   2 root     root        72 Jan  7  2003 devfs.d
drwxr-xr-x   2 root     root        48 Nov  5  2002 svga
drwxr-xr-x   2 root     root        80 Jul 10  2002 sasl
-rw-------   1 root     root       12K Jul  7  2002 .procmailrc.swp
drwxr-xr-x   2 root     root       120 Jul  4  2002 oaf
drwxr-xr-x   3 root     root        72 Jul  3  2002 sound
-rw-------   1 root     root         0 Jul  3  2002 .pwd.lock
drwxr-xr-x   2 root     root       112 Jul  3  2002 ssmtp

Decommissioned a couple of weeks ago!

BillK

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

William Kenworthy wrote:
>   Actually, thats a bit optimistic - 2002
>
> moriah ~ # ls -alth /var/backups/rattus/20110710/tree/etc/
>
> ....
>
> drwxr-xr-x   3 root     root       104 Sep  6  2003 hsf
> drwxr-xr-x   2 root     root        80 May 13  2003 sysconfig
> drwxr-xr-x   2 root     root        72 Jan  7  2003 devfs.d
> drwxr-xr-x   2 root     root        48 Nov  5  2002 svga
> drwxr-xr-x   2 root     root        80 Jul 10  2002 sasl
> -rw-------   1 root     root       12K Jul  7  2002 .procmailrc.swp
> drwxr-xr-x   2 root     root       120 Jul  4  2002 oaf
> drwxr-xr-x   3 root     root        72 Jul  3  2002 sound
> -rw-------   1 root     root         0 Jul  3  2002 .pwd.lock
> drwxr-xr-x   2 root     root       112 Jul  3  2002 ssmtp
>
> Decommissioned a couple of weeks ago!
>
> BillK
>

I do mine this way.  I joined the forums when I installed Gentoo.  So, 
this is pretty close to how long people have put up with me:

Joined: 19 Sep 2003
Posts: 1240
Location: Mississippi USA

I joined the mailing list shortly after that.  I'm amazed ya'll have put 
up with me this long.  lol

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat, 10 Sep 2011 18:34:42 +0200
Alex Schuster <wonko@wonkology.org> wrote:

> > Me either.  That's when I had to accept that I was a true chatter
> > box. O_O  I wonder if Neil knows this?  He may not realize how many
> > he sends either.  
> 
> Since I am on this list, I tend to confuse Alan and Neil. Is this
> only me?

Alan's > Me either.  That's when I had to accept that I was a true
chatter box.  
> O_O  I wonder if Neil knows this?  He may not realize how many he
> sends either.  

Since I am on this list, I tend to confuse Alan and Neil. Is this only
me? girlfriend says that Alan and Neil are both male bald middle-aged
pedantic old gits with a fascination for the writing of Douglas Adams.
And they are both grammar Nazis.

She is not in the least surprised you get them confused. If Neil ever
confesses to owning and riding motorcycles, she thinks she might get
them mixed up herself.

:-)

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sat, 10 Sep 2011 12:19:10 -0400
Michael Mol <mikemol@gmail.com> wrote:

> On Sat, Sep 10, 2011 at 12:09 PM, Dale <rdalek1967@gmail.com> wrote:
> > Mick wrote:
> > From my understanding, the dev is not listening.  That is another
> > thing that bothers me.  When devs stop listening to users, that
> > causes a problem. Remember hal?  How many people complained early
> > on about the config files? Lots.  I also don't like that a very few
> > people or just one person can make a decision like this that will
> > have a negative affect on a LOT and I mean a LOT of users.  That is
> > something that needs to be dealt with.  What I would like to see is
> > this, a good stable alternative that works well with a proper fix
> > and for that to push udev out and render it null.  I think that
> > would serve the dev right.  Listen to the people that use it or
> > people will use something else.  The mdev package comes to mind
> > here.  Maybe this will push it to take udevs place.  It seems there
> > is enough people that opposes this. If a few commercial and paying
> > people can help, it may just be the next better thing.
> 
> As I understand it, nothing of udev itself is in /usr, but instead
> packages and scripts which plug themselves into udev to be triggered
> by various events.
> 
> Perhaps the real solution is to circumvent udev and get those other
> packages and scripts to not put hotplug-active files under /usr.

That's my understanding too, and I agree with your conclusions. The
distros can easily (give enough man-power) deal with this too - they
simply have to modify their rpms/debs/pkgs/ebuilds to install specific
identified things to / instead of /usr. They *already* do this for
packages that natively install to peculiar locations.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Alan McKinnon wrote:
> Since I am on this list, I tend to confuse Alan and Neil. Is this only 
> me? girlfriend says that Alan and Neil are both male bald middle-aged 
> pedantic old gits with a fascination for the writing of Douglas Adams. 
> And they are both grammar Nazis. She is not in the least surprised you 
> get them confused. If Neil ever confesses to owning and riding 
> motorcycles, she thinks she might get them mixed up herself. :-) 

Well, I still got my hair.  ROFL  I guess we all have our oddities tho.  
;-)

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Keith Dart]

=== On Fri, 09/09, Alex Schuster wrote: ===
> What I fear much more is when good old grub is no longer supported
> and I have to use grub2, which I tried to understand, but failed.

===

Ya, it's horrid. But the {sys,ext}linux bootloader is still there and
maintained and I like it better. I use extlinux on all my Gentoo
systems now. 


-- Keith Dart

-- 

-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   Keith Dart <keith@dartworks.biz>
   public key: ID: 19017044
   <http://www.dartworks.biz/>
   =====================================================================

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Alan McKinnon writes:

> On Sat, 10 Sep 2011 18:34:42 +0200
> Alex Schuster <wonko@wonkology.org> wrote:

> > Since I am on this list, I tend to confuse Alan and Neil. Is this
> > only me?

At least I know by now that you are the South Africa guy.

> Alan's girlfriend says that Alan and Neil are both male bald middle-aged
> pedantic old gits with a fascination for the writing of Douglas Adams.

I like that :-) The last part.

> And they are both grammar Nazis.

And I thought that was Peter Humphrey... or are all of you the same
person? Who can tell.

> She is not in the least surprised you get them confused. If Neil ever
> confesses to owning and riding motorcycles, she thinks she might get
> them mixed up herself.

So I wonder what Neil will write about this.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Keith Dart writes:

> === On Fri, 09/09, Alex Schuster wrote: ===
> > What I fear much more is when good old grub is no longer supported
> > and I have to use grub2, which I tried to understand, but failed.
>
> Ya, it's horrid. But the {sys,ext}linux bootloader is still there and
> maintained and I like it better. I use extlinux on all my Gentoo
> systems now. 

Interesting. What are the advantages?

What I like most about Grub is the interactive shell. And that I don't
have to run a command like I had to do with Lilo after installing a new
kernel.

I guess Grub 1 will be around for a long time, and there will be no
need for me to switch soon. But I had to deal with Grub 2 on other
installations, and I had many problems. And I was disappointed because
configuring it is so much more complicated. It may have cooler features
and suport more file systems, but setting up grub 1 most of the times was
grub, root (hd0,0), setup (hd0), quit and setting up 2-3 lines in
grub.conf.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Keith Dart]

=== On Sun, 09/11, Alex Schuster wrote: ===
> Interesting. What are the advantages?

Mainly that it's simpler, as a bootloader should be. However it does
have some nice features, such as making nice looking, interactive
menus. You can also edit the config file by hand, if you need to, and
it's all contained on the boot partition.

The biggest problem with grub 2 is it adds a dependency on having your
main root partition already mounted in order to configure it. That may
not be available. Also, when you learn extlinux then you know syslinux,
isolinux, and pxelinux already which helps when configuring boot
loaders for those other media. 

> What I like most about Grub is the interactive shell. And that I don't
> have to run a command like I had to do with Lilo after installing a
> new kernel.

If you need a shell, boot a minimal kernel and shell from a ramdisk. No
good reason to bloat a bootloader with that. 


-- Keith Dart

-- 

-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   Keith Dart <keith@dartworks.biz>
   public key: ID: 19017044
   <http://www.dartworks.biz/>
   =====================================================================

Re: [gentoo-user] /dev/sda* missing at boot

[Paul Colquhoun]

On Fri, 9 Sep 2011 07:24:06 PM pk wrote:
> On 2011-09-09 10:53, Dale wrote:
> > Can I slap whoever started this?  The more I think on this, the worse it
> 
> Yes Dale, you have my permission! And while you're at it, slap him from
> me too! ;-)
> 
> It _may_ be this guy that's responsible for this crap:
> http://linuxplumbersconf.org/ocw/users/58
> 
> Also:
> http://comments.gmane.org/gmane.linux.hotplug.devel/16994


I've had a look at the stuff at those links, and some of what they link to in 
turn, and had a bit of a think about it.

Looking at "initramfs" as a modern Linux replacement for the "bootable / 
partition" of traditional Unix systems does make some sense, even though I 
think it could be made simpler.

Fot those opposed to initramfs, would you also object to /boot being
  1) a manditory seperate partition
  2) required to be ext2 (or one of a *very* short list)
  3) having /boot/{bin,sbin,lib} containing local copies of the absolute
      minimum boot requirements (i.e. initramfs in a real fs)

On the other hand, most of the problem seems to stem from software packages 
hooking into the early boot via udev rules, and not beiong careful where they 
put the executables and libraries that they reference.

Is udev (as it currently stands) really the best place for them to hook into?

Could udev be split into 2 passes, early-boot udev that only does system stuff 
(like mount filesystems out of /etc/fstab, setup keyboards & video), and late-
boot udev where other applications can put in any hooks they like, since the 
full system would then be available.

The late-boot udev may need to do a full rescan of everything that early-boot 
udev found, but didn't have the rules for yet, but I'm sure that the 2 passes 
could talk to each other and sort that out fairly simply.

Or possibly just add a whole new service to use just for hooking software 
packages into system events. Although this would probably end upneeding to be 
a udev clone anyway.


-- 
Reverend Paul Colquhoun, ULC.    http://andor.dropbear.id.au/~paulcol
 Before you criticize someone, you should walk a mile in their shoes.
Then, when you do, you'll be a mile away, and you'll have their shoes.

Re: [gentoo-user] /dev/sda* missing at boot

[Alan McKinnon]

On Sun, 11 Sep 2011 13:16:48 +1000
Paul Colquhoun <paulcol@andor.dropbear.id.au> wrote:

> I've had a look at the stuff at those links, and some of what they
> link to in turn, and had a bit of a think about it.
> 
> Looking at "initramfs" as a modern Linux replacement for the
> "bootable / partition" of traditional Unix systems does make some
> sense, even though I think it could be made simpler.
> 
> Fot those opposed to initramfs, would you also object to /boot being
>   1) a manditory seperate partition
>   2) required to be ext2 (or one of a *very* short list)
>   3) having /boot/{bin,sbin,lib} containing local copies of the
> absolute minimum boot requirements (i.e. initramfs in a real fs)

For my part, I don't object to any of those. The Unix boot system is
generic enough that one should be able to build whatever one wants.
Only a very few things are required:

the kernel must be accessible to the bootloader
the root partition must be accessible to the kernel
init must be available early

everything else is optional

How the distro (or user) makes this happen should be up to them, not up
to udev. I understand that udev opens up all manner of
future possibilities and these could be very useful. But I do object
to a single package breaking all the foundation assumptions, especially
when the package is now being used in ways not originally envisaged.

udev is a dynamic device node controller. It is not a hotplug framework
and should not be dictating how the rest of the stack must be arranged.

-- 
Alan McKinnnon
alan.mckinnon@gmail.com

Re: [gentoo-user] /dev/sda* missing at boot

[Mike Edenfield]

On 9/10/2011 5:28 PM, Alan McKinnon wrote:
> On Sat, 10 Sep 2011 12:19:10 -0400
> Michael Mol<mikemol@gmail.com>  wrote:
>
>> On Sat, Sep 10, 2011 at 12:09 PM, Dale<rdalek1967@gmail.com>  wrote:
>>> Mick wrote:
>>>  From my understanding, the dev is not listening.  That is another
>>> thing that bothers me.  When devs stop listening to users, that
>>> causes a problem. Remember hal?  How many people complained early

 From what I read, he's listening, he just isn't being 
swayed by the argument. From his perspective, udev "doesn't 
support" a split /,/usr because of the arbitrarily complex 
udev rules. This is causing users to fill their bug queue 
with errors when needed binaries are unavailable at boot, 
and thus their hardware doesn't work. Apparently he has 
concluded that the number of people who require a separate 
/usr partition but cannot use an initramfs is smaller than 
the number of people who need udev to have access to all of 
/usr.

Unfortunately it appears that he's taking a pretty extreme 
approach to solving the problem that will actually *break* 
the systems of that second group, which I don't quite 
understand the reasoning behind.

>> As I understand it, nothing of udev itself is in /usr, but instead
>> packages and scripts which plug themselves into udev to be triggered
>> by various events.
>>
>> Perhaps the real solution is to circumvent udev and get those other
>> packages and scripts to not put hotplug-active files under /usr.
>
> That's my understanding too, and I agree with your conclusions. The
> distros can easily (give enough man-power) deal with this too - they
> simply have to modify their rpms/debs/pkgs/ebuilds to install specific
> identified things to / instead of /usr. They *already* do this for
> packages that natively install to peculiar locations.

It would make perfect sense to me for the udev maintainer to 
simply declare a split /,/usr "not supported" and let us 
deal with the issues. The problem, if I'm reading correctly, 
is that he's taken things one step further and decided to 
move udev *itself* back into /usr.

Even still, I would think that a Gentoo patchset to revert 
the paths back to /lib would be a feasible workaround to 
this mess.

--Mike

Re: [gentoo-user] /dev/sda* missing at boot

[Pandu Poluan]

[-- Attachment #1: Type: text/plain, Size: 796 bytes --]

On Sep 11, 2011 3:25 PM, "Mike Edenfield" <kutulu@kutulu.org> wrote:
>
> It would make perfect sense to me for the udev maintainer to simply
declare a split /,/usr "not supported" and let us deal with the issues. The
problem, if I'm reading correctly, is that he's taken things one step
further and decided to move udev *itself* back into /usr.
>

100% agree!

> Even still, I would think that a Gentoo patchset to revert the paths back
to /lib would be a feasible workaround to this mess.
>
> --Mike
>

Yes, please!

And I'm sure there will be no shortage of testers among Gentoo users.

Heck, I hereby volunteer myself to be a tester if Gentoo devs go forth with
patching udev.

(I have several VMs on VMware ESX and XenServer where I can test an
initr*-free and separated-/usr environ)

Rgds,

[-- Attachment #2: Type: text/html, Size: 995 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Peter Humphrey]

On Saturday 10 September 2011 23:35:56 Alex Schuster wrote:
> Alan McKinnon writes:

> > And they are both grammar Nazis.
> 
> And I thought that was Peter Humphrey... or are all of you the same
> person? Who can tell.

First among equals? And seventh on the list!

> > She is not in the least surprised you get them confused. If Neil ever
> > confesses to owning and riding motorcycles, she thinks she might get
> > them mixed up herself.
> 
> So I wonder what Neil will write about this.

He seems to be lying low.

-- 
Rgds
Peter		Linux Counter 5290, 1994-04-23

Re: [gentoo-user] /dev/sda* missing at boot

[pk]

On 2011-09-10 18:09, Dale wrote:

> From my understanding, the dev is not listening.  That is another thing
> that bothers me.  When devs stop listening to users, that causes a

AFAIU he doesn't listen to people not running RHEL/Fedora (or any of the
big binary distros). For a binary distro, that most likely already are
using an initrd thingie, this works fine. In my mind it also makes it
more difficult to support your own kernel (patches etc.) under these
binary distros making you more dependent on the distro supplier. If you
want control of what goes into your machine then this works less well...
Now if you were a _big_ customer of RHEL that wanted to keep udev
working like it currently does I think you might have some more leverage
(i.e. if the developer refuses to keep things working, then someone at
Red Hat would probably step in at the benefit of their customer and do
the right thing(tm)).

Best regards

Peter K

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Keith Dart writes:

> === On Sun, 09/11, Alex Schuster wrote: ===
> > Interesting. What are the advantages?
> 
> Mainly that it's simpler, as a bootloader should be. However it does
> have some nice features, such as making nice looking, interactive
> menus. You can also edit the config file by hand, if you need to, and
> it's all contained on the boot partition.
> 
> The biggest problem with grub 2 is it adds a dependency on having your
> main root partition already mounted in order to configure it. That may
> not be available. Also, when you learn extlinux then you know syslinux,
> isolinux, and pxelinux already which helps when configuring boot
> loaders for those other media.

Thanks for the explanation. I like to learn, knowing how to use them
might come handy some time. 
I already installed syslinux recently, I think that was necessary for the
installation of systemrescuecd on USB. Which failed, after using the
installer, the stick was still empty. No idea what went wrong, I did not
dig further into this, I was too busy then.

> > What I like most about Grub is the interactive shell. And that I don't
> > have to run a command like I had to do with Lilo after installing a
> > new kernel.
> 
> If you need a shell, boot a minimal kernel and shell from a ramdisk. No
> good reason to bloat a bootloader with that. 

I still like how I could make Grub boot a system even when I did not know
on which partition it was. This happened a couple of times, like when I
had multiple hard drives that changed their order. Tab completion or the
find command were good to have then.
And about the bloat... 450 K being used for Grub in /boot is okay for me.
Which could probably be reduced further down to ~115K when removing
stage2{.old,_eltorito} and support for other file systems than ext2.

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Paul Colquhoun writes:

> Looking at "initramfs" as a modern Linux replacement for the
> "bootable / partition" of traditional Unix systems does make some
> sense, even though I think it could be made simpler.
> 
> Fot those opposed to initramfs, would you also object to /boot being
>   1) a manditory seperate partition
>   2) required to be ext2 (or one of a *very* short list)
>   3) having /boot/{bin,sbin,lib} containing local copies of the absolute
>       minimum boot requirements (i.e. initramfs in a real fs)

I had this on one machine. I used the stuff that Dirk Heinrich offered
[*] (he simply calls it initfs), and it sort of worked, but I also got
some errors. Anyway, I always wondered why this is not the standard way.
Sure, having a single intr{d,amfs} file is convenient, but every time I
want to have a look into it, I have to google the cpio syntax in order to
extract stuff. While, with an initfs, you simply see everything as plain
files in the /boot partition.

	Wonko

[*] http://www.mail-archive.com/gentoo-user@lists.gentoo.org/msg88055.html

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Paul Colquhoun wrote:
> On Fri, 9 Sep 2011 07:24:06 PM pk wrote:
>> On 2011-09-09 10:53, Dale wrote:
>>> Can I slap whoever started this?  The more I think on this, the worse it
>> Yes Dale, you have my permission! And while you're at it, slap him from
>> me too! ;-)
>>
>> It _may_ be this guy that's responsible for this crap:
>> http://linuxplumbersconf.org/ocw/users/58
>>
>> Also:
>> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
>
> I've had a look at the stuff at those links, and some of what they link to in
> turn, and had a bit of a think about it.
>
> Looking at "initramfs" as a modern Linux replacement for the "bootable /
> partition" of traditional Unix systems does make some sense, even though I
> think it could be made simpler.
>
> Fot those opposed to initramfs, would you also object to /boot being
>    1) a manditory seperate partition
>    2) required to be ext2 (or one of a *very* short list)
>    3) having /boot/{bin,sbin,lib} containing local copies of the absolute
>        minimum boot requirements (i.e. initramfs in a real fs)
>
> On the other hand, most of the problem seems to stem from software packages
> hooking into the early boot via udev rules, and not beiong careful where they
> put the executables and libraries that they reference.
>
> Is udev (as it currently stands) really the best place for them to hook into?
>
> Could udev be split into 2 passes, early-boot udev that only does system stuff
> (like mount filesystems out of /etc/fstab, setup keyboards&  video), and late-
> boot udev where other applications can put in any hooks they like, since the
> full system would then be available.
>
> The late-boot udev may need to do a full rescan of everything that early-boot
> udev found, but didn't have the rules for yet, but I'm sure that the 2 passes
> could talk to each other and sort that out fairly simply.
>
> Or possibly just add a whole new service to use just for hooking software
> packages into system events. Although this would probably end upneeding to be
> a udev clone anyway.
>

I always have /boot on a separate partition and it is always ext2.  So, 
that is done.  I also have a 200Mb /boot partition.  It sometimes gets 
about half full but I could just clean out old kernels more often.  I 
could always make /boot larger too.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 2661 bytes --]

On Sunday 11 Sep 2011 19:56:48 Dale wrote:
> Paul Colquhoun wrote:
> > On Fri, 9 Sep 2011 07:24:06 PM pk wrote:
> >> On 2011-09-09 10:53, Dale wrote:
> >>> Can I slap whoever started this?  The more I think on this, the worse
> >>> it
> >> 
> >> Yes Dale, you have my permission! And while you're at it, slap him from
> >> me too! ;-)
> >> 
> >> It _may_ be this guy that's responsible for this crap:
> >> http://linuxplumbersconf.org/ocw/users/58
> >> 
> >> Also:
> >> http://comments.gmane.org/gmane.linux.hotplug.devel/16994
> > 
> > I've had a look at the stuff at those links, and some of what they link
> > to in turn, and had a bit of a think about it.
> > 
> > Looking at "initramfs" as a modern Linux replacement for the "bootable /
> > partition" of traditional Unix systems does make some sense, even though
> > I think it could be made simpler.
> > 
> > Fot those opposed to initramfs, would you also object to /boot being
> > 
> >    1) a manditory seperate partition
> >    2) required to be ext2 (or one of a *very* short list)
> >    3) having /boot/{bin,sbin,lib} containing local copies of the absolute
> >    
> >        minimum boot requirements (i.e. initramfs in a real fs)
> > 
> > On the other hand, most of the problem seems to stem from software
> > packages hooking into the early boot via udev rules, and not beiong
> > careful where they put the executables and libraries that they
> > reference.
> > 
> > Is udev (as it currently stands) really the best place for them to hook
> > into?
> > 
> > Could udev be split into 2 passes, early-boot udev that only does system
> > stuff (like mount filesystems out of /etc/fstab, setup keyboards& 
> > video), and late- boot udev where other applications can put in any
> > hooks they like, since the full system would then be available.
> > 
> > The late-boot udev may need to do a full rescan of everything that
> > early-boot udev found, but didn't have the rules for yet, but I'm sure
> > that the 2 passes could talk to each other and sort that out fairly
> > simply.
> > 
> > Or possibly just add a whole new service to use just for hooking software
> > packages into system events. Although this would probably end upneeding
> > to be a udev clone anyway.
> 
> I always have /boot on a separate partition and it is always ext2.  So,
> that is done.  I also have a 200Mb /boot partition.  It sometimes gets
> about half full but I could just clean out old kernels more often.  I
> could always make /boot larger too.

It seems that I'm gonna have fun with a 35M /boot soon (and no LVM of course).
 ;-)
-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Mick wrote:
> On Sunday 11 Sep 2011 19:56:48 Dale wrote:
>
> I always have /boot on a separate partition and it is always ext2.  So,
> that is done.  I also have a 200Mb /boot partition.  It sometimes gets
> about half full but I could just clean out old kernels more often.  I
> could always make /boot larger too.
> It seems that I'm gonna have fun with a 35M /boot soon (and no LVM of course).
>   ;-)

I'm doing some thinking and reading.  I'm either going to go back to a 
rpm based thing and let something besides me deal with the init* stuff 
or stick around and dive into this init* crap and add LVM on top.  /boot 
would be the only thing not on LVM.  This makes me nervous as heck tho.  
I have read where if something goes wrong, you can lose everything.  I'm 
hoping I can make mine simple enough that I can manage any problems even 
if I can get no outside help.  From what I have read, usually it's when 
you can't figure out how to fix it that you lose everything.

I'm just not sure which I want to do right now.  I may put my spare 
drive to work here pretty soon tho. Either another distro or playing 
with the init* and LVM stuff.

< sighs >

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1: Type: text/plain, Size: 2288 bytes --]

On Sun, 11 Sep 2011 16:07:23 -0500, Dale wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> Mick wrote:
> > On Sunday 11 Sep 2011 19:56:48 Dale wrote:
> >
> > I always have /boot on a separate partition and it is always ext2.
> > So, that is done.  I also have a 200Mb /boot partition.  It
> > sometimes gets about half full but I could just clean out old
> > kernels more often.  I could always make /boot larger too.
> > It seems that I'm gonna have fun with a 35M /boot soon (and no LVM
> > of course). ;-)
> 
> I'm doing some thinking and reading.  I'm either going to go back to
> a rpm based thing and let something besides me deal with the init*
> stuff

IMO, better to use Debian or Slackware.  I went through "RPM Hell" back
in the days when I ran S.u.S.E. (complete with full-stops in the name)
and I will never go back.

> or stick around and dive into this init* crap and add LVM on
> top.

Watch this space.  You might read something to your advantage in the
next few days.

> /boot would be the only thing not on LVM.

Well, /boot cannot be on LVM, as the BIOS does not know about logical
volumes.

> This makes me
> nervous as heck tho. I have read where if something goes wrong, you
> can lose everything.

It's no worse than a normal partitioning system, just more flexible.
[Of course, that also means that it is more flexible for you to destroy
your DASD farm yourself.]

> I'm hoping I can make mine simple enough that I
> can manage any problems even if I can get no outside help.  From what
> I have read, usually it's when you can't figure out how to fix it
> that you lose everything.

Same as partitions: just keep backups.

I have some scripts that generate LVM rebuild scripts.  These scan the
current logical volumes and generate lvcreate commands into a script
that can rebuild your LVM set-up in seconds.  You (or anybody else) are
welcome to a copy if you wish.

After that, back up the contents using tar, dar, cpio or whatever your
favourite archiving tool happens to be.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

David W Noon wrote:
> On Sun, 11 Sep 2011 16:07:23 -0500, Dale wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> Mick wrote:
>>> On Sunday 11 Sep 2011 19:56:48 Dale wrote:
>>>
>>> I always have /boot on a separate partition and it is always ext2.
>>> So, that is done.  I also have a 200Mb /boot partition.  It
>>> sometimes gets about half full but I could just clean out old
>>> kernels more often.  I could always make /boot larger too.
>>> It seems that I'm gonna have fun with a 35M /boot soon (and no LVM
>>> of course). ;-)
>> I'm doing some thinking and reading.  I'm either going to go back to
>> a rpm based thing and let something besides me deal with the init*
>> stuff
> IMO, better to use Debian or Slackware.  I went through "RPM Hell" back
> in the days when I ran S.u.S.E. (complete with full-stops in the name)
> and I will never go back.

If I decide to switch, I'll do like I did before Gentoo.  Just read and 
see what best suites my needs.  I do hate the RPM stuff tho.  The 
updates for Mandrake was a nightmare.

>> or stick around and dive into this init* crap and add LVM on
>> top.
> Watch this space.  You might read something to your advantage in the
> next few days.

O_O  Both eyes wide open and watching.  I'm hoping to get new glasses in 
the next few days.  My post count may go up then.  lol   I may be a 
turbo charged chatter box.  ROFL

>
>> /boot would be the only thing not on LVM.
> Well, /boot cannot be on LVM, as the BIOS does not know about logical
> volumes.

I knew there was a reason I had to do that.  Sometimes I know something 
but not the reason behind it.

>> This makes me
>> nervous as heck tho. I have read where if something goes wrong, you
>> can lose everything.
> It's no worse than a normal partitioning system, just more flexible.
> [Of course, that also means that it is more flexible for you to destroy
> your DASD farm yourself.]

When I was reading about problems with LVM, I think it was mostly a lack 
of experience in the repair process.  Basically, something went wrong, 
typed in the wrong command and it got messy from there.  It wasn't LVM 
itself but the clueless geek in the chair.  That may be me before to 
long.  :/

>> I'm hoping I can make mine simple enough that I
>> can manage any problems even if I can get no outside help.  From what
>> I have read, usually it's when you can't figure out how to fix it
>> that you lose everything.
> Same as partitions: just keep backups.
>
> I have some scripts that generate LVM rebuild scripts.  These scan the
> current logical volumes and generate lvcreate commands into a script
> that can rebuild your LVM set-up in seconds.  You (or anybody else) are
> welcome to a copy if you wish.
>
> After that, back up the contents using tar, dar, cpio or whatever your
> favourite archiving tool happens to be.

I generally backup my /etc and world file on a USB stick.  I also have a 
backup of my scripts in /root somewhere around here.  I don't have 
enough space to backup everything tho.  Before DSL came along, I could 
backup to DVD-RWs from time to time but not now.  DSL is addictive.  lol

I'm hoping to find a 2 or 3Tb drive one day.  I can have a partition for 
back up and some data too.  If I get a faster DSL package, I may need 
two of those drives.

Looking forward to the new info you are working on.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[James Wall]

On Sun, Sep 11, 2011 at 4:46 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> On Sun, 11 Sep 2011 16:07:23 -0500, Dale wrote about Re:
> [gentoo-user] /dev/sda* missing at boot:
>
>> Mick wrote:
>> > On Sunday 11 Sep 2011 19:56:48 Dale wrote:
>> >
>> > I always have /boot on a separate partition and it is always ext2.
>> > So, that is done.  I also have a 200Mb /boot partition.  It
>> > sometimes gets about half full but I could just clean out old
>> > kernels more often.  I could always make /boot larger too.
>> > It seems that I'm gonna have fun with a 35M /boot soon (and no LVM
>> > of course). ;-)
>>
>> I'm doing some thinking and reading.  I'm either going to go back to
>> a rpm based thing and let something besides me deal with the init*
>> stuff
>
> IMO, better to use Debian or Slackware.  I went through "RPM Hell" back
> in the days when I ran S.u.S.E. (complete with full-stops in the name)
> and I will never go back.
>
>> or stick around and dive into this init* crap and add LVM on
>> top.
>
> Watch this space.  You might read something to your advantage in the
> next few days.
>
>> /boot would be the only thing not on LVM.
>
> Well, /boot cannot be on LVM, as the BIOS does not know about logical
> volumes.
>
>> This makes me
>> nervous as heck tho. I have read where if something goes wrong, you
>> can lose everything.
>
> It's no worse than a normal partitioning system, just more flexible.
> [Of course, that also means that it is more flexible for you to destroy
> your DASD farm yourself.]
>
>> I'm hoping I can make mine simple enough that I
>> can manage any problems even if I can get no outside help.  From what
>> I have read, usually it's when you can't figure out how to fix it
>> that you lose everything.
>
> Same as partitions: just keep backups.
>
> I have some scripts that generate LVM rebuild scripts.  These scan the
> current logical volumes and generate lvcreate commands into a script
> that can rebuild your LVM set-up in seconds.  You (or anybody else) are
> welcome to a copy if you wish.

I am interested in the backup scripts to help improve my backup/restore system.

> After that, back up the contents using tar, dar, cpio or whatever your
> favourite archiving tool happens to be.
> --
> Regards,
>
> Dave  [RLU #314465]
> *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
> dwnoon@ntlworld.com (David W Noon)
> *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
>



-- 
No trees were harmed in the sending of this message. However, a large
number of electrons were terribly inconvenienced.

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Saturday, September 10, 2011 02:54:58 AM Dale wrote:
> Alan McKinnon wrote:
> > You give me too much credit :-)
> > 
> > There's also Neil, Wonko, Volker, Stroller, Grant, meino.cramer, Mick,
> > Paul, Harry, Albert, Alex, Walter, Alan Mackenzie (awesome name!),
> > James, kashani, Pandu and about a 1000 more whose names I can't exactly
> > recall right now.
> > 
> > This here mailing-list has got the most varied and highest skills of
> > any technical list I've ever subscribed to. We have regular desktop
> > users, folks who work in server rooms, devs, owners of software
> > companies, regular sysadmins, fellows who ship embedded devices, and at
> > least one of everything in between.
> > 
> > I don't mean to go all fuzzy feel-good here, but it's an honour to be
> > able to communicate and interact with so many skilled people for so many
> > years.

I agree here. There are a few other lists with people with really good 
technical skills. But some of those are quite strict with what is on-topic and 
what isn't.
On this list, we tend to cover anything that is related to computers.

> That is true.  There are lots who post a lot here.  I just recall seeing
> some stats somewhere and me and you were the top two.  That was about a
> year ago so it may have changed.  Just had to go find that link again.
> Here it is:
> 
> http://archives.gentoo.org/stats/gentoo-user-per-year.xml

Nice, I'm in the top 10 :)

> We have a new comer.  lol
> 
> I think the mailing lists, and forums, are one of the key features of
> Gentoo.  The docs seemed to have slumped some but I think it was down to
> one or two people for a while.  I think someone jumped in the fire a few
> weeks ago tho.  Maybe they will catch up.  I'm sure it is hard to keep
> up with all the changes that are going on tho.  Gentoo has a LOT of
> stuff to document.

Documenting Gentoo is difficult. I think this list is a good start for 
documentation though.

> If we are so skilled, why is the Fedora dev not listening you reckon?

Is the Fedora dev aware of non-Fedora installations?

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Saturday, September 10, 2011 02:56:48 AM Dale wrote:
> Alan McKinnon wrote:
> > On Fri, 09 Sep 2011 20:25:22 -0500
> > 
> > Dale<rdalek1967@gmail.com>  wrote:
> >> Alan McKinnon wrote:
> >>> I'm lucky, I can vote with my feet. Out of 140, I have two servers
> >>> that *require* Linux. One runs Sybase ASE, the other runs Oracle.
> >>> Everything else works like a bomb on FreeBSD. kthankxbyeudev,
> >>> thanksfornotplayingnicely Not everyone else is so fortunate though.
> >> 
> >> I guess I understood more than I thought then.  Shocking.  I
> >> understand that but the udev guru doesn't.  ;-)
> >> 
> >> I may go the BSD route too if I leave Gentoo.  So, my feet works
> >> too.  I wonder if I would even be missed here?  :/
> >> 
> >> Dale
> > 
> > Noooooooo Dale you can't leeeeeeeeaaaaaaaaavvvvvveee!
> > 
> > Seriously, you're an institution around here, you would be sorely
> > missed.
> 
> I sometimes think people get tired of the chatter box.  lol   I wonder
> if I am on somebody's blacklist?  :/

If you are, that person is missing out on some good entertainment :)

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Friday, September 09, 2011 07:24:06 PM pk wrote:
> On 2011-09-09 10:53, Dale wrote:
> > Can I slap whoever started this?  The more I think on this, the worse it
> 
> Yes Dale, you have my permission! And while you're at it, slap him from
> me too! ;-)
> 
> It _may_ be this guy that's responsible for this crap:
> http://linuxplumbersconf.org/ocw/users/58
> 
> Also:
> http://comments.gmane.org/gmane.linux.hotplug.devel/16994

Interesting read, also that link for systemd.

What about the following as a gentoo-solution:

As long as filesystem-support for /usr is in the kernel, why can't "/usr" be 
mounted right after "/"?

Eg. instead of worrying with an init*, why not edit the boot-scripts to have 
"/usr" mounted before udev and colleagues start?

mount is still in /bin
fstab is still in /etc
Both should be available during boot.

A script that does:
mount /
check /etc/fstab to see if "/usr" is seperate
if yes: mount /usr

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Joost Roeleveld wrote:
> On Saturday, September 10, 2011 02:56:48 AM Dale wrote:
>> I sometimes think people get tired of the chatter box. lol I wonder 
>> if I am on somebody's blacklist? :/ 
> If you are, that person is missing out on some good entertainment :)
>
>

That may depend on my meds.  lol

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Sunday, September 11, 2011 08:44:20 PM James Wall wrote:
> On Sun, Sep 11, 2011 at 4:46 PM, David W Noon <dwnoon@ntlworld.com> wrote:
> > On Sun, 11 Sep 2011 16:07:23 -0500, Dale wrote about Re:
> > 
> > [gentoo-user] /dev/sda* missing at boot:
> >> Mick wrote:
> >> > On Sunday 11 Sep 2011 19:56:48 Dale wrote:
> >> > 
> >> > I always have /boot on a separate partition and it is always ext2.
> >> > So, that is done.  I also have a 200Mb /boot partition.  It
> >> > sometimes gets about half full but I could just clean out old
> >> > kernels more often.  I could always make /boot larger too.
> >> > It seems that I'm gonna have fun with a 35M /boot soon (and no LVM
> >> > of course). ;-)
> >> 
> >> I'm doing some thinking and reading.  I'm either going to go back to
> >> a rpm based thing and let something besides me deal with the init*
> >> stuff
> > 
> > IMO, better to use Debian or Slackware.  I went through "RPM Hell" back
> > in the days when I ran S.u.S.E. (complete with full-stops in the name)
> > and I will never go back.

Don't remind me, I used to install RPM-systems with the option "install 
everything" just to avoid having to find all the dependencies.
After install, I'd compile my own software (installing over distro-supplied 
files) or simply do a full new install. (Like I do with MS Windows...)

> >> or stick around and dive into this init* crap and add LVM on
> >> top.
> > 
> > Watch this space.  You might read something to your advantage in the
> > next few days.

If you're building something and needs testers, let me know.

> > I have some scripts that generate LVM rebuild scripts.  These scan the
> > current logical volumes and generate lvcreate commands into a script
> > that can rebuild your LVM set-up in seconds.  You (or anybody else) are
> > welcome to a copy if you wish.
> 
> I am interested in the backup scripts to help improve my backup/restore
> system.

Same here, I've been using LVM for a while and I generally remember how to fix 
things when it breaks. But as these occurences are now rare and far between, I 
always need to find my old notes again and then update them to new syntax.

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Joost Roeleveld writes:

> What about the following as a gentoo-solution:
> 
> As long as filesystem-support for /usr is in the kernel, why can't
> "/usr" be mounted right after "/"?
> 
> Eg. instead of worrying with an init*, why not edit the boot-scripts to
> have "/usr" mounted before udev and colleagues start?
> 
> mount is still in /bin
> fstab is still in /etc
> Both should be available during boot.

But there are no /dev/sd* entries yet for the device /usr is on. That's
what udev is for in the first place, creating them.

We could add those devices manually, like the essential /dev/console
and /dev/null that also have to be there before udev kicks in. Might be
simpler than creating the initramfs thing. But probably only with real
disk partitions. For LVM, many more devices will be necessary, and I
don't creating them all by hand might not be so easy.

When udev does so many things these days, couldn't udev itself mount
the /usr partition, and then continue with the rules
in /etc/udev/rules.d/?

But I really think that either udev should just not depend on stuff
in /usr, or consist of two stages, one for the essential device nodes,
and one that is run later, after /usr is mounted, dealing with stuff
in /etc/udev/rules.d. Which will not solve the problems with a bluetooth
keyboard, though. But for most of us it might work.

Just thinking,

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 783 bytes --]

On Sat, 10 Sep 2011 09:59:41 -0500, Dale wrote:

> >> http://archives.gentoo.org/stats/gentoo-user-per-year.xml

> > I had absolutely no idea I sent *that* much mail to gentoo-user :-)
> >  
> 
> Me either.  That's when I had to accept that I was a true chatter box.  
> O_O  I wonder if Neil knows this?  He may not realize how many he sends 
> either.  He comes in third several times.  Does that qualify as a 
> chatter box too?

I have never in my life been accused of talking too much (and if you
believe that, you'll believe anything). Although the figures look high,
it only works out to around 2 mails per day - or 3 per day if you only
post in work time :)


-- 
Neil Bothwick

Confucius says "He who posts with broken addresses gets no replies......"

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 480 bytes --]

On Sat, 10 Sep 2011 18:34:42 +0200, Alex Schuster wrote:

> > Me either.  That's when I had to accept that I was a true chatter
> > box. O_O  I wonder if Neil knows this?  He may not realize how many
> > he sends either.  
> 
> Since I am on this list, I tend to confuse Alan and Neil. Is this only
> me?

No, it's not only you. Dale confuses the hell out of me regularly ;-)


-- 
Neil Bothwick

Being politically correct means always having to say you're sorry.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[gentoo-user] Re: /dev/sda* missing at boot

[Nicolas Sebrecht]

The 09/09/11, Michael Schreckenbauer wrote:

> The question arose, when Canek mentioned bluetoothd, that udev seems to need 
> in some cases.

This is wrong. udev on its own does not require extra tools from /usr.
Though, the rules used by udev do use software in /usr. It's NOT a udev
fault _at all_.

This is how developers wrote software and because they wanted to hook
themselves early at boot time, using udev facility. They are PulseAudio,
NetworkManager, libatasmart, ALSA, D-Bus, CUPS, VirtualBox, usbmuxd,
bluetoothd and a LOT of other tools. It's even worse when you know that
some scripts are written in python. Everybody can write its own rules
without even think about direct (or hidden) /usr dependency.
Again, udev is NOT to blame.

>                If bluetoothd doesn't quite fit to /bin or /sbin (I tend to 
> agree here), but is needed before /usr is mounted, then it has to be put 
> *somewhere*. I don't say, that this is the way to go. Only searching for 
> alternatives to a forced initramfs.

So, what's the good way to fix all that mess? Certainly not moving most
of software to /. Fortunately, we can expect /usr to be mounted before
udev starts via the initramfs.

It does NOT mean everybody will require a initramfs. It means people
WANTING a seperate /usr will need a initramfs.

The good thing is that a lot of tools now in / will be granted back to
/usr. Let's clean up /. Also, it's a _good_ news for admins expecting to
maintain systems with a shared /usr (e.g. over the network).

-- 
Nicolas Sebrecht

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 789 bytes --]

On Sat, 10 Sep 2011 23:15:52 +0200, Alan McKinnon wrote:

> Since I am on this list, I tend to confuse Alan and Neil. Is this only
> me? girlfriend says that Alan and Neil are both male bald middle-aged
> pedantic old gits with a fascination for the writing of Douglas Adams.
> And they are both grammar Nazis.

I still strenuously deny being balding, although I won't be able to do so
for much longer :(

> She is not in the least surprised you get them confused. If Neil ever
> confesses to owning and riding motorcycles, she thinks she might get
> them mixed up herself.

I haven't ridden a bike for years, too many injuries from the days I used
to race them. Those walls on the Isle of Man are rather hard.


-- 
Neil Bothwick

"Criminal Lawyer" is a redundancy.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 288 bytes --]

On Sun, 11 Sep 2011 10:37:25 +0100, Peter Humphrey wrote:

> > So I wonder what Neil will write about this.  
> 
> He seems to be lying low.

Just in an area with very poor Internet access. I'm back in England now :)


-- 
Neil Bothwick

Top Oxymorons Number 31: Small crowd

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1309 bytes --]

On Mon, 12 Sep 2011 09:45:44 +0200, Joost Roeleveld wrote:

> As long as filesystem-support for /usr is in the kernel, why can't
> "/usr" be mounted right after "/"?
> 
> Eg. instead of worrying with an init*, why not edit the boot-scripts to
> have "/usr" mounted before udev and colleagues start?

Because it is udev that creates the device entries needed to mount /usr -
and that doesn't even touch other cases, like /usr being on a software
block device, like LVM or dmcrypt.

The problem here is that udev is trying to do too much. On the one hand
it handles the initial population of /dev/ and all that is needed to
mount the contents of fstab. On the other hand, it is trying to be an
all-encompassing device and hotplug manager. the latter function should
be started relatively late in the boot sequence, the former as soon as
possible.

I'd like to know why these functions cannot be separated, run the command
to populate /dev early on, then start the udev daemon after the
filesystems have been mounted.

Some sort of early boot rules file would need to be used to handle things
like setting up symlinks for block devices to avoid breaking some users'
fstabs.


-- 
Neil Bothwick

It may be that your sole purpose in life is simply to serve as a
warning to others.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Alex Schuster]

Neil Bothwick writes:

> On Sat, 10 Sep 2011 18:34:42 +0200, Alex Schuster wrote:

> > Since I am on this list, I tend to confuse Alan and Neil. Is this only
> > me?

Whoops, which should be: I tend to confuse Alan _with_ Neil. But then,
both may be right.

> No, it's not only you. Dale confuses the hell out of me regularly ;-)

Me too :)

	Wonko

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Monday, September 12, 2011 09:49:22 AM Neil Bothwick wrote:
> On Mon, 12 Sep 2011 09:45:44 +0200, Joost Roeleveld wrote:
> > As long as filesystem-support for /usr is in the kernel, why can't
> > "/usr" be mounted right after "/"?
> > 
> > Eg. instead of worrying with an init*, why not edit the boot-scripts to
> > have "/usr" mounted before udev and colleagues start?
> 
> Because it is udev that creates the device entries needed to mount /usr -
> and that doesn't even touch other cases, like /usr being on a software
> block device, like LVM or dmcrypt.

Thanks Alex and Neil.

I didn't think it through properly. Which is why I posted it here, rather then 
try to see how to get the scripts updated for it.

> The problem here is that udev is trying to do too much. On the one hand
> it handles the initial population of /dev/ and all that is needed to
> mount the contents of fstab. On the other hand, it is trying to be an
> all-encompassing device and hotplug manager. the latter function should
> be started relatively late in the boot sequence, the former as soon as
> possible.
> 
> I'd like to know why these functions cannot be separated, run the command
> to populate /dev early on, then start the udev daemon after the
> filesystems have been mounted.
> 
> Some sort of early boot rules file would need to be used to handle things
> like setting up symlinks for block devices to avoid breaking some users'
> fstabs.

Yes, which means "udev" would need to be split into:
* devd (which controls the /dev-tree)
* plugd (which handles all the "hotplug-events" where special things happen)

The communication between the 2 could be done using a simple /dev/udev_pipe 
device. "devd" throws events onto the pipe and plugd handles these events.

That would also make things easier to configure as the renaming and such is 
for "devd". But the commands to be executed can then be based on the actual 
name in /dev. Rather then on the kernel-name/id/..../whatever.

Any thoughts on this?

--
Joost

PS. I'm throwing ideas here, hopefully we can come to a sane and logical 
option here

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1171 bytes --]

On Mon, 12 Sep 2011 11:07:12 +0200, Joost Roeleveld wrote:

> > I'd like to know why these functions cannot be separated, run the
> > command to populate /dev early on, then start the udev daemon after
> > the filesystems have been mounted.
> > 
> > Some sort of early boot rules file would need to be used to handle
> > things like setting up symlinks for block devices to avoid breaking
> > some users' fstabs.  
> 
> Yes, which means "udev" would need to be split into:
> * devd (which controls the /dev-tree)
> * plugd (which handles all the "hotplug-events" where special things
> happen)
> 
> The communication between the 2 could be done using a
> simple /dev/udev_pipe device. "devd" throws events onto the pipe and
> plugd handles these events.

I wonder if it could be done more simply. udevd loads but only parses
those rule files marked as suitable for early boot time. Later in the
boot it switches to "full" mode and loads all rule files.

This is so simple it is either pure genius or completely naive and
unworkable. I know which option my money is on...


-- 
Neil Bothwick

Diarrhoea is hereditary, it runs in your genes.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] Re: /dev/sda* missing at boot

[Michael Schreckenbauer]

Hi,

On Monday, 12. September 2011 10:40:02 Nicolas Sebrecht wrote:
> The 09/09/11, Michael Schreckenbauer wrote:
> > The question arose, when Canek mentioned bluetoothd, that udev seems to
> > need in some cases.
> 
> This is wrong.
> udev on its own does not require extra tools from /usr.
> Though, the rules used by udev do use software in /usr. It's NOT a udev
> fault _at all_.

Well, this is details. Where's the diffference from user-point-of-view, whether 
it's udev itself or some scripts executed by udev?
And I tend to disagree, with the not udev's fault part.
udev treats all exit-codes from scripts as if the device were not present.
This includes errors of all kinds. How is this supposed to work at all?

> So, what's the good way to fix all that mess? Certainly not moving most
> of software to /. Fortunately, we can expect /usr to be mounted before
> udev starts via the initramfs.

That's *your* opinion. Most people on this list disagree.

> It does NOT mean everybody will require a initramfs. It means people
> WANTING a seperate /usr will need a initramfs.
> The good thing is that a lot of tools now in / will be granted back to
> /usr. Let's clean up /. Also, it's a _good_ news for admins expecting to
> maintain systems with a shared /usr (e.g. over the network).

Since when is a mandatory initramfs a good thing for admins?
Care to explain?

Regards,
Michael

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Monday, September 12, 2011 10:13:45 AM Neil Bothwick wrote:
> On Mon, 12 Sep 2011 11:07:12 +0200, Joost Roeleveld wrote:
> > > I'd like to know why these functions cannot be separated, run the
> > > command to populate /dev early on, then start the udev daemon after
> > > the filesystems have been mounted.
> > > 
> > > Some sort of early boot rules file would need to be used to handle
> > > things like setting up symlinks for block devices to avoid breaking
> > > some users' fstabs.
> > 
> > Yes, which means "udev" would need to be split into:
> > * devd (which controls the /dev-tree)
> > * plugd (which handles all the "hotplug-events" where special things
> > happen)
> > 
> > The communication between the 2 could be done using a
> > simple /dev/udev_pipe device. "devd" throws events onto the pipe and
> > plugd handles these events.
> 
> I wonder if it could be done more simply. udevd loads but only parses
> those rule files marked as suitable for early boot time. Later in the
> boot it switches to "full" mode and loads all rule files.
> 
> This is so simple it is either pure genius or completely naive and
> unworkable. I know which option my money is on...

This would depend on wether or not udev (or whatever program handles the 
events) can pick specific events out of the queue.

I think the events are placed on a queue waiting for some process to handle 
them and that process then does the following in an endless loop:
1) get event from queue
2) handle event

In order to split the 2 options, there needs to be something that sorts them 
between "init-level" and "run-level" events where "init-level" is what is 
needed/possible during boot.

As I currently understand it, the kernel does not support cherry-picking / 
multiple queues for hotplug-events and all devices cause a hotplug-event for 
the /dev-tree creation part of "udev".

A second queue will need to be handled somehow.
I also don't see why "udev" needs to get the additional code to handle 
delaying running external tools when this could be split off into seperate 
process.

This way, if the program/script that is configured in the udev-rules causes a 
system-crash, avoiding the handler for these to start up, will actually 
provide a better fail-safe.

The part that creates the dev-tree will still run and has become smaller and 
simpler.

Would a udev-fork work for Gentoo? 

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1473 bytes --]

On Mon, 12 Sep 2011 11:34:17 +0200, Joost Roeleveld wrote:

> > I wonder if it could be done more simply. udevd loads but only parses
> > those rule files marked as suitable for early boot time. Later in the
> > boot it switches to "full" mode and loads all rule files.
> > 
> > This is so simple it is either pure genius or completely naive and
> > unworkable. I know which option my money is on...  
> 
> This would depend on wether or not udev (or whatever program handles
> the events) can pick specific events out of the queue.
> 
> I think the events are placed on a queue waiting for some process to
> handle them and that process then does the following in an endless loop:
> 1) get event from queue
> 2) handle event
> 
> In order to split the 2 options, there needs to be something that sorts
> them between "init-level" and "run-level" events where "init-level" is
> what is needed/possible during boot.
 
If the rules are not loaded, the events are ignored. They are not
run handled until the full set of rules are loaded later on.

Alternatively, the first rules file parsed would do whatever is necessary
for other rules to be handled, such as doing whatever is necessary
to make all programs and libraries available. Then it would be the
responsible of who/whatever adds the rules to make sure the setup is
complete.


-- 
Neil Bothwick

To be sure of hitting the target, shoot first and call whatever you hit
the target.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Mike Edenfield]

On 9/12/2011 3:12 AM, Joost Roeleveld wrote:
> On Saturday, September 10, 2011 02:54:58 AM Dale wrote:
>> If we are so skilled, why is the Fedora dev not listening you reckon?
>
> Is the Fedora dev aware of non-Fedora installations?

He is, because a Gentoo user/dev explicitly pointed out the 
problems this will cause Gentoo.

His response, to me, appeared to be a heavy dose of "way 
more people use Fedora/Debian/etc than Gentoo so I'm 
tailoring my fix to those people" combined with a touch of 
"if you're running Gentoo you're smart enough to figure this 
out on your own". Possibly with a subtle, hidden hint of 
"that's what you get for not running Fedora", but I could be 
imagining that.

--Mike

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Monday, September 12, 2011 08:14:57 AM Mike Edenfield wrote:
> On 9/12/2011 3:12 AM, Joost Roeleveld wrote:
> > On Saturday, September 10, 2011 02:54:58 AM Dale wrote:
> >> If we are so skilled, why is the Fedora dev not listening you reckon?
> > 
> > Is the Fedora dev aware of non-Fedora installations?
> 
> He is, because a Gentoo user/dev explicitly pointed out the
> problems this will cause Gentoo.

Awareness comes at different levels.
It's like the difference of looking and seeing. :)

He seems to recall there is a world outside of Fedora, but doesn't seem to 
believe it...

> His response, to me, appeared to be a heavy dose of "way
> more people use Fedora/Debian/etc than Gentoo so I'm
> tailoring my fix to those people" combined with a touch of
> "if you're running Gentoo you're smart enough to figure this
> out on your own". Possibly with a subtle, hidden hint of
> "that's what you get for not running Fedora", but I could be
> imagining that.

Of that's how he sees it, then he is admitting that Gentoo-users are smarter 
then he is....
I like the compliment :)

--
Joost

Re: [gentoo-user] /dev/sda* missing at boot

[Pandu Poluan]

On Mon, Sep 12, 2011 at 19:28, Joost Roeleveld <joost@antarean.org> wrote:
> On Monday, September 12, 2011 08:14:57 AM Mike Edenfield wrote:
>> His response, to me, appeared to be a heavy dose of "way
>> more people use Fedora/Debian/etc than Gentoo so I'm
>> tailoring my fix to those people" combined with a touch of
>> "if you're running Gentoo you're smart enough to figure this
>> out on your own". Possibly with a subtle, hidden hint of
>> "that's what you get for not running Fedora", but I could be
>> imagining that.
>
> Of that's how he sees it, then he is admitting that Gentoo-users are smarter
> then he is....
> I like the compliment :)
>

That's a nice way of finding the silver lining, Joost :-D

That said...

Anyone up to forking udev? What will we be needing?

I can volunteer virtual servers (on top of XenServer and/or VMware --
take your pick).

And maybe one physical server.

Rgds,
-- 
FdS Pandu E Poluan
~ IT Optimizer ~

 • LOPSA Member #15248
 • Blog : http://pepoluan.tumblr.com
 • Linked-In : http://id.linkedin.com/in/pepoluan

Re: [gentoo-user] /dev/sda* missing at boot

[Michael Mol]

On Mon, Sep 12, 2011 at 10:47 AM, Pandu Poluan <pandu@poluan.info> wrote:
> On Mon, Sep 12, 2011 at 19:28, Joost Roeleveld <joost@antarean.org> wrote:
>> On Monday, September 12, 2011 08:14:57 AM Mike Edenfield wrote:
>>> His response, to me, appeared to be a heavy dose of "way
>>> more people use Fedora/Debian/etc than Gentoo so I'm
>>> tailoring my fix to those people" combined with a touch of
>>> "if you're running Gentoo you're smart enough to figure this
>>> out on your own". Possibly with a subtle, hidden hint of
>>> "that's what you get for not running Fedora", but I could be
>>> imagining that.
>>
>> Of that's how he sees it, then he is admitting that Gentoo-users are smarter
>> then he is....
>> I like the compliment :)
>>
>
> That's a nice way of finding the silver lining, Joost :-D
>
> That said...
>
> Anyone up to forking udev? What will we be needing?
>
> I can volunteer virtual servers (on top of XenServer and/or VMware --
> take your pick).
>
> And maybe one physical server.

Interested (it gives me an opportunity to learn a great deal about
another area of the system), though I've never hacked on anything like
udev, or anything early in the boot process, before. I'd probably be
limited to testing.

-- 
:wq

Re: [gentoo-user] /dev/sda* missing at boot

[Joost Roeleveld]

On Monday, September 12, 2011 11:29:12 AM Michael Mol wrote:
> On Mon, Sep 12, 2011 at 10:47 AM, Pandu Poluan <pandu@poluan.info> wrote:
> > On Mon, Sep 12, 2011 at 19:28, Joost Roeleveld <joost@antarean.org> wrote:
> >> On Monday, September 12, 2011 08:14:57 AM Mike Edenfield wrote:
> >>> His response, to me, appeared to be a heavy dose of "way
> >>> more people use Fedora/Debian/etc than Gentoo so I'm
> >>> tailoring my fix to those people" combined with a touch of
> >>> "if you're running Gentoo you're smart enough to figure this
> >>> out on your own". Possibly with a subtle, hidden hint of
> >>> "that's what you get for not running Fedora", but I could be
> >>> imagining that.
> >> 
> >> Of that's how he sees it, then he is admitting that Gentoo-users are
> >> smarter then he is....
> >> I like the compliment :)
> > 
> > That's a nice way of finding the silver lining, Joost :-D
> > 
> > That said...
> > 
> > Anyone up to forking udev? What will we be needing?
> > 
> > I can volunteer virtual servers (on top of XenServer and/or VMware --
> > take your pick).
> > 
> > And maybe one physical server.
> 
> Interested (it gives me an opportunity to learn a great deal about
> another area of the system), though I've never hacked on anything like
> udev, or anything early in the boot process, before. I'd probably be
> limited to testing.

I'm also interested. Not entirely sure how much I can help.
Testing, definitely. Coding, I'll try. :)

--
Joost

Re: [gentoo-user] Re: /dev/sda* missing at boot

[Dale]

Michael Schreckenbauer wrote:
> Hi,
>
> On Monday, 12. September 2011 10:40:02 Nicolas Sebrecht wrote:
>
>> So, what's the good way to fix all that mess? Certainly not moving most
>> of software to /. Fortunately, we can expect /usr to be mounted before
>> udev starts via the initramfs.
> That's *your* opinion. Most people on this list disagree.
>
>> It does NOT mean everybody will require a initramfs. It means people
>> WANTING a seperate /usr will need a initramfs.
>> The good thing is that a lot of tools now in / will be granted back to
>> /usr. Let's clean up /. Also, it's a _good_ news for admins expecting to
>> maintain systems with a shared /usr (e.g. over the network).
> Since when is a mandatory initramfs a good thing for admins?
> Care to explain?
>
> Regards,
> Michael
>
>

Couldn't agree more.

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[Dale]

Pandu Poluan wrote:
> On Mon, Sep 12, 2011 at 19:28, Joost Roeleveld<joost@antarean.org>  wrote:
>> On Monday, September 12, 2011 08:14:57 AM Mike Edenfield wrote:
>>> His response, to me, appeared to be a heavy dose of "way
>>> more people use Fedora/Debian/etc than Gentoo so I'm
>>> tailoring my fix to those people" combined with a touch of
>>> "if you're running Gentoo you're smart enough to figure this
>>> out on your own". Possibly with a subtle, hidden hint of
>>> "that's what you get for not running Fedora", but I could be
>>> imagining that.
>> Of that's how he sees it, then he is admitting that Gentoo-users are smarter
>> then he is....
>> I like the compliment :)
>>
> That's a nice way of finding the silver lining, Joost :-D
>
> That said...
>
> Anyone up to forking udev? What will we be needing?
>
> I can volunteer virtual servers (on top of XenServer and/or VMware --
> take your pick).
>
> And maybe one physical server.
>
> Rgds,

I noticed a new directory the other day.  I found out it belongs to 
cups.  It is named Resources.  If cups can use that, why not put the 
stuff udev needs in there?

Dale

:-)  :-)

Re: [gentoo-user] /dev/sda* missing at boot

[David W Noon]

[-- Attachment #1.1: Type: text/plain, Size: 1169 bytes --]

On Sun, 11 Sep 2011 20:44:20 -0500, James Wall wrote about Re:
[gentoo-user] /dev/sda* missing at boot:

> On Sun, Sep 11, 2011 at 4:46 PM, David W Noon <dwnoon@ntlworld.com>
> wrote:
[snip]
> > I have some scripts that generate LVM rebuild scripts.  These scan
> > the current logical volumes and generate lvcreate commands into a
> > script that can rebuild your LVM set-up in seconds.  You (or
> > anybody else) are welcome to a copy if you wish.
> 
> I am interested in the backup scripts to help improve my
> backup/restore system.

Attached.  I hope this list permits binary attachments.  Reply by
private email if it doesn't get through.

Note that it is a zsh script, so you'll need zsh installed.  The output
script will run under any shell.

I keep mine installed in /usr/local/bin/.  You can test the script by
running:

   lvm_rebuild.zsh | less

and you should see the output script displayed on the screen.
-- 
Regards,

Dave  [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
dwnoon@ntlworld.com (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

[-- Attachment #1.2: lvm_rebuild.tgz --]
[-- Type: application/x-compressed-tar, Size: 1050 bytes --]

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-user] /dev/sda* missing at boot

[Walter Dnes]

On Thu, Sep 08, 2011 at 04:03:53PM -0400, Canek Pel??ez Vald??s wrote

> I answered that already (actually, in that paragraph). But again: udev
> is not trivial, and it solves a (far from) trivial problem. If some
> developers think they can outsmart the kernel devs, please, lets try
> it. Maybe they will.

  A fraction of 1% of linux users need to run initramfs or keep /usr on
/.  Why should the remaining 99%+ be required to follow suit?

-- 
Walter Dnes <waltdnes@waltdnes.org>