[gentoo-user] boot problems

[gentoo-user] boot problems

[Stefan G. Weichinger]

I see boot problems on two of my machines ... very likely related to
lvm2 as far as I can tell so far.

Downgraded and even disabled systemd, re-emerged stuff ... I am still at
trying to find the reason. Maybe it is with the mdadm raids also ...

I just want to ask you other gentoo users if anyone else hit this issue
since around yesterday? Maybe someone else also scratches his head
already ...

Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On 3 May 2014 12:12:09 CEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>
>I see boot problems on two of my machines ... very likely related to
>lvm2 as far as I can tell so far.
>
>Downgraded and even disabled systemd, re-emerged stuff ... I am still
>at
>trying to find the reason. Maybe it is with the mdadm raids also ...
>
>I just want to ask you other gentoo users if anyone else hit this issue
>since around yesterday? Maybe someone else also scratches his head
>already ...
>
>Stefan

Can you provide more info on the kind of boot issues you are having?

Based on what you gave, only difference between your systems and mine is systemd. And my systems boot.
This leads to the assumption the issue is caused by systemd.

--
Joost

Ps. This is NOT an anti-systemd remark :)
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 12:27, schrieb J. Roeleveld:
> On 3 May 2014 12:12:09 CEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>>
>> I see boot problems on two of my machines ... very likely related to
>> lvm2 as far as I can tell so far.
>>
>> Downgraded and even disabled systemd, re-emerged stuff ... I am still
>> at
>> trying to find the reason. Maybe it is with the mdadm raids also ...
>>
>> I just want to ask you other gentoo users if anyone else hit this issue
>> since around yesterday? Maybe someone else also scratches his head
>> already ...
>>
>> Stefan
> 
> Can you provide more info on the kind of boot issues you are having?
> 
> Based on what you gave, only difference between your systems and mine is systemd. And my systems boot.
> This leads to the assumption the issue is caused by systemd.
> 
> --
> Joost
> 
> Ps. This is NOT an anti-systemd remark :)

Yeah, sure ;-)

Reporting issues/bugs from a system running with systemd very often
leads to these fundamental statements "see what happens ... !?" :-)

I am still trying to spot the reason. As mentioned in my other reply
currently it seems to also timeout/hang with openrc for me.

Right now I downgraded systemd, rebuilt dbus, dracut, procps ...

From a live-cd I am able to mount /, start the 2 raid-arrays, activate
VGs and LVs ... everything useable.

Currently I rebooted kernel 3.14.2 and get another:

"A start job is running for Activation of LVM2 logical volumes" ... from
systemd-212-r2 (tested -r3 and stable 208 as well)

Removing the "real_init" from the kernel line chooses openrc and also
leads to a hanging boot.

So a next step seems to comment the LVs from fstab and check if it boots
then.

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 12:33:11 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 12:27, schrieb J. Roeleveld:
> > On 3 May 2014 12:12:09 CEST, "Stefan G. Weichinger" <lists@xunil.at> 
wrote:
> >> I see boot problems on two of my machines ... very likely related to
> >> lvm2 as far as I can tell so far.
> >> 
> >> Downgraded and even disabled systemd, re-emerged stuff ... I am still
> >> at
> >> trying to find the reason. Maybe it is with the mdadm raids also ...
> >> 
> >> I just want to ask you other gentoo users if anyone else hit this issue
> >> since around yesterday? Maybe someone else also scratches his head
> >> already ...
> >> 
> >> Stefan
> > 
> > Can you provide more info on the kind of boot issues you are having?
> > 
> > Based on what you gave, only difference between your systems and mine is
> > systemd. And my systems boot. This leads to the assumption the issue is
> > caused by systemd.
> > 
> > --
> > Joost
> > 
> > Ps. This is NOT an anti-systemd remark :)
> 
> Yeah, sure ;-)
> 
> Reporting issues/bugs from a system running with systemd very often
> leads to these fundamental statements "see what happens ... !?" :-)

I tend to try not to.
If you had mentioned a random package X instead of systemd, my comment would 
have mentioned random package X.

I don't use systemd, but as it's pushed by Redhat and redhat, by default, uses 
LVM, I would expect that part to work correctly together.

> I am still trying to spot the reason. As mentioned in my other reply
> currently it seems to also timeout/hang with openrc for me.
> 
> Right now I downgraded systemd, rebuilt dbus, dracut, procps ...

lvm-utils?

> From a live-cd I am able to mount /, start the 2 raid-arrays, activate
> VGs and LVs ... everything useable.

Ok, to get this straight.
You have x amount of disks
those are in 2 raid-arrays
on top of that, you have LVM...

Can you check that LVM is WAITING for the MDADM arrays to be build?

This sounds like an issue I was having a while ago. I had to add an extra 
dependency to the start-up scripts at the time.

> Currently I rebooted kernel 3.14.2 and get another:
> 
> "A start job is running for Activation of LVM2 logical volumes" ... from
> systemd-212-r2 (tested -r3 and stable 208 as well)
> 
> Removing the "real_init" from the kernel line chooses openrc and also
> leads to a hanging boot.
> 
> So a next step seems to comment the LVs from fstab and check if it boots
> then.

Or enable interactive boot where you can press "I" to have every service ask 
if it needs to be started. Then skip LVM.
Do you have parallel-boot enabled with OpenRC?
If yes, disable that just to make sure that isn't causing the issue with 
OpenRC.
With systemd, I would recommend checking the dependencies between mdadm and 
lvm.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 12:54, schrieb J. Roeleveld:

> I don't use systemd, but as it's pushed by Redhat and redhat, by default, uses 
> LVM, I would expect that part to work correctly together.

Yes ... but I still don't fully get what services to enable to get it
set up correctly. lvmetad yes or no, various systemd-generators ... to
me it looks a bit confusing. I had a local lvm.service ... and disabled
it already.


>> I am still trying to spot the reason. As mentioned in my other reply
>> currently it seems to also timeout/hang with openrc for me.
>>
>> Right now I downgraded systemd, rebuilt dbus, dracut, procps ...
> 
> lvm-utils?

No such package in portage on my thinkpad ... (using this machine to reply).

>> From a live-cd I am able to mount /, start the 2 raid-arrays, activate
>> VGs and LVs ... everything useable.
> 
> Ok, to get this straight.
> You have x amount of disks
> those are in 2 raid-arrays
> on top of that, you have LVM...

one ssd for / and /home ... 2 hdds building 2 mdadm-arrays as PVs, VGs
on top ... LVs for data/music/VMs/etc


> Can you check that LVM is WAITING for the MDADM arrays to be build?

I will try to. Disabling it from a non-booting system is a bit
difficult. Commenting the fstab-lines containing LVs didn't help to skip
this step so far (seems there is this LVM-systemd-generator called)

>> Removing the "real_init" from the kernel line chooses openrc and also
>> leads to a hanging boot.
>>
>> So a next step seems to comment the LVs from fstab and check if it boots
>> then.
> 
> Or enable interactive boot where you can press "I" to have every service ask 
> if it needs to be started. Then skip LVM.
> Do you have parallel-boot enabled with OpenRC?

No.

> If yes, disable that just to make sure that isn't causing the issue with 
> OpenRC.
> With systemd, I would recommend checking the dependencies between mdadm and 
> lvm.

Additional scratching: even editing the grub2-line doesn't reliably
chose openrc for booting. Oh my ...

It would help to somehow set a smaller timeout for this start job to at
least get into emergency mode sooner. Right now it waits forever ("no
limit").

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 01:06:03 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 12:54, schrieb J. Roeleveld:
> > I don't use systemd, but as it's pushed by Redhat and redhat, by default,
> > uses LVM, I would expect that part to work correctly together.
> 
> Yes ... but I still don't fully get what services to enable to get it
> set up correctly. lvmetad yes or no, various systemd-generators ... to
> me it looks a bit confusing. I had a local lvm.service ... and disabled
> it already.
> 
> >> I am still trying to spot the reason. As mentioned in my other reply
> >> currently it seems to also timeout/hang with openrc for me.
> >> 
> >> Right now I downgraded systemd, rebuilt dbus, dracut, procps ...
> > 
> > lvm-utils?
> 
> No such package in portage on my thinkpad ... (using this machine to reply).

My age shows, there used to be one.
Following LVM is installed on my laptop:
===
 [I] sys-fs/lvm2
     Available versions:  2.02.97-r1 2.02.103{tbz2} ~2.02.104 ~2.02.105 
~2.02.105-r1 ~2.02.105-r2 {clvm cman device-mapper-only (+)lvm1 
lvm2create_initrd readline selinux static static-libs +thin +udev}
     Installed versions:  2.02.103{tbz2}(08:11:38 PM 03/05/2014)(lvm1 readline 
thin udev -clvm -cman -lvm2create_initrd -selinux -static -static-libs)

===

Not sure why, but lvm1 use-flag is enabled. (This is a recent fresh 
installation, must be a default somewhere)
Do you have that flag enabled?

> >> From a live-cd I am able to mount /, start the 2 raid-arrays, activate
> >> VGs and LVs ... everything useable.
> > 
> > Ok, to get this straight.
> > You have x amount of disks
> > those are in 2 raid-arrays
> > on top of that, you have LVM...
> 
> one ssd for / and /home ... 2 hdds building 2 mdadm-arrays as PVs, VGs
> on top ... LVs for data/music/VMs/etc

Sounds similar to my old server. Currently I use hardware raid on the server.
My desktops have 2 hdds in raid-0 (striping) with LVM ontop of that. With a 
small /boot partition containing kernel and an initramfs (built with 
genkernel)

> > Can you check that LVM is WAITING for the MDADM arrays to be build?
> 
> I will try to. Disabling it from a non-booting system is a bit
> difficult. Commenting the fstab-lines containing LVs didn't help to skip
> this step so far (seems there is this LVM-systemd-generator called)

Check the sequence of the services starting during boot?
(might need to disable any boot-splash if present)
Does it boot correctly when you physically unplug the drives?

> >> Removing the "real_init" from the kernel line chooses openrc and also
> >> leads to a hanging boot.
> >> 
> >> So a next step seems to comment the LVs from fstab and check if it boots
> >> then.
> > 
> > Or enable interactive boot where you can press "I" to have every service
> > ask if it needs to be started. Then skip LVM.
> > Do you have parallel-boot enabled with OpenRC?
> 
> No.
> 
> > If yes, disable that just to make sure that isn't causing the issue with
> > OpenRC.
> > With systemd, I would recommend checking the dependencies between mdadm
> > and
> > lvm.
> 
> Additional scratching: even editing the grub2-line doesn't reliably
> chose openrc for booting. Oh my ...
> 
> It would help to somehow set a smaller timeout for this start job to at
> least get into emergency mode sooner. Right now it waits forever ("no
> limit").

If you don't need the LVM disks to check the boot-config, try physically 
unplugging them. (either power or data cable should be sufficient)

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 12:12, schrieb Stefan G. Weichinger:
> 
> I see boot problems on two of my machines ... very likely related to
> lvm2 as far as I can tell so far.
> 
> Downgraded and even disabled systemd, re-emerged stuff ... I am still at
> trying to find the reason. Maybe it is with the mdadm raids also ...
> 
> I just want to ask you other gentoo users if anyone else hit this issue
> since around yesterday? Maybe someone else also scratches his head
> already ...

It seems to be related to the activation of the LVM2 logical volumes.
Even with "noauto" in fstab I see these issues.

Trying to boot with openrc leads to a timeout as well (or at least a
hanging boot).

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 12:28:03 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 12:12, schrieb Stefan G. Weichinger:
> > I see boot problems on two of my machines ... very likely related to
> > lvm2 as far as I can tell so far.
> > 
> > Downgraded and even disabled systemd, re-emerged stuff ... I am still at
> > trying to find the reason. Maybe it is with the mdadm raids also ...
> > 
> > I just want to ask you other gentoo users if anyone else hit this issue
> > since around yesterday? Maybe someone else also scratches his head
> > already ...
> 
> It seems to be related to the activation of the LVM2 logical volumes.
> Even with "noauto" in fstab I see these issues.
> 
> Trying to boot with openrc leads to a timeout as well (or at least a
> hanging boot).

Interesting.
In your other email you mentioned you can access the LVM volumes using a 
rescue-cd.
Are there any snapshots in there? Or did you do any recent changes to the LVs?
I am wondering if there might be an issue with the metadata causing issues.

Did you change any of the following packages recently:
kernel
lvm-utils
(or any of the libs used by lvm)

Did you try re-emerging lvm?

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 12:49, schrieb J. Roeleveld:

> Interesting.
> In your other email you mentioned you can access the LVM volumes using a 
> rescue-cd.
> Are there any snapshots in there? Or did you do any recent changes to the LVs?
> I am wondering if there might be an issue with the metadata causing issues.
> 
> Did you change any of the following packages recently:
> kernel
> lvm-utils
> (or any of the libs used by lvm)
> 
> Did you try re-emerging lvm?

Sure. I spent the night doing this on a second box here ... same
problems ...

Kernel 3.14.2 lately ... but when I chose an older one from grub2 it
leads to the same waiting system.

to explain a bit more:

the desktop here boots from an SSD, no RAID or LVM for the main OS.

The VGs/LVs are on mdadm-mirrors and contain data only, nothing crucial
for booting.

Many moving parts: dracut, kernel, lvm2, mdadm(?), systemd ...

I worked with exactly this box around 10hrs ago and it booted fine then.
I will check genlop what I upgraded at night ...

Got to eat now and take a rest ... I "love" to fix issues on a saturday
:-P  *sigh*

Thanks, Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 12:56:11 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 12:49, schrieb J. Roeleveld:
> > Interesting.
> > In your other email you mentioned you can access the LVM volumes using a
> > rescue-cd.
> > Are there any snapshots in there? Or did you do any recent changes to the
> > LVs? I am wondering if there might be an issue with the metadata causing
> > issues.
> > 
> > Did you change any of the following packages recently:
> > kernel
> > lvm-utils
> > (or any of the libs used by lvm)
> > 
> > Did you try re-emerging lvm?
> 
> Sure. I spent the night doing this on a second box here ... same
> problems ...
> 
> Kernel 3.14.2 lately ... but when I chose an older one from grub2 it
> leads to the same waiting system.
> 
> to explain a bit more:
> 
> the desktop here boots from an SSD, no RAID or LVM for the main OS.
> 
> The VGs/LVs are on mdadm-mirrors and contain data only, nothing crucial
> for booting.

Ok, but as it's part of the system, something is waiting for it.

> Many moving parts: dracut, kernel, lvm2, mdadm(?), systemd ...

dracut = initramfs
Did you update the initramfs recently? If out-of-sync with what is on the 
"main" environment, it could lead to issues.

> I worked with exactly this box around 10hrs ago and it booted fine then.
> I will check genlop what I upgraded at night ...

And rethink everything you did in the last 10 hours. Something caused an 
issue.
Also, check that the disks are fine and you don't have a degraded raid-
environment. The initramfs might try to have it rebuild. Which means the whole 
system can be slow.

> Got to eat now and take a rest ... I "love" to fix issues on a saturday

Enjoy your lunch, I just had mine.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 13:05, schrieb J. Roeleveld:

> dracut = initramfs
> Did you update the initramfs recently? If out-of-sync with what is on the 
> "main" environment, it could lead to issues.
> 
>> I worked with exactly this box around 10hrs ago and it booted fine then.
>> I will check genlop what I upgraded at night ...
> 
> And rethink everything you did in the last 10 hours. Something caused an 
> issue.
> Also, check that the disks are fine and you don't have a degraded raid-
> environment. The initramfs might try to have it rebuild. Which means the whole 
> system can be slow.
> 
>> Got to eat now and take a rest ... I "love" to fix issues on a saturday
> 
> Enjoy your lunch, I just had mine.

next steps:

booted into emergency mode directly.

LVs on the SSD (non-raid-based) show up fine.

No mdadm-RAIDs started yet ... when I run "systemctl start mdadm" to
pull up the arrays ... it hangs ... even looks exactly the same as in
the normal boot, 2 lines from systemd-fsck ... then blinking cursor.

Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
or maybe rebuild the initramfs (not so likely as it should boot with
older kernels/initramfs then).

Yes, lunch ..  thanks ;)

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 01:12:45 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 13:05, schrieb J. Roeleveld:
> > dracut = initramfs
> > Did you update the initramfs recently? If out-of-sync with what is on the
> > "main" environment, it could lead to issues.
> > 
> >> I worked with exactly this box around 10hrs ago and it booted fine then.
> >> I will check genlop what I upgraded at night ...
> > 
> > And rethink everything you did in the last 10 hours. Something caused an
> > issue.
> > Also, check that the disks are fine and you don't have a degraded raid-
> > environment. The initramfs might try to have it rebuild. Which means the
> > whole system can be slow.
> > 
> >> Got to eat now and take a rest ... I "love" to fix issues on a saturday
> > 
> > Enjoy your lunch, I just had mine.
> 
> next steps:
> 
> booted into emergency mode directly.
> 
> LVs on the SSD (non-raid-based) show up fine.

LVM working then.

> No mdadm-RAIDs started yet ... when I run "systemctl start mdadm" to
> pull up the arrays ... it hangs ... even looks exactly the same as in
> the normal boot, 2 lines from systemd-fsck ... then blinking cursor.

Sounds like an issue with the mdadm-raid devices.
Try setting those up WITHOUT systemctl (or init.d-scripts)

Then checking
- /proc/mdstat
- dmesg
- logging
Any of those might show an error.
Maybe ssh-in from a seperate machine to enable the raid-devices and have the 
local machine show a "tail -f" of the logs to see what happens.

> Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
> or maybe rebuild the initramfs (not so likely as it should boot with
> older kernels/initramfs then).

Which mdadm metadata format did you use when creating them?
Only 1 or 2 support auto-detection and kernel-assembly

> Yes, lunch ..  thanks ;)

13:20, late lunch :)

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 13:19, schrieb J. Roeleveld:

> Sounds like an issue with the mdadm-raid devices.
> Try setting those up WITHOUT systemctl (or init.d-scripts)
> 
> Then checking
> - /proc/mdstat
> - dmesg
> - logging
> Any of those might show an error.
> Maybe ssh-in from a seperate machine to enable the raid-devices and have the 
> local machine show a "tail -f" of the logs to see what happens.
> 
>> Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
>> or maybe rebuild the initramfs (not so likely as it should boot with
>> older kernels/initramfs then).
> 
> Which mdadm metadata format did you use when creating them?
> Only 1 or 2 support auto-detection and kernel-assembly

"mdadm -A --scan"  assembles both arrays correctly from within the
emergency mode. Both arrays non-degraded and in sync.

If I then try to change into the equivalent of runlevel 5, the system
waits and waits again ..


>> Yes, lunch ..  thanks ;)
> 
> 13:20, late lunch :)

I am stressed by the fact that my main work machine doesn't work
correctly and would prefer to have it fixed first.

But yes, I should eat now.

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 01:26:23 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 13:19, schrieb J. Roeleveld:
> > Sounds like an issue with the mdadm-raid devices.
> > Try setting those up WITHOUT systemctl (or init.d-scripts)
> > 
> > Then checking
> > - /proc/mdstat
> > - dmesg
> > - logging
> > Any of those might show an error.
> > Maybe ssh-in from a seperate machine to enable the raid-devices and have
> > the local machine show a "tail -f" of the logs to see what happens.
> > 
> >> Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
> >> or maybe rebuild the initramfs (not so likely as it should boot with
> >> older kernels/initramfs then).
> > 
> > Which mdadm metadata format did you use when creating them?
> > Only 1 or 2 support auto-detection and kernel-assembly
> 
> "mdadm -A --scan"  assembles both arrays correctly from within the
> emergency mode. Both arrays non-degraded and in sync.
> 
> If I then try to change into the equivalent of runlevel 5, the system
> waits and waits again ..

Sounds like an issue with the mdadm script/unit.
Where is Canek with his knowledge of systemd?

> >> Yes, lunch ..  thanks ;)
> > 
> > 13:20, late lunch :)
> 
> I am stressed by the fact that my main work machine doesn't work
> correctly and would prefer to have it fixed first.

I know the feeling, that's why I also have a laptop I can use. :)

> But yes, I should eat now.

Enjoy your lunch, and I'm sure it will be sorted.

--
Joost

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org> wrote:
[snip]
> Sounds like an issue with the mdadm script/unit.
> Where is Canek with his knowledge of systemd?

Canek was sleeping because it was 6:37 AM on a Saturday morning.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 15:46, schrieb Canek Peláez Valdés:
> On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org> wrote:
> [snip]
>> Sounds like an issue with the mdadm script/unit.
>> Where is Canek with his knowledge of systemd?
> 
> Canek was sleeping because it was 6:37 AM on a Saturday morning.

hehe ;-)

good morning then ... (and late congrats to your phd, btw ;-) )

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 8:48 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 03.05.2014 15:46, schrieb Canek Peláez Valdés:
>> On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org> wrote:
>> [snip]
>>> Sounds like an issue with the mdadm script/unit.
>>> Where is Canek with his knowledge of systemd?
>>
>> Canek was sleeping because it was 6:37 AM on a Saturday morning.
>
> hehe ;-)
>
> good morning then ... (and late congrats to your phd, btw ;-) )

Thanks Stefan.

Regarding your problem; as I said to Joost, I have no idea what could
possibly cause the problem, but from the experiments I conducted
trying to understand what happened with systemd+LVM2, I got the
impression that this kind of issue happens only with people having
very personalized or kinda old LVM2 setups.

I remember your fstab had the volumes listed directly. The
systemd+udev+dracut guys (and to some degree the kernel devs too, I
believe) are trying to move away from that. Could you try to set up
labels for all your mounted points and use that in fstab? It's a shot
in the dark, I don't really know if it will work.

Also, the usual requirements: boot with debug, rd.debug,
systemd.log_target=console, systemd.log_level=debug and please share
the logs.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 15:59, schrieb Canek Peláez Valdés:

> Regarding your problem; as I said to Joost, I have no idea what could
> possibly cause the problem, but from the experiments I conducted
> trying to understand what happened with systemd+LVM2, I got the
> impression that this kind of issue happens only with people having
> very personalized or kinda old LVM2 setups.

 ... kinda old ... yep ... as mentioned this box and its setup has grown
over the years.

I think it is rather straight but sure, maybe the syntax is outdated ->

> I remember your fstab had the volumes listed directly. The
> systemd+udev+dracut guys (and to some degree the kernel devs too, I
> believe) are trying to move away from that. Could you try to set up
> labels for all your mounted points and use that in fstab? It's a shot
> in the dark, I don't really know if it will work.

Hmm, ok, I will check out how to change this.

But isn't that later on in the sequence of boooting? The problem is that
the mdadm-RAIDs aren't assembled and then systemd waits for these arrays
to have the PVs ready for the LVM-VGs and LVs ...

So I suspect the initramfs more ... and the assembly of the RAIDs.


> Also, the usual requirements: boot with debug, rd.debug,
> systemd.log_target=console, systemd.log_level=debug and please share
> the logs.

Yes, thanks, I will return with more infos ...

Re: [gentoo-user] boot problems

[J. Roeleveld]

On 3 May 2014 15:48:03 CEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>Am 03.05.2014 15:46, schrieb Canek Peláez Valdés:
>> On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org>
>wrote:
>> [snip]
>>> Sounds like an issue with the mdadm script/unit.
>>> Where is Canek with his knowledge of systemd?
>> 
>> Canek was sleeping because it was 6:37 AM on a Saturday morning.
>
>hehe ;-)
>
>good morning then ... (and late congrats to your phd, btw ;-) )

Hmm...
I missed you finishing your phd. I knew you were busy with it.

Congratulations!
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 10:08 AM, J. Roeleveld <joost@antarean.org> wrote:
> On 3 May 2014 15:48:03 CEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>>Am 03.05.2014 15:46, schrieb Canek Peláez Valdés:
>>> On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org>
>>wrote:
>>> [snip]
>>>> Sounds like an issue with the mdadm script/unit.
>>>> Where is Canek with his knowledge of systemd?
>>>
>>> Canek was sleeping because it was 6:37 AM on a Saturday morning.
>>
>>hehe ;-)
>>
>>good morning then ... (and late congrats to your phd, btw ;-) )
>
> Hmm...
> I missed you finishing your phd. I knew you were busy with it.
>
> Congratulations!

Thanks.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[J. Roeleveld]

On 3 May 2014 15:46:07 CEST, "Canek Peláez Valdés" <caneko@gmail.com> wrote:
>On Sat, May 3, 2014 at 6:37 AM, J. Roeleveld <joost@antarean.org>
>wrote:
>[snip]
>> Sounds like an issue with the mdadm script/unit.
>> Where is Canek with his knowledge of systemd?
>
>Canek was sleeping because it was 6:37 AM on a Saturday morning.
>
>Regards.

I thought you might be.
Only thought of the time difference later though. But I think you occasionally have been online at that time before.

--
Joost
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 13:26, schrieb Stefan G. Weichinger:
> Am 03.05.2014 13:19, schrieb J. Roeleveld:
> 
>> Sounds like an issue with the mdadm-raid devices.
>> Try setting those up WITHOUT systemctl (or init.d-scripts)
>>
>> Then checking
>> - /proc/mdstat
>> - dmesg
>> - logging
>> Any of those might show an error.
>> Maybe ssh-in from a seperate machine to enable the raid-devices and have the 
>> local machine show a "tail -f" of the logs to see what happens.
>>
>>> Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
>>> or maybe rebuild the initramfs (not so likely as it should boot with
>>> older kernels/initramfs then).
>>
>> Which mdadm metadata format did you use when creating them?
>> Only 1 or 2 support auto-detection and kernel-assembly
> 
> "mdadm -A --scan"  assembles both arrays correctly from within the
> emergency mode. Both arrays non-degraded and in sync.
> 
> If I then try to change into the equivalent of runlevel 5, the system
> waits and waits again ..
> 
> 
>>> Yes, lunch ..  thanks ;)
>>
>> 13:20, late lunch :)
> 
> I am stressed by the fact that my main work machine doesn't work
> correctly and would prefer to have it fixed first.
> 
> But yes, I should eat now.

rebuilt mdadm, lvm2 and dracut from within emergency mode and then
recompiled kernel and let dracut rebuild the initramfs.

Booted with "rd.auto=1" and now I also get a funny start job
running/waiting for /dev/sda1 (/ on the SSD).

Oh my! :-)

pasta now.

Thanks so far, Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 01:39:06 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 13:26, schrieb Stefan G. Weichinger:
> > Am 03.05.2014 13:19, schrieb J. Roeleveld:
> >> Sounds like an issue with the mdadm-raid devices.
> >> Try setting those up WITHOUT systemctl (or init.d-scripts)
> >> 
> >> Then checking
> >> - /proc/mdstat
> >> - dmesg
> >> - logging
> >> Any of those might show an error.
> >> Maybe ssh-in from a seperate machine to enable the raid-devices and have
> >> the local machine show a "tail -f" of the logs to see what happens.
> >> 
> >>> Maybe -> dracut ... tell it to assemble the arrays via kernel-option ...
> >>> or maybe rebuild the initramfs (not so likely as it should boot with
> >>> older kernels/initramfs then).
> >> 
> >> Which mdadm metadata format did you use when creating them?
> >> Only 1 or 2 support auto-detection and kernel-assembly
> > 
> > "mdadm -A --scan"  assembles both arrays correctly from within the
> > emergency mode. Both arrays non-degraded and in sync.
> > 
> > If I then try to change into the equivalent of runlevel 5, the system
> > waits and waits again ..
> > 
> >>> Yes, lunch ..  thanks ;)
> >> 
> >> 13:20, late lunch :)
> > 
> > I am stressed by the fact that my main work machine doesn't work
> > correctly and would prefer to have it fixed first.
> > 
> > But yes, I should eat now.
> 
> rebuilt mdadm, lvm2 and dracut from within emergency mode and then
> recompiled kernel and let dracut rebuild the initramfs.
> 
> Booted with "rd.auto=1" and now I also get a funny start job
> running/waiting for /dev/sda1 (/ on the SSD).
> 
> Oh my! :-)

That's no good.
Something is not playing well.
I would check the boot-scripts/unit-files/dependencies

> pasta now.

Pasta, which kind?

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:03, schrieb J. Roeleveld:

>> Booted with "rd.auto=1" and now I also get a funny start job
>> running/waiting for /dev/sda1 (/ on the SSD).
>>
>> Oh my! :-)
> 
> That's no good.
> Something is not playing well.
> I would check the boot-scripts/unit-files/dependencies

Yep, I do .. see other reply over there (1min ago).

>> pasta now.
> 
> Pasta, which kind?

Plain spaghetti with veggie sugo ... not very inspired.

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 02:09:39 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 14:03, schrieb J. Roeleveld:
> >> Booted with "rd.auto=1" and now I also get a funny start job
> >> running/waiting for /dev/sda1 (/ on the SSD).
> >> 
> >> Oh my! :-)
> > 
> > That's no good.
> > Something is not playing well.
> > I would check the boot-scripts/unit-files/dependencies
> 
> Yep, I do .. see other reply over there (1min ago).

Saw it

> >> pasta now.
> > 
> > Pasta, which kind?
> 
> Plain spaghetti with veggie sugo ... not very inspired.

Can still be nice.
My daughter also likes plain pasta without sauce. But that must be her Italian 
side ;)

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:19, schrieb J. Roeleveld:
> On Saturday, May 03, 2014 02:09:39 PM Stefan G. Weichinger wrote:
>> Am 03.05.2014 14:03, schrieb J. Roeleveld:
>>>> Booted with "rd.auto=1" and now I also get a funny start job
>>>> running/waiting for /dev/sda1 (/ on the SSD).
>>>>
>>>> Oh my! :-)
>>>
>>> That's no good.
>>> Something is not playing well.
>>> I would check the boot-scripts/unit-files/dependencies
>>
>> Yep, I do .. see other reply over there (1min ago).
> 
> Saw it
> 
>>>> pasta now.
>>>
>>> Pasta, which kind?
>>
>> Plain spaghetti with veggie sugo ... not very inspired.
> 
> Can still be nice.
> My daughter also likes plain pasta without sauce. But that must be her Italian 
> side ;)

You know better than me ;-)
No italian roots here ... still like that food.

:-)

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 02:38:08 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 14:19, schrieb J. Roeleveld:
> > On Saturday, May 03, 2014 02:09:39 PM Stefan G. Weichinger wrote:
> >>> Pasta, which kind?
> >> 
> >> Plain spaghetti with veggie sugo ... not very inspired.
> > 
> > Can still be nice.
> > My daughter also likes plain pasta without sauce. But that must be her
> > Italian side ;)
> 
> You know better than me ;-)

Her mother knows even better.

> No italian roots here ... still like that food.

Italian food is quite diverse and can be found in a more-or-less consistent 
way in most countries.

Then again, I do have a wide taste when it comes to different foods.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 13:39, schrieb Stefan G. Weichinger:

> Booted with "rd.auto=1" and now I also get a funny start job
> running/waiting for /dev/sda1 (/ on the SSD).
> 
> Oh my! :-)
> 
> pasta now.

So, back from speed-lunch now ;-)

While cooking the pasta I got a bit further:

Box boots now with kernel 3.14.1 and with commented LVs in fstab.

When I login and check there are no mdadm-raids assembled.

When I "mdadm -A --scan" they get correctly assembled:


# cat /proc/mdstat
Personalities : [raid1]
md4 : active raid1 sdb3[0] sdc6[2]
      52395904 blocks super 1.2 [2/2] [UU]

md1 : active raid1 sdb6[0] sdc3[1]
      623963072 blocks [2/2] [UU]

(Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
historically grown somehow ...)

"vgchange -ay" then gets me all my LVs. Great so far.

So the question is, what part of the whole setup should now assemble the
arrays?

I think, dracut, right?

So I will now test booting with these funny "rd.auto" kernel line
parameters ...

Has mdadm.service to be enabled as well? Or is that redundant in a way?

Stefan

Re: [gentoo-user] boot problems

[Todd Goodman]

* Stefan G. Weichinger <lists@xunil.at> [140503 08:09]:
> Am 03.05.2014 13:39, schrieb Stefan G. Weichinger:
> 
> > Booted with "rd.auto=1" and now I also get a funny start job
> > running/waiting for /dev/sda1 (/ on the SSD).
> > 
> > Oh my! :-)
> > 
> > pasta now.
> 
> So, back from speed-lunch now ;-)
> 
> While cooking the pasta I got a bit further:
> 
> Box boots now with kernel 3.14.1 and with commented LVs in fstab.
> 
> When I login and check there are no mdadm-raids assembled.
> 
> When I "mdadm -A --scan" they get correctly assembled:
> 
> 
> # cat /proc/mdstat
> Personalities : [raid1]
> md4 : active raid1 sdb3[0] sdc6[2]
>       52395904 blocks super 1.2 [2/2] [UU]
> 
> md1 : active raid1 sdb6[0] sdc3[1]
>       623963072 blocks [2/2] [UU]
> 
> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
> historically grown somehow ...)
> 
> "vgchange -ay" then gets me all my LVs. Great so far.
> 
> So the question is, what part of the whole setup should now assemble the
> arrays?
> 
> I think, dracut, right?
> 
> So I will now test booting with these funny "rd.auto" kernel line
> parameters ...
> 
> Has mdadm.service to be enabled as well? Or is that redundant in a way?
> 
> Stefan

FWIW, I have a similar problem with mdadm and dracut and do something
similarly to what's described in:

http://rich0gentoo.wordpress.com/2012/01/21/a-quick-dracut-module/

Todd

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 08:17:37 AM Todd Goodman wrote:

<snipped>

> FWIW, I have a similar problem with mdadm and dracut and do something
> similarly to what's described in:
> 
> http://rich0gentoo.wordpress.com/2012/01/21/a-quick-dracut-module/

After more than 2 years still necessary to do it like that?

--
Joost

Re: [gentoo-user] boot problems

[Todd Goodman]

* J. Roeleveld <joost@antarean.org> [140503 08:24]:
> On Saturday, May 03, 2014 08:17:37 AM Todd Goodman wrote:
> 
> <snipped>
> 
> > FWIW, I have a similar problem with mdadm and dracut and do something
> > similarly to what's described in:
> > 
> > http://rich0gentoo.wordpress.com/2012/01/21/a-quick-dracut-module/
> 
> After more than 2 years still necessary to do it like that?
> 
> --
> Joost

I wouldn't think so, but when I set up a new server somewhat recently I
needed to do the same.

I've heard of others who don't need to do anything special so it's
likely some misconfiguration on my part.

Todd

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 15:08, schrieb Todd Goodman:

> I wouldn't think so, but when I set up a new server somewhat recently I
> needed to do the same.
> 
> I've heard of others who don't need to do anything special so it's
> likely some misconfiguration on my part.

In my case it seems to be related to the rather new and maybe buggy
releases of systemd and dracut together ....

S

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:17, schrieb Todd Goodman:
> * Stefan G. Weichinger <lists@xunil.at> [140503 08:09]:
>> Am 03.05.2014 13:39, schrieb Stefan G. Weichinger:
>>
>>> Booted with "rd.auto=1" and now I also get a funny start job
>>> running/waiting for /dev/sda1 (/ on the SSD).
>>>
>>> Oh my! :-)
>>>
>>> pasta now.
>>
>> So, back from speed-lunch now ;-)
>>
>> While cooking the pasta I got a bit further:
>>
>> Box boots now with kernel 3.14.1 and with commented LVs in fstab.
>>
>> When I login and check there are no mdadm-raids assembled.
>>
>> When I "mdadm -A --scan" they get correctly assembled:
>>
>>
>> # cat /proc/mdstat
>> Personalities : [raid1]
>> md4 : active raid1 sdb3[0] sdc6[2]
>>       52395904 blocks super 1.2 [2/2] [UU]
>>
>> md1 : active raid1 sdb6[0] sdc3[1]
>>       623963072 blocks [2/2] [UU]
>>
>> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
>> historically grown somehow ...)
>>
>> "vgchange -ay" then gets me all my LVs. Great so far.
>>
>> So the question is, what part of the whole setup should now assemble the
>> arrays?
>>
>> I think, dracut, right?
>>
>> So I will now test booting with these funny "rd.auto" kernel line
>> parameters ...
>>
>> Has mdadm.service to be enabled as well? Or is that redundant in a way?
>>
>> Stefan
> 
> FWIW, I have a similar problem with mdadm and dracut and do something
> similarly to what's described in:
> 
> http://rich0gentoo.wordpress.com/2012/01/21/a-quick-dracut-module/

Thanks for the link .... I will keep it in mind ... for now I try to
stay as close to the stuff in portage as possible (and it worked until
yesterday ;-) ).

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 02:08:26 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 13:39, schrieb Stefan G. Weichinger:
> > Booted with "rd.auto=1" and now I also get a funny start job
> > running/waiting for /dev/sda1 (/ on the SSD).
> > 
> > Oh my! :-)
> > 
> > pasta now.
> 
> So, back from speed-lunch now ;-)
> 
> While cooking the pasta I got a bit further:
> 
> Box boots now with kernel 3.14.1 and with commented LVs in fstab.
> 
> When I login and check there are no mdadm-raids assembled.
> 
> When I "mdadm -A --scan" they get correctly assembled:
> 
> 
> # cat /proc/mdstat
> Personalities : [raid1]
> md4 : active raid1 sdb3[0] sdc6[2]
>       52395904 blocks super 1.2 [2/2] [UU]
> 
> md1 : active raid1 sdb6[0] sdc3[1]
>       623963072 blocks [2/2] [UU]
> 
> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
> historically grown somehow ...)

Seen stranger, it works.
What do the logs show when enabling them? Any "odd" messages?

> "vgchange -ay" then gets me all my LVs. Great so far.
> 
> So the question is, what part of the whole setup should now assemble the
> arrays?
> 
> I think, dracut, right?

Yes, something there.

> So I will now test booting with these funny "rd.auto" kernel line
> parameters ...

With genkernel, I have a lvm and mdadm boot parameter.
With dracut, that might also be necessary.

I would prefer the initramfs to do it automagically without extra parameters.

> Has mdadm.service to be enabled as well? Or is that redundant in a way?

That's systemd-specific. Don't know.

If dracut sorts that, mdadm.service might be too much.
But, as those are not boot-critical, they might wait for the real system to be 
running.
In that case, you do need mdadm.service as well.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:17, schrieb J. Roeleveld:

>> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
>> historically grown somehow ...)
> 
> Seen stranger, it works.
> What do the logs show when enabling them? Any "odd" messages?

Just looked it up, nothing odd. To me it seems that the assembling just
doesn't happen at all.

Maybe really take the step with the dracut module or a local separate
systemd unit file to do that ....

> With genkernel, I have a lvm and mdadm boot parameter.
> With dracut, that might also be necessary.

Doesn't work yet.


> If dracut sorts that, mdadm.service might be too much.
> But, as those are not boot-critical, they might wait for the real system to be 
> running.
> In that case, you do need mdadm.service as well.

rd.auto=1 alone does not do the trick so far.

It's quite hard to keep you informed what steps I take here ...

dozens of reboots now ...

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 02:43:39 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 14:17, schrieb J. Roeleveld:
> >> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
> >> historically grown somehow ...)
> > 
> > Seen stranger, it works.
> > What do the logs show when enabling them? Any "odd" messages?
> 
> Just looked it up, nothing odd. To me it seems that the assembling just
> doesn't happen at all.
> 
> Maybe really take the step with the dracut module or a local separate
> systemd unit file to do that ....
> 
> > With genkernel, I have a lvm and mdadm boot parameter.
> > With dracut, that might also be necessary.
> 
> Doesn't work yet.
> 
> > If dracut sorts that, mdadm.service might be too much.
> > But, as those are not boot-critical, they might wait for the real system
> > to be running.
> > In that case, you do need mdadm.service as well.
> 
> rd.auto=1 alone does not do the trick so far.
> 
> It's quite hard to keep you informed what steps I take here ...
> 
> dozens of reboots now ...

If you boot without those in fstab, then do the mdadm and lvm steps manually, 
do you then have it " working "?
If yes, I would leave it for now and wait for Canek to come online in a few 
hours. (Mexico should be awake at around 4 or 5pm I think)

He knows systemd and dracut better then most people on here.

We are now getting into systemd and dracut territory and apart from trying 
different tools, I have no further ideas.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:52, schrieb J. Roeleveld:

> If you boot without those in fstab, then do the mdadm and lvm steps manually, 
> do you then have it " working "?

yes ... everything up and accessible then.

> If yes, I would leave it for now and wait for Canek to come online in a few 
> hours. (Mexico should be awake at around 4 or 5pm I think)

ok ... yes.

> He knows systemd and dracut better then most people on here.
> 
> We are now getting into systemd and dracut territory and apart from trying 
> different tools, I have no further ideas.

Thanks so far, anyway.

I will try some things until then.

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 7:52 AM, J. Roeleveld <joost@antarean.org> wrote:
[...]
> If you boot without those in fstab, then do the mdadm and lvm steps manually,
> do you then have it " working "?
> If yes, I would leave it for now and wait for Canek to come online in a few
> hours. (Mexico should be awake at around 4 or 5pm I think)
>
> He knows systemd and dracut better then most people on here.

Thanks for the confidence, but I don't think I will be able to help
very much: I'm still running systemd 208. I defended my PhD thesis
last week, and I *really* wanted to avoid a situation similar to the
one Stefan is having, so I haven't updated any of my systems on weeks.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 15:50, schrieb Canek Peláez Valdés:
> On Sat, May 3, 2014 at 7:52 AM, J. Roeleveld <joost@antarean.org> wrote:
> [...]
>> If you boot without those in fstab, then do the mdadm and lvm steps manually,
>> do you then have it " working "?
>> If yes, I would leave it for now and wait for Canek to come online in a few
>> hours. (Mexico should be awake at around 4 or 5pm I think)
>>
>> He knows systemd and dracut better then most people on here.
> 
> Thanks for the confidence, but I don't think I will be able to help
> very much: I'm still running systemd 208. I defended my PhD thesis
> last week, and I *really* wanted to avoid a situation similar to the
> one Stefan is having, so I haven't updated any of my systems on weeks.

Yep. Clever decision ;-)

I could downgrade to 208 as well, this mainly only removes
systemd-networkd from my box which was nice to have .. but anyway.

systemd-208 would also allow me to downgrade dracut, and I assume these
2 steps would definitely change something.

Right now I do *backups* and take a break ... I currently know how to
get the box up and working manually so this somehow takes the bigger
stress out of the game. Resetting the box dozens of times does thrill me
a bit with RAIDs and stuff.

I will keep the list informed if I find out something new.

Thanks all, Stefan

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 15:57, schrieb Stefan G. Weichinger:
> Am 03.05.2014 15:50, schrieb Canek Peláez Valdés:
>> On Sat, May 3, 2014 at 7:52 AM, J. Roeleveld <joost@antarean.org> wrote:
>> [...]
>>> If you boot without those in fstab, then do the mdadm and lvm steps manually,
>>> do you then have it " working "?
>>> If yes, I would leave it for now and wait for Canek to come online in a few
>>> hours. (Mexico should be awake at around 4 or 5pm I think)
>>>
>>> He knows systemd and dracut better then most people on here.
>>
>> Thanks for the confidence, but I don't think I will be able to help
>> very much: I'm still running systemd 208. I defended my PhD thesis
>> last week, and I *really* wanted to avoid a situation similar to the
>> one Stefan is having, so I haven't updated any of my systems on weeks.
> 
> Yep. Clever decision ;-)
> 
> I could downgrade to 208 as well, this mainly only removes
> systemd-networkd from my box which was nice to have .. but anyway.
> 
> systemd-208 would also allow me to downgrade dracut, and I assume these
> 2 steps would definitely change something.
> 
> Right now I do *backups* and take a break ... I currently know how to
> get the box up and working manually so this somehow takes the bigger
> stress out of the game. Resetting the box dozens of times does thrill me
> a bit with RAIDs and stuff.
> 
> I will keep the list informed if I find out something new.

back on systemd 208-r3

I somehow get lost more and more ...

fstab with uuids and labels only now

Dracut wants to resume from the swap-partition somehow ... parameter
"noresume" helps

Grub2 still has "root=/dev/sda" in its line and I don't get where that
comes from

Canek, I use your kerninst, could you point out how to edit
kerninst.conf to get grub2 read something like "root=UUID=..." or so?

Oh my, what a saturday.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 18:40, schrieb Stefan G. Weichinger:
> Am 03.05.2014 15:57, schrieb Stefan G. Weichinger:
>> Am 03.05.2014 15:50, schrieb Canek Peláez Valdés:
>>> On Sat, May 3, 2014 at 7:52 AM, J. Roeleveld <joost@antarean.org> wrote:
>>> [...]
>>>> If you boot without those in fstab, then do the mdadm and lvm steps manually,
>>>> do you then have it " working "?
>>>> If yes, I would leave it for now and wait for Canek to come online in a few
>>>> hours. (Mexico should be awake at around 4 or 5pm I think)
>>>>
>>>> He knows systemd and dracut better then most people on here.
>>>
>>> Thanks for the confidence, but I don't think I will be able to help
>>> very much: I'm still running systemd 208. I defended my PhD thesis
>>> last week, and I *really* wanted to avoid a situation similar to the
>>> one Stefan is having, so I haven't updated any of my systems on weeks.
>>
>> Yep. Clever decision ;-)
>>
>> I could downgrade to 208 as well, this mainly only removes
>> systemd-networkd from my box which was nice to have .. but anyway.
>>
>> systemd-208 would also allow me to downgrade dracut, and I assume these
>> 2 steps would definitely change something.
>>
>> Right now I do *backups* and take a break ... I currently know how to
>> get the box up and working manually so this somehow takes the bigger
>> stress out of the game. Resetting the box dozens of times does thrill me
>> a bit with RAIDs and stuff.
>>
>> I will keep the list informed if I find out something new.
> 
> back on systemd 208-r3
> 
> I somehow get lost more and more ...
> 
> fstab with uuids and labels only now
> 
> Dracut wants to resume from the swap-partition somehow ... parameter
> "noresume" helps
> 
> Grub2 still has "root=/dev/sda" in its line and I don't get where that
> comes from

correction, it was "sda1" and my fault in /etc/default/grub (hiding UUIDs)

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 18:51, schrieb Stefan G. Weichinger:
> Am 03.05.2014 18:40, schrieb Stefan G. Weichinger:
>> Am 03.05.2014 15:57, schrieb Stefan G. Weichinger:
>>> Am 03.05.2014 15:50, schrieb Canek Peláez Valdés:
>>>> On Sat, May 3, 2014 at 7:52 AM, J. Roeleveld <joost@antarean.org> wrote:
>>>> [...]
>>>>> If you boot without those in fstab, then do the mdadm and lvm steps manually,
>>>>> do you then have it " working "?
>>>>> If yes, I would leave it for now and wait for Canek to come online in a few
>>>>> hours. (Mexico should be awake at around 4 or 5pm I think)
>>>>>
>>>>> He knows systemd and dracut better then most people on here.
>>>>
>>>> Thanks for the confidence, but I don't think I will be able to help
>>>> very much: I'm still running systemd 208. I defended my PhD thesis
>>>> last week, and I *really* wanted to avoid a situation similar to the
>>>> one Stefan is having, so I haven't updated any of my systems on weeks.
>>>
>>> Yep. Clever decision ;-)
>>>
>>> I could downgrade to 208 as well, this mainly only removes
>>> systemd-networkd from my box which was nice to have .. but anyway.
>>>
>>> systemd-208 would also allow me to downgrade dracut, and I assume these
>>> 2 steps would definitely change something.
>>>
>>> Right now I do *backups* and take a break ... I currently know how to
>>> get the box up and working manually so this somehow takes the bigger
>>> stress out of the game. Resetting the box dozens of times does thrill me
>>> a bit with RAIDs and stuff.
>>>
>>> I will keep the list informed if I find out something new.
>>
>> back on systemd 208-r3
>>
>> I somehow get lost more and more ...
>>
>> fstab with uuids and labels only now
>>
>> Dracut wants to resume from the swap-partition somehow ... parameter
>> "noresume" helps
>>
>> Grub2 still has "root=/dev/sda" in its line and I don't get where that
>> comes from
> 
> correction, it was "sda1" and my fault in /etc/default/grub (hiding UUIDs)

An older kernel (3.13.4) boots OK ... the newer ones are screwed up
(their initramfs) ...

Back to the start.

As I don't boot from LVM or mdadm-raid I consider to not need any
"DRACUT_MODULES=" in /etc/portage/make.conf, right?

What do you guys have in /etc/dracut.conf ?

Canek, I use your kerninst-tool ...

And there are settings in /etc/default/grub as well ... I somehow get
the feeling that there are way too much bells and whistles in the game.

I want to clean up as much as possible so that this more or less wasted
day at least makes some sense in the end.

Thanks all, Stefan

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 12:00 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
[snip]
> An older kernel (3.13.4) boots OK ... the newer ones are screwed up
> (their initramfs) ...
>
> Back to the start.
>
> As I don't boot from LVM or mdadm-raid I consider to not need any
> "DRACUT_MODULES=" in /etc/portage/make.conf, right?
>
> What do you guys have in /etc/dracut.conf ?
>
> Canek, I use your kerninst-tool ...
>
> And there are settings in /etc/default/grub as well ... I somehow get
> the feeling that there are way too much bells and whistles in the game.
>
> I want to clean up as much as possible so that this more or less wasted
> day at least makes some sense in the end.

I'll give you my confs in my LVM2 virtual machine; my normal machines
don't use LVM2, so I don't think they'll help in any way.

Also, if my normal machines are lagging on updates, the virtual ones
are even more behind: my LVM vm has systemd 204!

I'm removing any comments:

/etc/default/grub.conf:

GRUB_DISTRIBUTOR="Gentoo"
GRUB_DEFAULT=0
GRUB_HIDDEN_TIMEOUT=0
GRUB_HIDDEN_TIMEOUT_QUIET=true
GRUB_TIMEOUT=10
GRUB_CMDLINE_LINUX="init=/usr/lib/systemd/systemd quiet nosplash"
GRUB_PRELOAD_MODULES="lvm mdraid1x"

Notice: no root, no real_root, nothing.

/etc/dracut.conf:

add_dracutmodules+="crypt lvm mdraid systemd"
add_drivers+="autofs4 ipv6 dm-crypt aes sha256"
fscks="umount mount /sbin/fsck* e2fsck"

/etc/fstab:

LABEL=Root / ext4 acl,noatime 0 1
LABEL=User /usr ext4 acl,noatime 0 2
LABEL=Boot /boot ext2 acl,noatime 0 2
LABEL=Home /home ext4 acl,noatime 0 2
/dev/mapper/swap none swap sw 0 0
shm /dev/shm tmpfs nodev,nosuid,noexec 0 0
tmpfs /tmp tmpfs defaults,nosuid,size=100% 0 0

/etc/crypttab:

home /dev/vg/vol5
swap /dev/vg/vol2 /dev/urandom swap

I didn't touched /etc/lvm/lvm.conf, nor /etc/mdadm.conf.

My drive topology:

NAME                 MAJ:MIN RM  SIZE RO TYPE  MOUNTPOINT
sr0                   11:0    1 1024M  0 rom
vda                  253:0    0    5G  0 disk
`-vda1               253:1    0    5G  0 part
  `-md127              9:127  0   15G  0 raid5
    |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
    |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
    |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
    |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
    | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
    `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
      `-home (dm-6)  254:6    0  2.8G  0 crypt /home
vdb                  253:16   0    5G  0 disk
`-vdb1               253:17   0    5G  0 part
  `-md127              9:127  0   15G  0 raid5
    |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
    |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
    |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
    |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
    | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
    `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
      `-home (dm-6)  254:6    0  2.8G  0 crypt /home
vdc                  253:32   0    5G  0 disk
`-vdc1               253:33   0    5G  0 part
  `-md127              9:127  0   15G  0 raid5
    |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
    |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
    |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
    |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
    | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
    `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
      `-home (dm-6)  254:6    0  2.8G  0 crypt /home
vdd                  253:48   0    5G  0 disk
`-vdd1               253:49   0    5G  0 part
  `-md127              9:127  0   15G  0 raid5
    |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
    |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
    |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
    |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
    | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
    `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
      `-home (dm-6)  254:6    0  2.8G  0 crypt /home

Do you want to see any other config?

Let me try to update this test VM to see if I can hit the same issue as you.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 19:25, schrieb Canek Peláez Valdés:

> I'll give you my confs in my LVM2 virtual machine; my normal machines
> don't use LVM2, so I don't think they'll help in any way.
> 
> Also, if my normal machines are lagging on updates, the virtual ones
> are even more behind: my LVM vm has systemd 204!

;-)

> I'm removing any comments:
> 
> /etc/default/grub.conf:
> 
> GRUB_DISTRIBUTOR="Gentoo"
> GRUB_DEFAULT=0
> GRUB_HIDDEN_TIMEOUT=0
> GRUB_HIDDEN_TIMEOUT_QUIET=true
> GRUB_TIMEOUT=10
> GRUB_CMDLINE_LINUX="init=/usr/lib/systemd/systemd quiet nosplash"
> GRUB_PRELOAD_MODULES="lvm mdraid1x"
> 
> Notice: no root, no real_root, nothing.

Yes, and I have to say, it should be that way!
I don't like the fact that all these configs get more and more
complicated when on the other hand technology should make things easier ...

sure, more complicated use cases on one hand ... but in the end I want
to boot from one plain /dev/sda1 ... mdadm and lvm are *optional* in my
case right now.

> /etc/dracut.conf:
> 
> add_dracutmodules+="crypt lvm mdraid systemd"
> add_drivers+="autofs4 ipv6 dm-crypt aes sha256"
> fscks="umount mount /sbin/fsck* e2fsck"
> 
> /etc/fstab:
> 
> LABEL=Root / ext4 acl,noatime 0 1
> LABEL=User /usr ext4 acl,noatime 0 2
> LABEL=Boot /boot ext2 acl,noatime 0 2
> LABEL=Home /home ext4 acl,noatime 0 2
> /dev/mapper/swap none swap sw 0 0
> shm /dev/shm tmpfs nodev,nosuid,noexec 0 0
> tmpfs /tmp tmpfs defaults,nosuid,size=100% 0 0
> 
> /etc/crypttab:
> 
> home /dev/vg/vol5
> swap /dev/vg/vol2 /dev/urandom swap
> 
> I didn't touched /etc/lvm/lvm.conf, nor /etc/mdadm.conf.
> 
> My drive topology:
> 
> NAME                 MAJ:MIN RM  SIZE RO TYPE  MOUNTPOINT
> sr0                   11:0    1 1024M  0 rom
> vda                  253:0    0    5G  0 disk
> `-vda1               253:1    0    5G  0 part
>   `-md127              9:127  0   15G  0 raid5
>     |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
>     |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
>     |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
>     |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
>     | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
>     `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
>       `-home (dm-6)  254:6    0  2.8G  0 crypt /home
> vdb                  253:16   0    5G  0 disk
> `-vdb1               253:17   0    5G  0 part
>   `-md127              9:127  0   15G  0 raid5
>     |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
>     |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
>     |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
>     |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
>     | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
>     `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
>       `-home (dm-6)  254:6    0  2.8G  0 crypt /home
> vdc                  253:32   0    5G  0 disk
> `-vdc1               253:33   0    5G  0 part
>   `-md127              9:127  0   15G  0 raid5
>     |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
>     |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
>     |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
>     |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
>     | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
>     `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
>       `-home (dm-6)  254:6    0  2.8G  0 crypt /home
> vdd                  253:48   0    5G  0 disk
> `-vdd1               253:49   0    5G  0 part
>   `-md127              9:127  0   15G  0 raid5
>     |-vg-vol1 (dm-0) 254:0    0  200M  0 lvm   /boot
>     |-vg-vol3 (dm-1) 254:1    0    2G  0 lvm   /
>     |-vg-vol4 (dm-2) 254:2    0    8G  0 lvm   /usr
>     |-vg-vol2 (dm-3) 254:3    0    2G  0 lvm
>     | `-swap (dm-5)  254:5    0    2G  0 crypt [SWAP]
>     `-vg-vol5 (dm-4) 254:4    0  2.8G  0 lvm
>       `-home (dm-6)  254:6    0  2.8G  0 crypt /home
> 
> Do you want to see any other config?
> 
> Let me try to update this test VM to see if I can hit the same issue as you.
> 
> Regards.
> 


Sure, thanks so far. I am taking another break now and return hopefully
regenerated ;-)

S

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 19:25, schrieb Canek Peláez Valdés:

> Do you want to see any other config?

The version numbers would be interesting:

grub
dracut
lvm2
systemd
...

Thanks! I configured my machine similar to yours and try the next booting ..

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 1:30 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 03.05.2014 19:25, schrieb Canek Peláez Valdés:
>
>> Do you want to see any other config?
>
> The version numbers would be interesting:
>
> grub
> dracut
> lvm2
> systemd

sys-boot/grub-2.00_p5107-r1:2
sys-kernel/dracut-033:0
sys-apps/systemd-204:0
sys-fs/lvm2-2.02.100-r2:0

I'm updating my infrastructure, but it will take a while...

> Thanks! I configured my machine similar to yours and try the next booting ..

No problem.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 20:33, schrieb Canek Peláez Valdés:

> sys-boot/grub-2.00_p5107-r1:2
> sys-kernel/dracut-033:0
> sys-apps/systemd-204:0
> sys-fs/lvm2-2.02.100-r2:0

great, thanks

I am able to boot a kernel 3.13.4 with its initramfs which was built
back then with dracut-034-r1. No mdadm-raids assembled as well, but at
least it finds sda1 ;-)

my fstab now also uses LABELs as you suggested.

I now rebuild a kernel 3.13.11 (as I don't have any sources for the
older one I can't build nvidia-drivers to run Gnome on it) with some of
your dracut-changes and dracut-036-r4 (downgraded from 037).

> I'm updating my infrastructure, but it will take a while...

Take your time ... thanks !

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 08:37:49 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 20:33, schrieb Canek Peláez Valdés:
> > sys-boot/grub-2.00_p5107-r1:2
> > sys-kernel/dracut-033:0
> > sys-apps/systemd-204:0
> > sys-fs/lvm2-2.02.100-r2:0
> 
> great, thanks
> 
> I am able to boot a kernel 3.13.4 with its initramfs which was built
> back then with dracut-034-r1. No mdadm-raids assembled as well, but at
> least it finds sda1 ;-)
> 
> my fstab now also uses LABELs as you suggested.
> 
> I now rebuild a kernel 3.13.11 (as I don't have any sources for the
> older one I can't build nvidia-drivers to run Gnome on it) with some of
> your dracut-changes and dracut-036-r4 (downgraded from 037).

Which one do you need:
shell ~ # ls /opt/distfiles/linux-*
/opt/distfiles/linux-2.6.18.tar.bz2 
 /opt/distfiles/linux-2.6.39.tar.bz2 
 /opt/distfiles/linux-3.12.tar.xz  
/opt/distfiles/linux-3.3.tar.bz2  
/opt/distfiles/linux-3.6.tar.bz2  
/opt/distfiles/linux-3.8.tar.xz
/opt/distfiles/linux-2.6.36.tar.bz2  
/opt/distfiles/linux-3.0.tar.bz2     
/opt/distfiles/linux-3.1.tar.bz2  
/opt/distfiles/linux-3.4.tar.bz2  
/opt/distfiles/linux-3.7.tar.bz2
/opt/distfiles/linux-2.6.38.tar.bz2  
/opt/distfiles/linux-3.10.tar.xz     
/opt/distfiles/linux-3.2.tar.bz2  
/opt/distfiles/linux-3.5.tar.bz2  
/opt/distfiles/linux-3.7.tar.xz

I always use gentoo-sources, so I will also have the gentoo-patches for those.

I might also still have the packages around as I build packages on a seperate 
machine regularly before updating the actual machines.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 20:47, schrieb J. Roeleveld:
> On Saturday, May 03, 2014 08:37:49 PM Stefan G. Weichinger wrote:
>> Am 03.05.2014 20:33, schrieb Canek Peláez Valdés:
>>> sys-boot/grub-2.00_p5107-r1:2
>>> sys-kernel/dracut-033:0
>>> sys-apps/systemd-204:0
>>> sys-fs/lvm2-2.02.100-r2:0
>>
>> great, thanks
>>
>> I am able to boot a kernel 3.13.4 with its initramfs which was built
>> back then with dracut-034-r1. No mdadm-raids assembled as well, but at
>> least it finds sda1 ;-)
>>
>> my fstab now also uses LABELs as you suggested.
>>
>> I now rebuild a kernel 3.13.11 (as I don't have any sources for the
>> older one I can't build nvidia-drivers to run Gnome on it) with some of
>> your dracut-changes and dracut-036-r4 (downgraded from 037).
> 
> Which one do you need:
> shell ~ # ls /opt/distfiles/linux-*
> /opt/distfiles/linux-2.6.18.tar.bz2 
>  /opt/distfiles/linux-2.6.39.tar.bz2 
>  /opt/distfiles/linux-3.12.tar.xz  
> /opt/distfiles/linux-3.3.tar.bz2  
> /opt/distfiles/linux-3.6.tar.bz2  
> /opt/distfiles/linux-3.8.tar.xz
> /opt/distfiles/linux-2.6.36.tar.bz2  
> /opt/distfiles/linux-3.0.tar.bz2     
> /opt/distfiles/linux-3.1.tar.bz2  
> /opt/distfiles/linux-3.4.tar.bz2  
> /opt/distfiles/linux-3.7.tar.bz2
> /opt/distfiles/linux-2.6.38.tar.bz2  
> /opt/distfiles/linux-3.10.tar.xz     
> /opt/distfiles/linux-3.2.tar.bz2  
> /opt/distfiles/linux-3.5.tar.bz2  
> /opt/distfiles/linux-3.7.tar.xz
> 
> I always use gentoo-sources, so I will also have the gentoo-patches for those.
> 
> I might also still have the packages around as I build packages on a seperate 
> machine regularly before updating the actual machines.

Thanks for the offer, I am not yet there ;-)

new infos ... new and fixed ebuilds for systemd in portage right now:

http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/sys-apps/systemd/ChangeLog?view=markup

and a shiny new bug for dracut as well:

https://bugs.gentoo.org/show_bug.cgi?id=509492

So even more moving parts for me no :-P

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 08:51:14 PM Stefan G. Weichinger wrote:
> Am 03.05.2014 20:47, schrieb J. Roeleveld:
> > On Saturday, May 03, 2014 08:37:49 PM Stefan G. Weichinger wrote:
> >> Am 03.05.2014 20:33, schrieb Canek Peláez Valdés:
> >>> sys-boot/grub-2.00_p5107-r1:2
> >>> sys-kernel/dracut-033:0
> >>> sys-apps/systemd-204:0
> >>> sys-fs/lvm2-2.02.100-r2:0
> >> 
> >> great, thanks
> >> 
> >> I am able to boot a kernel 3.13.4 with its initramfs which was built
> >> back then with dracut-034-r1. No mdadm-raids assembled as well, but at
> >> least it finds sda1 ;-)
> >> 
> >> my fstab now also uses LABELs as you suggested.
> >> 
> >> I now rebuild a kernel 3.13.11 (as I don't have any sources for the
> >> older one I can't build nvidia-drivers to run Gnome on it) with some of
> >> your dracut-changes and dracut-036-r4 (downgraded from 037).
> > 
> > Which one do you need:
> > shell ~ # ls /opt/distfiles/linux-*
> > /opt/distfiles/linux-2.6.18.tar.bz2
> > 
> >  /opt/distfiles/linux-2.6.39.tar.bz2
> >  /opt/distfiles/linux-3.12.tar.xz
> > 
> > /opt/distfiles/linux-3.3.tar.bz2
> > /opt/distfiles/linux-3.6.tar.bz2
> > /opt/distfiles/linux-3.8.tar.xz
> > /opt/distfiles/linux-2.6.36.tar.bz2
> > /opt/distfiles/linux-3.0.tar.bz2
> > /opt/distfiles/linux-3.1.tar.bz2
> > /opt/distfiles/linux-3.4.tar.bz2
> > /opt/distfiles/linux-3.7.tar.bz2
> > /opt/distfiles/linux-2.6.38.tar.bz2
> > /opt/distfiles/linux-3.10.tar.xz
> > /opt/distfiles/linux-3.2.tar.bz2
> > /opt/distfiles/linux-3.5.tar.bz2
> > /opt/distfiles/linux-3.7.tar.xz
> > 
> > I always use gentoo-sources, so I will also have the gentoo-patches for
> > those.
> > 
> > I might also still have the packages around as I build packages on a
> > seperate machine regularly before updating the actual machines.
> 
> Thanks for the offer, I am not yet there ;-)

Ok. I hardly ever delete anything these days. So those are likely to remain 
available in their current form for the foreseeable future.

> new infos ... new and fixed ebuilds for systemd in portage right now:
> 
> http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/sys-apps/systemd/Cha
> ngeLog?view=markup
> 
> and a shiny new bug for dracut as well:
> 
> https://bugs.gentoo.org/show_bug.cgi?id=509492
> 
> So even more moving parts for me no :-P

That bug sounds interesting.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 20:56, schrieb J. Roeleveld:

>> Thanks for the offer, I am not yet there ;-)
> 
> Ok. I hardly ever delete anything these days. So those are likely to remain 
> available in their current form for the foreseeable future.

My current goal is to get a working 3.13.11 with systemd and dracut.
From there on to 3.14.x ....


>> and a shiny new bug for dracut as well:
>>
>> https://bugs.gentoo.org/show_bug.cgi?id=509492
>>
>> So even more moving parts for me no :-P
> 
> That bug sounds interesting.

It might fix the lvm-issues, yes. But for the last 2 hrs my booting in
general is messed up ... ;-)

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 1:51 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
[snip]
> and a shiny new bug for dracut as well:
>
> https://bugs.gentoo.org/show_bug.cgi?id=509492

That looks like a reasonable explanation. But it could be something else.

> So even more moving parts for me no :-P

In my stable machines I only have two keyworded packages:

sys-kernel/dracut
sys-kernel/vanilla-sources

The vanilla-sources package will always be unstable, apparently, so
nothing to do there. However, I whish dracut got stabilized at some
point.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
[...]
> Let me try to update this test VM to see if I can hit the same issue as you.

Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB

This is going to take a while...

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 01:30:43 PM Canek Peláez Valdés wrote:
> On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com>
> wrote: [...]
> 
> > Let me try to update this test VM to see if I can hit the same issue as
> > you.
> Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
> 9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB

Downgrade?
Which one is that?

> This is going to take a while...

Always, especially that download if you're unlucky with the mirrors.

--
Joost

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 1:35 PM, J. Roeleveld <joost@antarean.org> wrote:
> On Saturday, May 03, 2014 01:30:43 PM Canek Peláez Valdés wrote:
>> On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com>
>> wrote: [...]
>>
>> > Let me try to update this test VM to see if I can hit the same issue as
>> > you.
>> Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
>> 9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB
>
> Downgrade?
> Which one is that?

media-libs/exiftool-9.12, from 9.120. I suppose a change in version numbering?

>> This is going to take a while...
>
> Always, especially that download if you're unlucky with the mirrors.

The download is over, I'm now compiling.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 21:33, schrieb Canek Peláez Valdés:
> On Sat, May 3, 2014 at 1:35 PM, J. Roeleveld <joost@antarean.org> wrote:
>> On Saturday, May 03, 2014 01:30:43 PM Canek Peláez Valdés wrote:
>>> On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com>
>>> wrote: [...]
>>>
>>>> Let me try to update this test VM to see if I can hit the same issue as
>>>> you.
>>> Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
>>> 9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB
>>
>> Downgrade?
>> Which one is that?
> 
> media-libs/exiftool-9.12, from 9.120. I suppose a change in version numbering?
> 
>>> This is going to take a while...
>>
>> Always, especially that download if you're unlucky with the mirrors.
> 
> The download is over, I'm now compiling.

I don't know which sub-thread to follow / reply to, so I take it
chronologically.

A bit of success here:

booted 3.13.11 with latest systemd and dracut.

I had to manually "rm
/usr/lib/systemd/system-generators/lvm2-activation-generator" to get
that done (and comment the LVs in fstab).

I now see an assembled mdadm raid after booting, but no activated LVs.

So lvm2 smells ...

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 2:38 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 03.05.2014 21:33, schrieb Canek Peláez Valdés:
>> On Sat, May 3, 2014 at 1:35 PM, J. Roeleveld <joost@antarean.org> wrote:
>>> On Saturday, May 03, 2014 01:30:43 PM Canek Peláez Valdés wrote:
>>>> On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com>
>>>> wrote: [...]
>>>>
>>>>> Let me try to update this test VM to see if I can hit the same issue as
>>>>> you.
>>>> Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
>>>> 9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB
>>>
>>> Downgrade?
>>> Which one is that?
>>
>> media-libs/exiftool-9.12, from 9.120. I suppose a change in version numbering?
>>
>>>> This is going to take a while...
>>>
>>> Always, especially that download if you're unlucky with the mirrors.
>>
>> The download is over, I'm now compiling.
>
> I don't know which sub-thread to follow / reply to, so I take it
> chronologically.
>
> A bit of success here:
>
> booted 3.13.11 with latest systemd and dracut.
>
> I had to manually "rm
> /usr/lib/systemd/system-generators/lvm2-activation-generator" to get
> that done (and comment the LVs in fstab).

I don't think deleting any generator is a good idea...

> I now see an assembled mdadm raid after booting, but no activated LVs.
>
> So lvm2 smells ...

As per the bug you posted, can you show us the contents of your
initramfs, to check that the LVM binaries are included? And confirm
that the binaries are the same as in your normal system?

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 21:41, schrieb Canek Peláez Valdés:
> On Sat, May 3, 2014 at 2:38 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
>> Am 03.05.2014 21:33, schrieb Canek Peláez Valdés:
>>> On Sat, May 3, 2014 at 1:35 PM, J. Roeleveld <joost@antarean.org> wrote:
>>>> On Saturday, May 03, 2014 01:30:43 PM Canek Peláez Valdés wrote:
>>>>> On Sat, May 3, 2014 at 12:25 PM, Canek Peláez Valdés <caneko@gmail.com>
>>>>> wrote: [...]
>>>>>
>>>>>> Let me try to update this test VM to see if I can hit the same issue as
>>>>>> you.
>>>>> Total: 124 packages (107 upgrades, 1 downgrade, 5 new, 2 in new slots,
>>>>> 9 reinstalls, 1 uninstall), Size of downloads: 663,051 kB
>>>>
>>>> Downgrade?
>>>> Which one is that?
>>>
>>> media-libs/exiftool-9.12, from 9.120. I suppose a change in version numbering?
>>>
>>>>> This is going to take a while...
>>>>
>>>> Always, especially that download if you're unlucky with the mirrors.
>>>
>>> The download is over, I'm now compiling.
>>
>> I don't know which sub-thread to follow / reply to, so I take it
>> chronologically.
>>
>> A bit of success here:
>>
>> booted 3.13.11 with latest systemd and dracut.
>>
>> I had to manually "rm
>> /usr/lib/systemd/system-generators/lvm2-activation-generator" to get
>> that done (and comment the LVs in fstab).
> 
> I don't think deleting any generator is a good idea...
> 
>> I now see an assembled mdadm raid after booting, but no activated LVs.
>>
>> So lvm2 smells ...
> 
> As per the bug you posted, can you show us the contents of your
> initramfs, to check that the LVM binaries are included? And confirm
> that the binaries are the same as in your normal system?

Hm, yes ... I was a bit faster and already applied that patch ... and
had a successful boot with activated and mounted LVs ... currently
rebuilding some gdm/gnome-shell .... etc

I will reboot without all the debug-stuff for testing ... and then maybe
remove the patch and test 3.14.x

... and I will get a beer now. This is gentoo-testers, not gentoo-users
here :-P

S

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 21:53, schrieb Stefan G. Weichinger:

>> As per the bug you posted, can you show us the contents of your
>> initramfs, to check that the LVM binaries are included? And confirm
>> that the binaries are the same as in your normal system?
> 
> Hm, yes ... I was a bit faster and already applied that patch ... and
> had a successful boot with activated and mounted LVs ... currently
> rebuilding some gdm/gnome-shell .... etc
> 
> I will reboot without all the debug-stuff for testing ... and then maybe
> remove the patch and test 3.14.x
> 
> ... and I will get a beer now. This is gentoo-testers, not gentoo-users
> here :-P

Back again, booted successfully ... with latest gentoo-sources, systemd,
dracut, lvm2 ... and the mentioned patch.

If you check the link there is already a small discussion started ... it
is 10pm here and I am happy and tired after this day of fighting and
booting ... :-)

-- so I will skip testing the removal of the patch for now, if you allow ;-)

So I got some decent cleanup today ... removed one of the RAID-arrays
(obsolete/unused), redid my fstab completely and modified the whole
dracut/grub2/kerninst-setup ... phew!

So my lvm2-lines aren't "kinda old" anymore, see fstab down there for
checks ...

I wonder if I should keep or remove the option
"comment=systemd.automount" ? I thought it would help to start up faster
while the non-important (in terms of boot-process) filesystems get
fscked and mounted in background. Something to test tomorrow ;-)


LABEL=ROOT	/		ext4 noatime,comment=systemd.automount       0 1

/dev/disk/by-uuid/10ccf5a2-0a37-42a6-886a-146e01cab301	none		swap	
defaults	0 0

tmpfs			/dev/shm	tmpfs		nodev,nosuid,noexec	0 0

tmpfs	/tmp			tmpfs	mode=1777	0 0
tmpfs	/var/tmp/portage	tmpfs	size=8000m,mode=1777	0 0

/dev/cdrw               /media/cdrecorder       auto
user,exec,noauto,managed 0 0

LABEL=HOME /home           ext4 noatime,user_xattr 0 2


LABEL=distfiles /usr/portage/distfiles	xfs
noatime,comment=systemd.automount 0 2
LABEL=dropbox 	/mnt/dropbox		ext4 	noatime,comment=systemd.automount 0 2
LABEL=gentoo32	/mnt/gentoo32		ext4 	noatime,comment=systemd.automount 0 2
LABEL=media	/mnt/media		ext4 	noatime,comment=systemd.automount 0 2
LABEL=music	/mnt/music		ext4 	noatime,comment=systemd.automount 0 2
LABEL=platz	/mnt/platz		ext4 	noatime,comment=systemd.automount 0 2
LABEL=oopsfiles /mnt/oopsfiles		ext4 	noatime,comment=systemd.automount 0 2



Thanks to all of you who responded and helped along this odyssey ;-) ...
I will check back for open ends in this thread and reply some missing
parts later (next days).

Looking forward to how these bugs develop ...


Greetings, best regards, Stefan

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 22:23, schrieb Stefan G. Weichinger:

> So I got some decent cleanup today ... removed one of the RAID-arrays
> (obsolete/unused), redid my fstab completely and modified the whole
> dracut/grub2/kerninst-setup ... phew!

[...]

> I wonder if I should keep or remove the option
> "comment=systemd.automount" ? I thought it would help to start up faster
> while the non-important (in terms of boot-process) filesystems get
> fscked and mounted in background. Something to test tomorrow ;-)

... note:

checked the man-pages and now have:

LABEL=distfiles /usr/portage/distfiles	xfs 	noatime,x-systemd.automount 0 2

(noauto optional)

--- if I am at it I could then attack UEFI booting again ... and
partition my SSD with GPT :-P

Good night now ... or have a fine day, Stefan

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Sat, May 3, 2014 at 2:41 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
[snip]
> Let me try to update this test VM to see if I can hit the same issue as
> you.

OK, I updated everything, and I certainly hit an issue. With
sys-kernel/dracut-037, I cannot boot; it gets stuck after loading the
initramfs and it complains about not able to mounting /. This happens
even with the patch on [1].

With sys-kernel/dracut-036-r4, everything works perfectly. I don't
know what the problem is exactly, but as a workaround I would
recommend staying on With sys-kernel/dracut-036-r4.

Regards.

[1] http://git.kernel.org/cgit/boot/dracut/dracut.git/commit/?id=12819a579900b9691e2bfaf14e76fbb025851530
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 05.05.2014 07:09, schrieb Canek Peláez Valdés:
> On Sat, May 3, 2014 at 2:41 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> [snip]
>> Let me try to update this test VM to see if I can hit the same issue as
>> you.
> 
> OK, I updated everything, and I certainly hit an issue. With
> sys-kernel/dracut-037, I cannot boot; it gets stuck after loading the
> initramfs and it complains about not able to mounting /. This happens
> even with the patch on [1].
> 
> With sys-kernel/dracut-036-r4, everything works perfectly. I don't
> know what the problem is exactly, but as a workaround I would
> recommend staying on With sys-kernel/dracut-036-r4.
> 
> Regards.
> 
> [1] http://git.kernel.org/cgit/boot/dracut/dracut.git/commit/?id=12819a579900b9691e2bfaf14e76fbb025851530

Interesting.

I had my VG not activated again when I booted today.

Going back to sys-kernel/dracut-036-r4 means downgrading systemd as well
here (at least with the current ebuilds in portage) ... I will skip that
for now as today I need this box up and running for work ;-)

Do you think we should file a bug for ... something? ;-)

Stefan

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Mon, May 5, 2014 at 1:18 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 05.05.2014 07:09, schrieb Canek Peláez Valdés:
>> On Sat, May 3, 2014 at 2:41 PM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> [snip]
>>> Let me try to update this test VM to see if I can hit the same issue as
>>> you.
>>
>> OK, I updated everything, and I certainly hit an issue. With
>> sys-kernel/dracut-037, I cannot boot; it gets stuck after loading the
>> initramfs and it complains about not able to mounting /. This happens
>> even with the patch on [1].
>>
>> With sys-kernel/dracut-036-r4, everything works perfectly. I don't
>> know what the problem is exactly, but as a workaround I would
>> recommend staying on With sys-kernel/dracut-036-r4.
>>
>> Regards.
>>
>> [1] http://git.kernel.org/cgit/boot/dracut/dracut.git/commit/?id=12819a579900b9691e2bfaf14e76fbb025851530
>
> Interesting.
>
> I had my VG not activated again when I booted today.
>
> Going back to sys-kernel/dracut-036-r4 means downgrading systemd as well
> here (at least with the current ebuilds in portage) ... I will skip that
> for now as today I need this box up and running for work ;-)

I think I already mentioned this: dracut is the only package
keyworded, so I run the stable version of systemd: 208-r3.

> Do you think we should file a bug for ... something? ;-)

There is for sure a regression with the latest dracut. That should be
reported, yes.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:

[ It's been more than a week since I last participated in the thread,
so I'm just replying to my last participation. ]

Stefan, have you tried to run "dracut --print-cmdline" and add that to
your kernel command line?

By the last thread related to systemd+dracut, that solved my problems
when using dracut 037. Could you try to see if it solves your issues?

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Thu, May 15, 2014 at 1:21 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>
> [ It's been more than a week since I last participated in the thread,
> so I'm just replying to my last participation. ]
>
> Stefan, have you tried to run "dracut --print-cmdline" and add that to
> your kernel command line?
>
> By the last thread related to systemd+dracut, that solved my problems
> when using dracut 037. Could you try to see if it solves your issues?

Also, I just noticed the --hostonly-cmdline option. Have you tried that?

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 08:49, schrieb Canek Peláez Valdés:
> On Thu, May 15, 2014 at 1:21 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>> On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>
>> [ It's been more than a week since I last participated in the thread,
>> so I'm just replying to my last participation. ]
>>
>> Stefan, have you tried to run "dracut --print-cmdline" and add that to
>> your kernel command line?
>>
>> By the last thread related to systemd+dracut, that solved my problems
>> when using dracut 037. Could you try to see if it solves your issues?
> 
> Also, I just noticed the --hostonly-cmdline option. Have you tried that?

Nope.

I am away from LVM & RAID now as mentioned in the other thread ... btrfs
everywhere ;-)

Stefan

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 09:08, schrieb Stefan G. Weichinger:
> Am 15.05.2014 08:49, schrieb Canek Peláez Valdés:
>> On Thu, May 15, 2014 at 1:21 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>> On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>>
>>> [ It's been more than a week since I last participated in the thread,
>>> so I'm just replying to my last participation. ]
>>>
>>> Stefan, have you tried to run "dracut --print-cmdline" and add that to
>>> your kernel command line?
>>>
>>> By the last thread related to systemd+dracut, that solved my problems
>>> when using dracut 037. Could you try to see if it solves your issues?
>>
>> Also, I just noticed the --hostonly-cmdline option. Have you tried that?
> 
> Nope.
> 
> I am away from LVM & RAID now as mentioned in the other thread ... btrfs
> everywhere ;-)

Aside from that: I always use your tool "kerninst" so I have "-H" set as
well.

Stefan

Re: [gentoo-user] boot problems

[covici]

Stefan G. Weichinger <lists@xunil.at> wrote:

> Am 15.05.2014 09:08, schrieb Stefan G. Weichinger:
> > Am 15.05.2014 08:49, schrieb Canek Peláez Valdés:
> >> On Thu, May 15, 2014 at 1:21 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> >>> On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> >>>
> >>> [ It's been more than a week since I last participated in the thread,
> >>> so I'm just replying to my last participation. ]
> >>>
> >>> Stefan, have you tried to run "dracut --print-cmdline" and add that to
> >>> your kernel command line?
> >>>
> >>> By the last thread related to systemd+dracut, that solved my problems
> >>> when using dracut 037. Could you try to see if it solves your issues?
> >>
> >> Also, I just noticed the --hostonly-cmdline option. Have you tried that?
> > 
> > Nope.
> > 
> > I am away from LVM & RAID now as mentioned in the other thread ... btrfs
> > everywhere ;-)
> 
> Aside from that: I always use your tool "kerninst" so I have "-H" set as
> well.
What is kerninst?   I do not see it in the repository.

-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 11:58, schrieb covici@ccs.covici.com:

> What is kerninst?   I do not see it in the repository.


https://github.com/canek-pelaez/kerninst

... but it uses GRUB2

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Thu, May 15, 2014 at 4:14 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 15.05.2014 09:08, schrieb Stefan G. Weichinger:
>> Am 15.05.2014 08:49, schrieb Canek Peláez Valdés:
>>> On Thu, May 15, 2014 at 1:21 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>>> On Mon, May 5, 2014 at 1:40 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
>>>>
>>>> [ It's been more than a week since I last participated in the thread,
>>>> so I'm just replying to my last participation. ]
>>>>
>>>> Stefan, have you tried to run "dracut --print-cmdline" and add that to
>>>> your kernel command line?
>>>>
>>>> By the last thread related to systemd+dracut, that solved my problems
>>>> when using dracut 037. Could you try to see if it solves your issues?
>>>
>>> Also, I just noticed the --hostonly-cmdline option. Have you tried that?
>>
>> Nope.
>>
>> I am away from LVM & RAID now as mentioned in the other thread ... btrfs
>> everywhere ;-)
>
> Aside from that: I always use your tool "kerninst" so I have "-H" set as
> well.

That's the issue; since version 037, --hostonly and --hostonly-cmdline
are *separated*. You need to specify both.

With -H, you don't get the kernel cmdline, and therefore your kernel
cannot load your LVM volumes since it doesn't know their... names? I
don't knot the terminology.

In any case, you need to set --hostonly-cmdline (or
hostonly_cmdline="yes" in the config file), *besides* -H.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:

> With -H, you don't get the kernel cmdline, and therefore your kernel
> cannot load your LVM volumes since it doesn't know their... names? I
> don't knot the terminology.
> 
> In any case, you need to set --hostonly-cmdline (or
> hostonly_cmdline="yes" in the config file), *besides* -H.

ok ... I pulled your changes (kerninst) from github ... on the web I see
it, but it doesn't get into my copy here ... strange.

As I don't need it right now, I will (a) wait or (b) edit manually.

No problem.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 20:27, schrieb Stefan G. Weichinger:
> Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:
> 
>> With -H, you don't get the kernel cmdline, and therefore your kernel
>> cannot load your LVM volumes since it doesn't know their... names? I
>> don't knot the terminology.
>>
>> In any case, you need to set --hostonly-cmdline (or
>> hostonly_cmdline="yes" in the config file), *besides* -H.
> 
> ok ... I pulled your changes (kerninst) from github ... on the web I see
> it, but it doesn't get into my copy here ... strange.
> 
> As I don't need it right now, I will (a) wait or (b) edit manually.

forget that. I had local changes ... git pull works.

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Thu, May 15, 2014 at 1:27 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:
>
>> With -H, you don't get the kernel cmdline, and therefore your kernel
>> cannot load your LVM volumes since it doesn't know their... names? I
>> don't knot the terminology.
>>
>> In any case, you need to set --hostonly-cmdline (or
>> hostonly_cmdline="yes" in the config file), *besides* -H.
>
> ok ... I pulled your changes (kerninst) from github ... on the web I see
> it, but it doesn't get into my copy here ... strange.
>
> As I don't need it right now, I will (a) wait or (b) edit manually.
>
> No problem.

I actually *removed* -H from kerninst. That should be configured in
the user's dracut.conf; now I have:

hostonly="yes"
hostonly_cmdline="yes"

in my dracut.conf.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 15.05.2014 20:33, schrieb Canek Peláez Valdés:
> On Thu, May 15, 2014 at 1:27 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
>> Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:
>>
>>> With -H, you don't get the kernel cmdline, and therefore your kernel
>>> cannot load your LVM volumes since it doesn't know their... names? I
>>> don't knot the terminology.
>>>
>>> In any case, you need to set --hostonly-cmdline (or
>>> hostonly_cmdline="yes" in the config file), *besides* -H.
>>
>> ok ... I pulled your changes (kerninst) from github ... on the web I see
>> it, but it doesn't get into my copy here ... strange.
>>
>> As I don't need it right now, I will (a) wait or (b) edit manually.
>>
>> No problem.
> 
> I actually *removed* -H from kerninst. That should be configured in
> the user's dracut.conf; now I have:
> 
> hostonly="yes"
> hostonly_cmdline="yes"
> 
> in my dracut.conf.

Yes, I understood ... thanks.

Aside from that a more general question:

Does it it any way help to have a *small* (= as small as possible)
initramfs?

Maybe on embedded systems but on the big multi-GB-ram-machines we use it
doesn't make much difference, right?

I ask because in all my reorganizing furor I also thought that now with
btrfs only I could get rid of "lvm mdraid" as dracut-modules. I can try
... ;-) (don't call me "ricer")

Additional in this context: does it make a noticeable difference which
"Kernel compression mode" you choose? I assume it is again an issue for
systems with (a) small boot-partitions and/or (b) slower CPUs to select
something special here.

I checked and see that I use LZ4 anyway already ... seems to be the
fastest to unpack as far as I understand the help text.

-

And then, who writes the howto condensed out of this thread? ;-)
Much to learn and understand as always, I appreciate it a lot.

Stefan

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Fri, May 16, 2014 at 4:46 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 15.05.2014 20:33, schrieb Canek Peláez Valdés:
>> On Thu, May 15, 2014 at 1:27 PM, Stefan G. Weichinger <lists@xunil.at> wrote:
>>> Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:
>>>
>>>> With -H, you don't get the kernel cmdline, and therefore your kernel
>>>> cannot load your LVM volumes since it doesn't know their... names? I
>>>> don't knot the terminology.
>>>>
>>>> In any case, you need to set --hostonly-cmdline (or
>>>> hostonly_cmdline="yes" in the config file), *besides* -H.
>>>
>>> ok ... I pulled your changes (kerninst) from github ... on the web I see
>>> it, but it doesn't get into my copy here ... strange.
>>>
>>> As I don't need it right now, I will (a) wait or (b) edit manually.
>>>
>>> No problem.
>>
>> I actually *removed* -H from kerninst. That should be configured in
>> the user's dracut.conf; now I have:
>>
>> hostonly="yes"
>> hostonly_cmdline="yes"
>>
>> in my dracut.conf.
>
> Yes, I understood ... thanks.
>
> Aside from that a more general question:
>
> Does it it any way help to have a *small* (= as small as possible)
> initramfs?
>
> Maybe on embedded systems but on the big multi-GB-ram-machines we use it
> doesn't make much difference, right?

AFAIU, no, it doesn't. As long as the (uncompressed) initramfs fits
into the RAM, its size doesn't matter.

> I ask because in all my reorganizing furor I also thought that now with
> btrfs only I could get rid of "lvm mdraid" as dracut-modules. I can try
> ... ;-) (don't call me "ricer")

Whatever gets rid of LVM is good on my book. I've never understood why
people uses it, and in my experience it only brings headaches.
Besides, I've heard from many people that btrfs is the way to go in
the future. I'm not ready to make the change yet, but I will at some
point.

> Additional in this context: does it make a noticeable difference which
> "Kernel compression mode" you choose? I assume it is again an issue for
> systems with (a) small boot-partitions and/or (b) slower CPUs to select
> something special here.

Given the size of the kernel, I don't thin the difference can be
humanly measured.

> I checked and see that I use LZ4 anyway already ... seems to be the
> fastest to unpack as far as I understand the help text.

It will be a difference of microseconds, if not nanoseconds. I
honestly don't think it matters at all.

> And then, who writes the howto condensed out of this thread? ;-)
> Much to learn and understand as always, I appreciate it a lot.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Alan McKinnon]

On 16/05/2014 12:04, Canek Peláez Valdés wrote:
> Whatever gets rid of LVM is good on my book. I've never understood why
> people uses it, and in my experience it only brings headaches.
> Besides, I've heard from many people that btrfs is the way to go in
> the future. I'm not ready to make the change yet, but I will at some
> point.


LVM is an excellent solution for what it was designed to do, which is to
deal with stuff like this:

Oops. I misjudged how big /var/log needed to be and now I need to add
50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
Now I need 5 hour downtime to play 15-pieces with fdisk.

LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
you have to hold the PV/VG/LV model in your head, but it also *fixes*
the issue with rigid MSDOS partition style.

Modern filesystems like ZFS and btrfs sidestep the need for LVM in a
really elegant and wonderful way, none of which changes the fact that
ZFS/btrfs weren't around when LVM was first coded.


-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] boot problems

[covici]

Alan McKinnon <alan.mckinnon@gmail.com> wrote:

> On 16/05/2014 12:04, Canek Peláez Valdés wrote:
> > Whatever gets rid of LVM is good on my book. I've never understood why
> > people uses it, and in my experience it only brings headaches.
> > Besides, I've heard from many people that btrfs is the way to go in
> > the future. I'm not ready to make the change yet, but I will at some
> > point.
> 
> 
> LVM is an excellent solution for what it was designed to do, which is to
> deal with stuff like this:
> 
> Oops. I misjudged how big /var/log needed to be and now I need to add
> 50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
> Now I need 5 hour downtime to play 15-pieces with fdisk.
> 
> LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
> you have to hold the PV/VG/LV model in your head, but it also *fixes*
> the issue with rigid MSDOS partition style.
> 
> Modern filesystems like ZFS and btrfs sidestep the need for LVM in a
> really elegant and wonderful way, none of which changes the fact that
> ZFS/btrfs weren't around when LVM was first coded.

So is btrfs ready for production -- all the tools work, etc. to the
level that the ext2/3/4 work?  Also, what kernel do you need for this to
function -- and last question, how to convert an lvm volume to btrfs, or
do you just have to make some space somewhere and copy the files?

So far, I have liked lvm, what's the advantage of btrfs over lvm?


-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]

On Fri, 16 May 2014 07:14:27 -0400, covici@ccs.covici.com wrote:

> So far, I have liked lvm, what's the advantage of btrfs over lvm?

I have only looked at btrfs, with a consideration for switching from ZFS,
but it seems to offer the same advantages as ZFS. That is, it makes
things even easier than LVM does. with LVM you can easily resize volumes
and the filesystems on them, but it is still two or three steps, more if
you add RAID into the equation. The modern filesystems do it all at once.
If you need a bigger var, you just tell it so. And it is exactly the same
process for shrinking a volume, something that can be tricky with LVM
because of the need to handle volume and filesystem separately.


-- 
Neil Bothwick

You know the end of the world is near when the Spice Girls start
reproducing.

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

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

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

Am 16.05.2014 14:03, schrieb Neil Bothwick:
> On Fri, 16 May 2014 07:14:27 -0400, covici@ccs.covici.com wrote:
> 
>> So far, I have liked lvm, what's the advantage of btrfs over
>> lvm?
> 
> I have only looked at btrfs, with a consideration for switching
> from ZFS, but it seems to offer the same advantages as ZFS. That
> is, it makes things even easier than LVM does. with LVM you can
> easily resize volumes and the filesystems on them, but it is still
> two or three steps, more if you add RAID into the equation. The
> modern filesystems do it all at once. If you need a bigger var, you
> just tell it so. And it is exactly the same process for shrinking a
> volume, something that can be tricky with LVM because of the need
> to handle volume and filesystem separately.

btrfs and zfs are removing the various layers we all had to deal with:

partitions, logical volumes, raid-arrays, filesystems, and then
snapshots etc.

With these modern filesystems you are able to basically say:

"I have these physical devices/disks, create me a pool of storage with
these properties" and then just use that pool in a flexible and
dynamic way.

Your disk based storage is then usable in a way RAM is, you add it and
it is available and you can then use it where you like it.

No (or let's say "much less" ...) fixed and hard barriers like
partition sizes, if you need space for /var, use it ... if you want to
set quotas on /home, just set them for the subvolume, if you add
another pair of harddisks, tell btrfs to redistribute redundancy
information ("re-balance").

(I see that Alan right now answered basically the same ;-) ).

You get checksums for your blocks and the possibility to repair rotted
blocks ... you get snapshots within the filesystem, no more slow
rsnapshot-crontabs ...

I used zfs-fuse back then and learned about the concepts, and it blew
my mind already years ago ;-)

zfs on linux ... it works fine for me on one server, but I never
really wanted it on my main machines (desktop and laptops) although I
once even wrote some "how to use zfs on your fully encrypted laptop"
for a magazine. It always feels like "suboptimal because it is not in
the kernel" to me (think licensing issues here).

btrfs is officially in the kernel, still marked "experimental" because
it is in active development, after all I read over the last days it
should be quite stable to use if you don't run very complex setups or
so ... and doing regular backups should be usual for the people in
this list, I assume? Distros like SLES come with btrfs as default fs
(soon).

I migrated ~3 machines to btrfs in the last days and I really love
getting rid of all the partitions and raids that grew over the years
... for now it is cleaned up and flexible and so far solid.

btrfs and zfs have different concepts for various aspects, but
basically the same goals. I definitely recommend to get in touch with
this generation of filesystems.


Stefan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=+F6u
-----END PGP SIGNATURE-----

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 948 bytes --]

On Fri, 16 May 2014 14:35:08 +0200, Stefan G. Weichinger wrote:

> zfs on linux ... it works fine for me on one server, but I never
> really wanted it on my main machines (desktop and laptops) although I
> once even wrote some "how to use zfs on your fully encrypted laptop"
> for a magazine. It always feels like "suboptimal because it is not in
> the kernel" to me (think licensing issues here).

That's why I'm looking at btrfs. ZFS is great, it does all I want it to.
But it is not in the kernel, which is not a major issue. More important
is that it is based on an old version of ZFS, later versions are still
closed source. That's a shame, because they support neat things like
encryption (yet another separate layer got rid of) and it means ZFS on
Linux can't really go anywhere beyond bug fixes and minor tweaks.


-- 
Neil Bothwick

There's a fine line between fishing and standing on the shore looking
like an idiot.

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

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

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

Am 16.05.2014 14:43, schrieb Neil Bothwick:
> On Fri, 16 May 2014 14:35:08 +0200, Stefan G. Weichinger wrote:
> 
>> zfs on linux ... it works fine for me on one server, but I never 
>> really wanted it on my main machines (desktop and laptops)
>> although I once even wrote some "how to use zfs on your fully
>> encrypted laptop" for a magazine. It always feels like
>> "suboptimal because it is not in the kernel" to me (think
>> licensing issues here).
> 
> That's why I'm looking at btrfs. ZFS is great, it does all I want
> it to. But it is not in the kernel, which is not a major issue.
> More important is that it is based on an old version of ZFS, later
> versions are still closed source. That's a shame, because they
> support neat things like encryption (yet another separate layer got
> rid of) and it means ZFS on Linux can't really go anywhere beyond
> bug fixes and minor tweaks.

Yes, this way one gets stuck somehow with ZFSonLinux.

btrfs also does not yet support encryption ... I assume that will come
over the time, I don't know if this is still correct:

https://btrfs.wiki.kernel.org/index.php/Project_ideas#Encryption

But the features available already are great ...

https://btrfs.wiki.kernel.org/index.php/Main_Page#Features

;-)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=MTN5
-----END PGP SIGNATURE-----

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1426 bytes --]

On Fri, 16 May 2014 15:03:43 +0200, Stefan G. Weichinger wrote:

> > That's why I'm looking at btrfs. ZFS is great, it does all I want
> > it to. But it is not in the kernel, which is not a major issue.
> > More important is that it is based on an old version of ZFS, later
> > versions are still closed source. That's a shame, because they
> > support neat things like encryption (yet another separate layer got
> > rid of) and it means ZFS on Linux can't really go anywhere beyond
> > bug fixes and minor tweaks.  
> 
> Yes, this way one gets stuck somehow with ZFSonLinux.
> 
> btrfs also does not yet support encryption ... I assume that will come
> over the time, I don't know if this is still correct:
> 
> https://btrfs.wiki.kernel.org/index.php/Project_ideas#Encryption

If it doesn't happen,. it is still on a par with ZFS on that front. But
there is always the possibility that someone will want it enough to
implement it... looks in the general direction of SUSE.

I like the idea of the version of RAID1 they have, where two copies of
the data are stored, even if you have multiple disks. It looks like it
should provide RAID5 like capacities without some of the overhead.

I've been moving my ZFS partitions around, which reminded me how good 
zfs send/zfs recv are, so I can give btrfs RAID a good try out soon.


-- 
Neil Bothwick

LISP: Lots of Infuriating & Silly Parentheses

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

Re: [gentoo-user] boot problems

[covici]

Stefan G. Weichinger <lists@xunil.at> wrote:

> Am 16.05.2014 14:03, schrieb Neil Bothwick:
> > On Fri, 16 May 2014 07:14:27 -0400, covici@ccs.covici.com wrote:
> > 
> >> So far, I have liked lvm, what's the advantage of btrfs over
> >> lvm?
> > 
> > I have only looked at btrfs, with a consideration for switching
> > from ZFS, but it seems to offer the same advantages as ZFS. That
> > is, it makes things even easier than LVM does. with LVM you can
> > easily resize volumes and the filesystems on them, but it is still
> > two or three steps, more if you add RAID into the equation. The
> > modern filesystems do it all at once. If you need a bigger var, you
> > just tell it so. And it is exactly the same process for shrinking a
> > volume, something that can be tricky with LVM because of the need
> > to handle volume and filesystem separately.
> 
> btrfs and zfs are removing the various layers we all had to deal with:
> 
> partitions, logical volumes, raid-arrays, filesystems, and then
> snapshots etc.
> 
> With these modern filesystems you are able to basically say:
> 
> "I have these physical devices/disks, create me a pool of storage with
> these properties" and then just use that pool in a flexible and
> dynamic way.
> 
> Your disk based storage is then usable in a way RAM is, you add it and
> it is available and you can then use it where you like it.
> 
> No (or let's say "much less" ...) fixed and hard barriers like
> partition sizes, if you need space for /var, use it ... if you want to
> set quotas on /home, just set them for the subvolume, if you add
> another pair of harddisks, tell btrfs to redistribute redundancy
> information ("re-balance").
> 
> (I see that Alan right now answered basically the same ;-) ).
> 
> You get checksums for your blocks and the possibility to repair rotted
> blocks ... you get snapshots within the filesystem, no more slow
> rsnapshot-crontabs ...
> 
> I used zfs-fuse back then and learned about the concepts, and it blew
> my mind already years ago ;-)
> 
> zfs on linux ... it works fine for me on one server, but I never
> really wanted it on my main machines (desktop and laptops) although I
> once even wrote some "how to use zfs on your fully encrypted laptop"
> for a magazine. It always feels like "suboptimal because it is not in
> the kernel" to me (think licensing issues here).
> 
> btrfs is officially in the kernel, still marked "experimental" because
> it is in active development, after all I read over the last days it
> should be quite stable to use if you don't run very complex setups or
> so ... and doing regular backups should be usual for the people in
> this list, I assume? Distros like SLES come with btrfs as default fs
> (soon).
> 
> I migrated ~3 machines to btrfs in the last days and I really love
> getting rid of all the partitions and raids that grew over the years
> ... for now it is cleaned up and flexible and so far solid.
> 
> btrfs and zfs have different concepts for various aspects, but
> basically the same goals. I definitely recommend to get in touch with
> this generation of filesystems.

Thanks much for that explanation.

So where do I find some documentation for btrfs and its user space tools?
-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 16.05.2014 14:54, schrieb covici@ccs.covici.com:

> Thanks much for that explanation.
> 
> So where do I find some documentation for btrfs and its user space tools?

There are many howtos and wiki-pages ... some examples, gentoo-related:

https://btrfs.wiki.kernel.org/index.php/Getting_started

http://www.funtoo.org/BTRFS_Fun

Stefan

Re: [gentoo-user] boot problems

[covici]

Stefan G. Weichinger <lists@xunil.at> wrote:

> Am 16.05.2014 14:54, schrieb covici@ccs.covici.com:
> 
> > Thanks much for that explanation.
> > 
> > So where do I find some documentation for btrfs and its user space tools?
> 
> There are many howtos and wiki-pages ... some examples, gentoo-related:
> 
> https://btrfs.wiki.kernel.org/index.php/Getting_started
Thanks, I will look at this in my copious free time.

-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: [gentoo-user] boot problems

[Alan McKinnon]

On 16/05/2014 13:14, covici@ccs.covici.com wrote:
> Alan McKinnon <alan.mckinnon@gmail.com> wrote:
> 
>> On 16/05/2014 12:04, Canek Peláez Valdés wrote:
>>> Whatever gets rid of LVM is good on my book. I've never understood why
>>> people uses it, and in my experience it only brings headaches.
>>> Besides, I've heard from many people that btrfs is the way to go in
>>> the future. I'm not ready to make the change yet, but I will at some
>>> point.
>>
>>
>> LVM is an excellent solution for what it was designed to do, which is to
>> deal with stuff like this:
>>
>> Oops. I misjudged how big /var/log needed to be and now I need to add
>> 50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
>> Now I need 5 hour downtime to play 15-pieces with fdisk.
>>
>> LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
>> you have to hold the PV/VG/LV model in your head, but it also *fixes*
>> the issue with rigid MSDOS partition style.
>>
>> Modern filesystems like ZFS and btrfs sidestep the need for LVM in a
>> really elegant and wonderful way, none of which changes the fact that
>> ZFS/btrfs weren't around when LVM was first coded.
> 
> So is btrfs ready for production -- all the tools work, etc. to the
> level that the ext2/3/4 work?  Also, what kernel do you need for this to
> function -- and last question, how to convert an lvm volume to btrfs, or
> do you just have to make some space somewhere and copy the files?
> 
> So far, I have liked lvm, what's the advantage of btrfs over lvm?



I don't have enough experience with btrfs to answer, but I believe it's
much like ZFS in many ways. So here goes:

ZFS dispenses with the entire concept of partitions and rigidly
allocated areas of storage on a disk. All you really have is "storage".
You can divide it up into chunks and sections that look and feel like
volumes and partitions but that is not how it's implemented. You don't
create a 50G partition for logs, you tell the system to give you 50G of
space you will put logs in. And that "space" is something you can mount,
and apply permissions and quotas to.

It's a lot like having the best parts of partitions and directories in
one unit with none of the rigidity and downsides, and the whole lot is
done in a virtual manner by software.

You can drop the entire hierarchy of disk/partition/pv/vg/lv/fs right
out of your head with these new modern systems, and just not ever have
to deal with that complexity at all.

-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] boot problems

[Marc Joliet]

[-- Attachment #1: Type: text/plain, Size: 1170 bytes --]

Am Fri, 16 May 2014 07:14:27 -0400
schrieb covici@ccs.covici.com:

[...]
> Also, what kernel do you need for this to
> function -- and last question, how to convert an lvm volume to btrfs, or
> do you just have to make some space somewhere and copy the files?

I am not aware of being able to directly convert an LV; copying your files from
a backup (or whatever) to a newly created btrfs is AFAIK your only option. You
can directly convert Ext4 partitions, though (and even revert), though I have
not tried this myself.

As to the kernel version, the recommendation is to use the most recent stable
versions of the kernel and btrfs-progs.

> So far, I have liked lvm, what's the advantage of btrfs over lvm?

I also liked LVM while I had it, but I like btrfs even more :) . But yeah, this
has already been thoroughly answered. Additionally, you can also check the
thread "planned btrfs conversion: questions", which I started, where others also
posted their experiences with btrfs (and ZFS, too, to a smaller degree).

-- 
Marc Joliet
--
"People who think they know everything really annoy those of us who know we
don't" - Bjarne Stroustrup

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[gentoo-user] Re: LVM

[Stefan G. Weichinger]

Am 16.05.2014 13:06, schrieb Alan McKinnon:

> LVM is an excellent solution for what it was designed to do, which is to
> deal with stuff like this:
> 
> Oops. I misjudged how big /var/log needed to be and now I need to add
> 50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
> Now I need 5 hour downtime to play 15-pieces with fdisk.
> 
> LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
> you have to hold the PV/VG/LV model in your head, but it also *fixes*
> the issue with rigid MSDOS partition style.
> 
> Modern filesystems like ZFS and btrfs sidestep the need for LVM in a
> really elegant and wonderful way, none of which changes the fact that
> ZFS/btrfs weren't around when LVM was first coded.

exactly. I loved LVM when it was new as it was a way to get the
mentioned capability to resize filesystems and underlying "partitions".

And I still use it for servers, creating a VG on the mdadm-RAID-array
and only providing a part of it for the customers ... if they then fill
up their samba-shares with cat pictures I can easily ssh in and give
them some more space in a minute ... that is nice to have!

OK, I also had some issues with LVM over the years ... but not in a
regular way, more when physical volumes got flaky or so. In general it
just works for me (and show me one piece of tech where you are
guaranteed to not have issues with ...)

But sure, now I also think of using btrfs on one of the next fileservers
I deliver ... and instead of using rsnapshots to give customers a
readonly history of their data there could be btrfs-snapshots.

time changes, things develop ...

Stefan

Re: [gentoo-user] boot problems

[Marc Joliet]

[-- Attachment #1: Type: text/plain, Size: 1715 bytes --]

Am Fri, 16 May 2014 13:06:41 +0200
schrieb Alan McKinnon <alan.mckinnon@gmail.com>:

> On 16/05/2014 12:04, Canek Peláez Valdés wrote:
> > Whatever gets rid of LVM is good on my book. I've never understood why
> > people uses it, and in my experience it only brings headaches.
> > Besides, I've heard from many people that btrfs is the way to go in
> > the future. I'm not ready to make the change yet, but I will at some
> > point.
> 
> 
> LVM is an excellent solution for what it was designed to do, which is to
> deal with stuff like this:
> 
> Oops. I misjudged how big /var/log needed to be and now I need to add
> 50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
> Now I need 5 hour downtime to play 15-pieces with fdisk.
> 
> LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
> you have to hold the PV/VG/LV model in your head, but it also *fixes*
> the issue with rigid MSDOS partition style.

This is precisely why I switched to RAID + LVM several years ago, instead of
just RAID. No, wait, that's not correct: I remember now that I in fact started
with just LVM on two differently-sized disks.

But even without a RAID underneath, you can manage multiple disks (PVs) in one
or more VGs and do stuff like move partitions between disks.

> Modern filesystems like ZFS and btrfs sidestep the need for LVM in a
> really elegant and wonderful way, none of which changes the fact that
> ZFS/btrfs weren't around when LVM was first coded.

And this is one of the reasons why I switched to btrfs now :) .

-- 
Marc Joliet
--
"People who think they know everything really annoy those of us who know we
don't" - Bjarne Stroustrup

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-user] boot problems

[Walter Dnes]

On Fri, May 16, 2014 at 01:06:41PM +0200, Alan McKinnon wrote
> 
> LVM is an excellent solution for what it was designed to do, which is to
> deal with stuff like this:
> 
> Oops. I misjudged how big /var/log needed to be and now I need to add
> 50G to that partition. But it's sda6 and I have up to sda8. Arggghhhhh!
> Now I need 5 hour downtime to play 15-pieces with fdisk.
> 
> LVM makes that 2 commands and 12 seconds. Yes, it's a bit complex and
> you have to hold the PV/VG/LV model in your head, but it also *fixes*
> the issue with rigid MSDOS partition style.

  The concept of "Logical Volume Management" is also useful for an
encryption/decryption layer on top of an encrypted USB key.  lvm2 is a
mandatory dependancy for cryptsetup.

-- 
Walter Dnes <waltdnes@waltdnes.org>
I don't run "desktop environments"; I run useful applications

Re: [gentoo-user] boot problems

[Tanstaafl]

On 5/16/2014 6:04 AM, Canek Peláez Valdés <caneko@gmail.com> wrote:
> Whatever gets rid of LVM is good on my book. I've never understood why
> people uses it, and in my experience it only brings headaches.

One very big reason to have been using it on linux - since it is only 
relatively recently that zfs has been an option, and now btrfs is 
getting there - and about the only reason I use it, is for snapshots, 
which bring consistent point-in-time backups for things like mail 
servers....

Others use it for it filesystem resizing abilities, but I've only had to 
do this once or twice and that was a long time ago - but, I was able to 
do it... :)

Yes, zfs, and now btrfs, now offer much better support for both of these 
and then some, but they weren't around/available for linux until 
relatively recently, which LVM was stable and mature.

Re: [gentoo-user] boot problems

[Bruce Schultz]

On 16 May 2014 7:46:29 PM AEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>Am 15.05.2014 20:33, schrieb Canek Peláez Valdés:
>> On Thu, May 15, 2014 at 1:27 PM, Stefan G. Weichinger
><lists@xunil.at> wrote:
>>> Am 15.05.2014 20:05, schrieb Canek Peláez Valdés:
>>>
>>>> With -H, you don't get the kernel cmdline, and therefore your
>kernel
>>>> cannot load your LVM volumes since it doesn't know their... names?
>I
>>>> don't knot the terminology.
>>>>
>>>> In any case, you need to set --hostonly-cmdline (or
>>>> hostonly_cmdline="yes" in the config file), *besides* -H.
>>>
>>> ok ... I pulled your changes (kerninst) from github ... on the web I
>see
>>> it, but it doesn't get into my copy here ... strange.
>>>
>>> As I don't need it right now, I will (a) wait or (b) edit manually.
>>>
>>> No problem.
>> 
>> I actually *removed* -H from kerninst. That should be configured in
>> the user's dracut.conf; now I have:
>> 
>> hostonly="yes"
>> hostonly_cmdline="yes"
>> 
>> in my dracut.conf.
>
>Yes, I understood ... thanks.
>
>Aside from that a more general question:
>
>Does it it any way help to have a *small* (= as small as possible)
>initramfs?
>
>Maybe on embedded systems but on the big multi-GB-ram-machines we use
>it
>doesn't make much difference, right?
>
>I ask because in all my reorganizing furor I also thought that now with
>btrfs only I could get rid of "lvm mdraid" as dracut-modules. I can try
>... ;-) (don't call me "ricer")

If you have a multi-disk btrfs, I think you need to add the btrfs dracut module. At least that's how I remember it, but its been a while & my memory could be failing me, or it could well have changed since then.

Bruce
--
:B

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 16.05.2014 13:56, schrieb Bruce Schultz:

>> I ask because in all my reorganizing furor I also thought that now
>> with btrfs only I could get rid of "lvm mdraid" as dracut-modules.
>> I can try ... ;-) (don't call me "ricer")
> 
> If you have a multi-disk btrfs, I think you need to add the btrfs
> dracut module. At least that's how I remember it, but its been a
> while & my memory could be failing me, or it could well have changed
> since then.

I currently have in /etc/dracut.conf:

add_dracutmodules+="bash systemd"
hostonly="yes"
hostonly_cmdline="yes"

and am able to boot via grub2 and efi, that means,

/boot/efi (vfat) on /dev/sda1 (sda is the SSD)

and

/ as btrfs-subvol on /dev/sda2 (with /boot as directory on it).

So no multi-disk btrfs for / or /boot on this machine.

-

removing "lvm mdraid" from dracut slimmed down the initrd from around
6.8 MB to 5.5 MB ... nice, but not necessary as Canek mentioned.

I just "play" with it to learn and understand even better.

I am not sure what the module "systemd" does or is good for, and

https://www.kernel.org/pub/linux/utils/boot/dracut/dracut.html#dracut.kernel

does look scary and didn't tell me more about that module yet.

I will read more when I find the time, for now it works very well here.


Stefan

Re: [gentoo-user] boot problems

[Canek Peláez Valdés]

On Fri, May 16, 2014 at 7:10 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
> Am 16.05.2014 13:56, schrieb Bruce Schultz:
>
>>> I ask because in all my reorganizing furor I also thought that now
>>> with btrfs only I could get rid of "lvm mdraid" as dracut-modules.
>>> I can try ... ;-) (don't call me "ricer")
>>
>> If you have a multi-disk btrfs, I think you need to add the btrfs
>> dracut module. At least that's how I remember it, but its been a
>> while & my memory could be failing me, or it could well have changed
>> since then.
>
> I currently have in /etc/dracut.conf:
>
> add_dracutmodules+="bash systemd"
> hostonly="yes"
> hostonly_cmdline="yes"
>
> and am able to boot via grub2 and efi, that means,
>
> /boot/efi (vfat) on /dev/sda1 (sda is the SSD)
>
> and
>
> / as btrfs-subvol on /dev/sda2 (with /boot as directory on it).
>
> So no multi-disk btrfs for / or /boot on this machine.
>
> -
>
> removing "lvm mdraid" from dracut slimmed down the initrd from around
> 6.8 MB to 5.5 MB ... nice, but not necessary as Canek mentioned.
>
> I just "play" with it to learn and understand even better.
>
> I am not sure what the module "systemd" does or is good for, and
>
> https://www.kernel.org/pub/linux/utils/boot/dracut/dracut.html#dracut.kernel
>
> does look scary and didn't tell me more about that module yet.

With the systemd module, dracut uses systemd as init. So you have your
initramfs, which executes systemd to setup the machine up until /usr
is mounted, and then that instance of systemd calls the instance of
systemd in the hard drive.

If you don't use the systemd module, your initramfs uses a custom init
script to mount /usr and similar stuff. It seems to work, but I trust
more systemd to do it, and I certainly prefer not to rely on any
script during boot up.

Also, for non-trivial setups, I'm sure systemd handles them better
that a custom init script; that's why I recommended John to use the
systemd module.

Regards.
-- 
Canek Peláez Valdés
Profesor de asignatura, Facultad de Ciencias
Universidad Nacional Autónoma de México

Re: [gentoo-user] boot problems

[Peter Humphrey]

On Saturday 03 May 2014 14:17:53 J. Roeleveld wrote:

> With genkernel, I have a lvm and mdadm boot parameter.
> With dracut, that might also be necessary.
> 
> I would prefer the initramfs to do it automagically without extra
> parameters.

Hope I'm not butting in here, but...

Although I don't run systemd nor do I have an initramfs, the grub.conf entry 
for my LVM2 setup is just these two lines:

title=Gentoo Linux 3.12.13
        kernel /boot/kernel-x86_64-3.12.13-gentoo root=/dev/md5 net.ifnames=0

I've noticed several times (often much to my annoyance before I discovered 
what to do about it*) that starting of the raid arrays is automatic, 
apparently done by the kernel though I could be wrong about that. In fact I 
was astonished to find not long ago that I'd been running for a year or two 
with neither lvm2 nor mdraid installed!

* SystemRescueCD and the Gentoo minimal installation CD both start any raid 
arrays they find and apply their own names to them. It is then impossible, or 
so I thought, to resume an interrupted installation process. Of course, all I 
had to do was "mdadm --stop /dev/md127" etc.

-- 
Regards
Peter

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Saturday, May 03, 2014 06:09:21 PM Peter Humphrey wrote:
> On Saturday 03 May 2014 14:17:53 J. Roeleveld wrote:
> > With genkernel, I have a lvm and mdadm boot parameter.
> > With dracut, that might also be necessary.
> > 
> > I would prefer the initramfs to do it automagically without extra
> > parameters.
> 
> Hope I'm not butting in here, but...
> 
> Although I don't run systemd nor do I have an initramfs, the grub.conf entry
> for my LVM2 setup is just these two lines:
> 
> title=Gentoo Linux 3.12.13
>         kernel /boot/kernel-x86_64-3.12.13-gentoo root=/dev/md5
> net.ifnames=0

That works with metadata=0.9 when creating the raid-1 device and not having " 
/ " on LVM.

I only have /boot on a raid-1 with metadata=0.9.
All the other partitions are LVs with the lvm layer ontop of a raid-0.
(no important data is stored locally on the desktop machines)

> I've noticed several times (often much to my annoyance before I discovered
> what to do about it*) that starting of the raid arrays is automatic,
> apparently done by the kernel though I could be wrong about that. In fact I
> was astonished to find not long ago that I'd been running for a year or two
> with neither lvm2 nor mdraid installed!

Something must have handled the LVM part. Afaik, there is no kernel auto-
detect for LVM.

> * SystemRescueCD and the Gentoo minimal installation CD both start any raid
> arrays they find and apply their own names to them. It is then impossible,
> or so I thought, to resume an interrupted installation process. Of course,
> all I had to do was "mdadm --stop /dev/md127" etc.

Yes, I noticed that annoyance myself. I would much prefer it to default to 
more logical names.

--
Joost

Re: [gentoo-user] boot problems

[Peter Humphrey]

On Saturday 03 May 2014 20:40:47 J. Roeleveld wrote:
> On Saturday, May 03, 2014 06:09:21 PM Peter Humphrey wrote:
> > Hope I'm not butting in here, but...
> > 
> > Although I don't run systemd nor do I have an initramfs, the grub.conf
> > entry for my LVM2 setup is just these two lines:
> > 
> > title=Gentoo Linux 3.12.13
> >         kernel /boot/kernel-x86_64-3.12.13-gentoo root=/dev/md5
> > net.ifnames=0
> 
> That works with metadata=0.9 when creating the raid-1 device and not having
> " / " on LVM.

Yes, I should have added that I have /boot on straight ext2 on /dev/sda1 (sdb1 
is ready in case I decide to raid it). The file-system root is on raid1 and 
everything else is in LVs on /dev/md7, also raid1. I did it that way to keep 
booting simple. And I've used 0.9 metadata throughout.

> I only have /boot on a raid-1 with metadata=0.9.
> All the other partitions are LVs with the lvm layer ontop of a raid-0.
> (no important data is stored locally on the desktop machines)

I wanted the little extra safety of raid1 because this is where I keep all my 
data.

> > I've noticed several times (often much to my annoyance before I discovered
> > what to do about it*) that starting of the raid arrays is automatic,
> > apparently done by the kernel though I could be wrong about that. In fact 
> > I was astonished to find not long ago that I'd been running for a year or
> > two with neither lvm2 nor mdraid installed!
> 
> Something must have handled the LVM part. Afaik, there is no kernel auto-
> detect for LVM.

Yes, that's why I mentioned it. If it's not the kernel I don't know what else 
it could have been. Udev? I don't see anything relevant under /etc/udev.

> > * SystemRescueCD and the Gentoo minimal installation CD both start any
> > raid
> > arrays they find and apply their own names to them. It is then impossible,
> > or so I thought, to resume an interrupted installation process. Of course,
> > all I had to do was "mdadm --stop /dev/md127" etc.
> 
> Yes, I noticed that annoyance myself. I would much prefer it to default to
> more logical names.

If the docs had included that little snippet I'd have saved myself many a 
frustrating hour. I'll only look stupid if I tell you how many  ;-)

Anyway, I don't want to hijack the thread. I just wanted to point out that 
raid arrays don't need lvm2 or mdraid present to auto-start, at least not on 
my openrc box which also has no initramfs.

-- 
Regards
Peter

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 22:50, schrieb Peter Humphrey:

>> Yes, I noticed that annoyance myself. I would much prefer it to default to
>> more logical names.
> 
> If the docs had included that little snippet I'd have saved myself many a 
> frustrating hour. I'll only look stupid if I tell you how many  ;-)
> 
> Anyway, I don't want to hijack the thread. I just wanted to point out that 
> raid arrays don't need lvm2 or mdraid present to auto-start, at least not on 
> my openrc box which also has no initramfs.

Thanks for your contribution. I spent nearly the whole day digging
around this issue ...

I wonder if I speak for more users when I say that all this is kind of
confusing sometimes ... and as mentioned I take the opportunity to clean
up this (my!) mess here now as it grew over the years.

I am not so far to skip the initramfs -> I don't *know* that, I just
tested removing the line from grub2 and it failed finding the root-fs.

For booting from a plain partition on an SSD I think I shouldn't need an
initramfs? Does it have to do with MBR/GPT as well (the SSD is
still/again MBR, as UEFI booting broke badly for me back then) ?

Maybe I learn more soon ;-)

Stefan

Re: [gentoo-user] boot problems

[Peter Humphrey]

On Saturday 03 May 2014 23:04:49 Stefan G. Weichinger wrote:
> Am 03.05.2014 22:50, schrieb Peter Humphrey:
> >> Yes, I noticed that annoyance myself. I would much prefer it to default
> >> to
> >> more logical names.
> > 
> > If the docs had included that little snippet I'd have saved myself many a
> > frustrating hour. I'll only look stupid if I tell you how many  ;-)
> > 
> > Anyway, I don't want to hijack the thread. I just wanted to point out that
> > raid arrays don't need lvm2 or mdraid present to auto-start, at least not
> > on my openrc box which also has no initramfs.
> 
> Thanks for your contribution.

My pleasure.

> I spent nearly the whole day digging around this issue ...

You did better than I did recently: I spent four days at it.

> I wonder if I speak for more users when I say that all this is kind of
> confusing sometimes ...

I'm with you there, Stefan. I find the whole RAID and LVM area deeply 
mysterious, and the docs I've seen only say what to do, not why. I'd still 
like to find a proper explanation of how it all works.

> I am not so far to skip the initramfs -> I don't *know* that, I just
> tested removing the line from grub2 and it failed finding the root-fs.

I've never had an initramfs, seeing no need in my case to keep /usr on its own 
partition.

> For booting from a plain partition on an SSD I think I shouldn't need an
> initramfs? Does it have to do with MBR/GPT as well (the SSD is
> still/again MBR, as UEFI booting broke badly for me back then) ?

As far as I know, the only thing that /requires/ an initramfs is having a 
separate /usr. And I can't help you with GPT or UEFI - sorry.

> Maybe I learn more soon ;-)

I sometimes say that life is just one long journey of discovery  :-)

-- 
Regards
Peter

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 04.05.2014 10:53, schrieb Peter Humphrey:

>> I spent nearly the whole day digging around this issue ...
> 
> You did better than I did recently: I spent four days at it.

phew! ;-)


>> I wonder if I speak for more users when I say that all this is kind
>> of confusing sometimes ...
> 
> I'm with you there, Stefan. I find the whole RAID and LVM area deeply
>  mysterious, and the docs I've seen only say what to do, not why. I'd
> still like to find a proper explanation of how it all works.

I think I understand how it works (in a somewhat higher level of being
between the plain user and the interested admin ... not understanding
all the exact details on the lowest level) .... but I always feel that
it is hard to "do it right", even when I follow howtos and dig through docs.

Maybe it is related to being interested in latest development, for sure
I have to accept hitting bugs and issues when I run unstable software.

But on the other hand:

over the years the "best practise" changes ... for example what your
fstab should/could look like or how to partition your harddrive.

Do I have to change things because it's better that way, is it worth the
effort ... ? Should I go away from RAID because LVM could stripe/mirror
by itself? Should I go away from LVM because it's kinda old technology?
... all these things to consider.

And then you get into issues with block sizes and stuff, where I always
wonder why *I* have to type all these parameters ... why doesn't modern
software just come with this knowledge inside?

 .... you know

*sigh* ;-)


>> I am not so far to skip the initramfs -> I don't *know* that, I
>> just tested removing the line from grub2 and it failed finding the
>> root-fs.
> 
> I've never had an initramfs, seeing no need in my case to keep /usr
> on its own partition.

I don't have that either ...

>> For booting from a plain partition on an SSD I think I shouldn't
>> need an initramfs? Does it have to do with MBR/GPT as well (the SSD
>> is still/again MBR, as UEFI booting broke badly for me back then)
>> ?
> 
> As far as I know, the only thing that /requires/ an initramfs is
> having a separate /usr. And I can't help you with GPT or UEFI -
> sorry.

As mentioned: I don't know if it has any benefits in my case.

My desktop once was set up to boot gentoo via UEFI (Grub2), worked OK,
then something happened and I spent hours to fix it, then went back to
BIOS/MBR. I just thought I could set that up now that I clean through my
disks and partitioning.

One of my thinkpads boots via UEFI, that was rather straight to set up
and works fine.

>> Maybe I learn more soon ;-)
> 
> I sometimes say that life is just one long journey of discovery  :-)

Definitely!

Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 11:15:22 AM Stefan G. Weichinger wrote:
> Do I have to change things because it's better that way, is it worth the
> effort ... ? Should I go away from RAID because LVM could stripe/mirror
> by itself? Should I go away from LVM because it's kinda old technology?
> ... all these things to consider.

I wouldn't use the stripe/mirror support in LVM as I don't think it is used 
often and I feel that functionality doesn't belong in LVM.
If you want to move it all into a single layer, I would suggest ZFS instead.

> And then you get into issues with block sizes and stuff, where I always
> wonder why *I* have to type all these parameters ... why doesn't modern
> software just come with this knowledge inside?
> 
>  .... you know
> 
> *sigh* ;-)

I agree, and I feel that has actually improved over time with modern tools 
defaulting to 4k sectors.

> >> I am not so far to skip the initramfs -> I don't *know* that, I
> >> just tested removing the line from grub2 and it failed finding the
> >> root-fs.
> > 
> > I've never had an initramfs, seeing no need in my case to keep /usr
> > on its own partition.
> 
> I don't have that either ...

Then it should work, provided you have all the required drivers inside your 
kernel and not as modules.
I also believe an initramfs is needed when using LABELs for the root-fs.

> >> For booting from a plain partition on an SSD I think I shouldn't
> >> need an initramfs? Does it have to do with MBR/GPT as well (the SSD
> >> is still/again MBR, as UEFI booting broke badly for me back then)
> >> ?
> > 
> > As far as I know, the only thing that /requires/ an initramfs is
> > having a separate /usr. And I can't help you with GPT or UEFI -
> > sorry.
> 
> As mentioned: I don't know if it has any benefits in my case.
> 
> My desktop once was set up to boot gentoo via UEFI (Grub2), worked OK,
> then something happened and I spent hours to fix it, then went back to
> BIOS/MBR. I just thought I could set that up now that I clean through my
> disks and partitioning.
> 
> One of my thinkpads boots via UEFI, that was rather straight to set up
> and works fine.

At the moment, I don't see, from a simple user perspective, any real difference 
between booting using UEFI and BIOS/MBR.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 04.05.2014 12:49, schrieb J. Roeleveld:

> I wouldn't use the stripe/mirror support in LVM as I don't think it is used 
> often and I feel that functionality doesn't belong in LVM.
> If you want to move it all into a single layer, I would suggest ZFS instead.

Oh, yes, I like ZFS and its features and used it in some cases already.
But I didn't yet take the step to set up ZFS-root on my work machines.

>> And then you get into issues with block sizes and stuff, where I always
>> wonder why *I* have to type all these parameters ... why doesn't modern
>> software just come with this knowledge inside?
>>
>>  .... you know
>>
>> *sigh* ;-)
> 
> I agree, and I feel that has actually improved over time with modern tools 
> defaulting to 4k sectors.

Yes, but it always feels like "I missed something" when you look at the
various layers: partitioning at correct sectors, RAIDs with their
parameters, creating PVs with PEs aligned ... and then the filesystems.
I never get the feeling that I really did it right from the base to the
top.

> Then it should work, provided you have all the required drivers inside your 
> kernel and not as modules.
> I also believe an initramfs is needed when using LABELs for the root-fs.

Interesting. I don't really care if I have an initramfs or not, as long
as things work ... The feature with LABELs is nice for preparing
installations in VMs and then move it to physical hardware (eg.
/dev/vda1 then becomes /dev/sda1 or /dev/md0 and booting fails).

> At the moment, I don't see, from a simple user perspective, any real difference 
> between booting using UEFI and BIOS/MBR.

UEFI is one thing, GPT partitioning another. Being able to have more
than the 4 primary partitions of MBR looks good to me ...

I will see where my motivation leads to ;-)

Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 03:07:28 PM Stefan G. Weichinger wrote:
> Am 04.05.2014 12:49, schrieb J. Roeleveld:
> > I wouldn't use the stripe/mirror support in LVM as I don't think it is
> > used
> > often and I feel that functionality doesn't belong in LVM.
> > If you want to move it all into a single layer, I would suggest ZFS
> > instead.
> Oh, yes, I like ZFS and its features and used it in some cases already.
> But I didn't yet take the step to set up ZFS-root on my work machines.

I haven't yet, but it's on the list of items to look into at some point.
I'd like to know if, with ZFS, it is possible to create block-devices like LVs 
which I can then attach to VMs. Or if I have to use files instead.

> >> And then you get into issues with block sizes and stuff, where I always
> >> wonder why *I* have to type all these parameters ... why doesn't modern
> >> software just come with this knowledge inside?
> >> 
> >>  .... you know
> >> 
> >> *sigh* ;-)
> > 
> > I agree, and I feel that has actually improved over time with modern tools
> > defaulting to 4k sectors.
> 
> Yes, but it always feels like "I missed something" when you look at the
> various layers: partitioning at correct sectors, RAIDs with their
> parameters, creating PVs with PEs aligned ... and then the filesystems.
> I never get the feeling that I really did it right from the base to the
> top.

True, but if all the tools are working correctly, it all should cascade down 
when getting the "sector" sizes from the previous layer.
Not sure if this actually works.

I think it does as I get the following on my server:
***
# gdisk -l /dev/sda
GPT fdisk (gdisk) version 0.8.8

Partition table scan:
  MBR: protective
  BSD: not present
  APM: not present
  GPT: present

Found valid GPT with protective MBR; using GPT.
Disk /dev/sda: 11718749184 sectors, 5.5 TiB
Logical sector size: 512 bytes
Disk identifier (GUID): 936FDBE4-A736-41CF-B9A5-51069940D3DB
Partition table holds up to 128 entries
First usable sector is 34, last usable sector is 11718749150
Partitions will be aligned on 2048-sector boundaries
Total free space is 2014 sectors (1007.0 KiB)

Number  Start (sector)    End (sector)  Size       Code  Name
   1            2048          411647   200.0 MiB   EF00  EFI System
   2          411648         2101247   825.0 MiB   8300  Linux filesystem
   3         2101248     11718749150   5.5 TiB     8E00  Linux LVM

***

That's a hardware raid device with 4 * 3TB disks with raid-6.
I don't like the fact that a 2nd disk-failure can kill a raid-10 when both 
disks are in the same mirror-set.

gdisk automatically aligns on 2048 sector boundaries, that is more then enough 
for the 4k-sectors and the block/stripe sizes employed by the raid-controller.

> > Then it should work, provided you have all the required drivers inside
> > your
> > kernel and not as modules.
> > I also believe an initramfs is needed when using LABELs for the root-fs.
> 
> Interesting. I don't really care if I have an initramfs or not, as long
> as things work ... The feature with LABELs is nice for preparing
> installations in VMs and then move it to physical hardware (eg.
> /dev/vda1 then becomes /dev/sda1 or /dev/md0 and booting fails).

UUIDs, I believe, do work natively. And those are stored inside the partition 
itself. Which means they should also work. But are not as easy to locate. (eg. 
you don't specify them yourself)

> > At the moment, I don't see, from a simple user perspective, any real
> > difference between booting using UEFI and BIOS/MBR.
> 
> UEFI is one thing, GPT partitioning another. Being able to have more
> than the 4 primary partitions of MBR looks good to me ...
> 
> I will see where my motivation leads to ;-)

See the partitioning on my server above.
It boots using BIOS as I haven't been able to boot Xen using UEFI yet.

Support should be there now, but not been able to test that yet.
GPT is supported by grub-1 (and grub2) and with the MBR-support inside GPT, 
booting works from BIOS/MBR.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 04.05.2014 20:40, schrieb J. Roeleveld:
> On Sunday, May 04, 2014 03:07:28 PM Stefan G. Weichinger wrote:

>> Oh, yes, I like ZFS and its features and used it in some cases already.
>> But I didn't yet take the step to set up ZFS-root on my work machines.
> 
> I haven't yet, but it's on the list of items to look into at some point.
> I'd like to know if, with ZFS, it is possible to create block-devices like LVs 
> which I can then attach to VMs. Or if I have to use files instead.

I think you would have to use files on top of ZFS ... but I am not
up-to-date in that area.

People run KVM-hosts with storage on ZFS ... nice with the snapshots etc ...

> I think it does as I get the following on my server:
> ***
> # gdisk -l /dev/sda
> GPT fdisk (gdisk) version 0.8.8
> 
> Partition table scan:
>   MBR: protective
>   BSD: not present
>   APM: not present
>   GPT: present
> 
> Found valid GPT with protective MBR; using GPT.
> Disk /dev/sda: 11718749184 sectors, 5.5 TiB
> Logical sector size: 512 bytes
> Disk identifier (GUID): 936FDBE4-A736-41CF-B9A5-51069940D3DB
> Partition table holds up to 128 entries
> First usable sector is 34, last usable sector is 11718749150
> Partitions will be aligned on 2048-sector boundaries
> Total free space is 2014 sectors (1007.0 KiB)
> 
> Number  Start (sector)    End (sector)  Size       Code  Name
>    1            2048          411647   200.0 MiB   EF00  EFI System
>    2          411648         2101247   825.0 MiB   8300  Linux filesystem
>    3         2101248     11718749150   5.5 TiB     8E00  Linux LVM
> 
> ***
> 
> That's a hardware raid device with 4 * 3TB disks with raid-6.
> I don't like the fact that a 2nd disk-failure can kill a raid-10 when both 
> disks are in the same mirror-set.
> 
> gdisk automatically aligns on 2048 sector boundaries, that is more then enough 
> for the 4k-sectors and the block/stripe sizes employed by the raid-controller.

Yes, this is for UEFI booting, thanks.

I tried to migrate to GPT/BIOS-booting today but failed. It seems my
mainboard/BIOS has problems detecting that ... I vaguely remember this
from trying it back then.

So I am back on a freshly partitioned and formatted SSD with plain old
MBR now.

I also wanted to partition the SSD according to the Erase Block Size of
6144 kB by this way ... dunno if this is still needed or has any real
speed benefits.

Maybe I take another approach to migrate to UEFI/GPT in the next days,
now that I have my rsynced filesystems at hand (I got rid of more LVs
and stuff so it gets pretty slim now).


> UUIDs, I believe, do work natively. And those are stored inside the partition 
> itself. Which means they should also work. But are not as easy to locate. (eg. 
> you don't specify them yourself)

Yep. LABELs are human readable ... big advantage.

> See the partitioning on my server above.
> It boots using BIOS as I haven't been able to boot Xen using UEFI yet.

So then the EFI partition is useless ...

> Support should be there now, but not been able to test that yet.
> GPT is supported by grub-1 (and grub2) and with the MBR-support inside GPT, 
> booting works from BIOS/MBR.

... if your BIOS isn't crappy ;-)

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 09:03:08 PM Stefan G. Weichinger wrote:
> Am 04.05.2014 20:40, schrieb J. Roeleveld:
> > On Sunday, May 04, 2014 03:07:28 PM Stefan G. Weichinger wrote:
> >> Oh, yes, I like ZFS and its features and used it in some cases already.
> >> But I didn't yet take the step to set up ZFS-root on my work machines.
> > 
> > I haven't yet, but it's on the list of items to look into at some point.
> > I'd like to know if, with ZFS, it is possible to create block-devices like
> > LVs which I can then attach to VMs. Or if I have to use files instead.
> I think you would have to use files on top of ZFS ... but I am not
> up-to-date in that area.
> 
> People run KVM-hosts with storage on ZFS ... nice with the snapshots etc ...

Does KVM support running snapshots?
Eg. also storing the memory and registers?
I have not found any indication that KVM supports that.

Without that, KVM is useless to me.


> > That's a hardware raid device with 4 * 3TB disks with raid-6.
> > I don't like the fact that a 2nd disk-failure can kill a raid-10 when both
> > disks are in the same mirror-set.
> > 
> > gdisk automatically aligns on 2048 sector boundaries, that is more then
> > enough for the 4k-sectors and the block/stripe sizes employed by the
> > raid-controller.
> Yes, this is for UEFI booting, thanks.
> 
> I tried to migrate to GPT/BIOS-booting today but failed. It seems my
> mainboard/BIOS has problems detecting that ... I vaguely remember this
> from trying it back then.

I thought the MBR info that's possible with GPT would make it possible with 
any BIOS?
Provided the /boot partition is early enough on the disk.

> So I am back on a freshly partitioned and formatted SSD with plain old
> MBR now.
> 
> I also wanted to partition the SSD according to the Erase Block Size of
> 6144 kB by this way ... dunno if this is still needed or has any real
> speed benefits.

Again, I would expect current tools should do that automagically?

> Maybe I take another approach to migrate to UEFI/GPT in the next days,
> now that I have my rsynced filesystems at hand (I got rid of more LVs
> and stuff so it gets pretty slim now).

Less LVs is simpler. More is more flexible.

> > UUIDs, I believe, do work natively. And those are stored inside the
> > partition itself. Which means they should also work. But are not as easy
> > to locate. (eg. you don't specify them yourself)
> 
> Yep. LABELs are human readable ... big advantage.

I would like LABELs to be supported directly by the kernel.

> > See the partitioning on my server above.
> > It boots using BIOS as I haven't been able to boot Xen using UEFI yet.
> 
> So then the EFI partition is useless ...

True, but I leave it there as repartitioning requires extended downtime. And I 
do occasionally test new versions to see if I can get it to work.

> > Support should be there now, but not been able to test that yet.
> > GPT is supported by grub-1 (and grub2) and with the MBR-support inside
> > GPT,
> > booting works from BIOS/MBR.
> 
> ... if your BIOS isn't crappy ;-)

Try updating? :)

Seriously, do you have the following:

***
# man gdisk
artemis ~ # gdisk -l /dev/sda
GPT fdisk (gdisk) version 0.8.8

Partition table scan:
  MBR: protective

***

That last line of what I copied is the bit that should make it possible.

Also, the /boot partition needs to have the mbr-boot flag (or whatever it's 
called) enabled.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 04.05.2014 21:22, schrieb J. Roeleveld:
> On Sunday, May 04, 2014 09:03:08 PM Stefan G. Weichinger wrote:

>> People run KVM-hosts with storage on ZFS ... nice with the
>> snapshots etc ...
> 
> Does KVM support running snapshots? Eg. also storing the memory and
> registers? I have not found any indication that KVM supports that.
> 
> Without that, KVM is useless to me.

KVM != ZFS ... you could use ZFS with other hypervisors.

AFAI understand you can take snapshots using libvirt (which in turn
controls the underlying qemu-kvm etc).

> I thought the MBR info that's possible with GPT would make it
> possible with any BIOS? Provided the /boot partition is early enough
> on the disk.

I had no separate /boot ... just sda1 as "BIOS boot partition",
unformatted and with partition type ef02 (and sda2 as my gentoo-root).

>> I also wanted to partition the SSD according to the Erase Block
>> Size of 6144 kB by this way ... dunno if this is still needed or
>> has any real speed benefits.
> 
> Again, I would expect current tools should do that automagically?

Maybe.

>> Maybe I take another approach to migrate to UEFI/GPT in the next
>> days, now that I have my rsynced filesystems at hand (I got rid of
>> more LVs and stuff so it gets pretty slim now).
> 
> Less LVs is simpler. More is more flexible.

Exactly. But it was time to clean up a bit.

>>> booting works from BIOS/MBR.
>> 
>> ... if your BIOS isn't crappy ;-)
> 
> Try updating? :)

Never found an update for this box.

-> DMI: Hewlett-Packard HP Elite 7300 Series MT/2AB5, BIOS 7.12 10/12/2011

hints welcome ;-)


> Seriously, do you have the following:
> 
> *** # man gdisk artemis ~ # gdisk -l /dev/sda GPT fdisk (gdisk)
> version 0.8.8
> 
> Partition table scan: MBR: protective
> 
> ***
> 
> That last line of what I copied is the bit that should make it
> possible.
> 
> Also, the /boot partition needs to have the mbr-boot flag (or
> whatever it's called) enabled.

Thanks, but I had that as well ... I will skip that stuff and try going
the UEFI-way later this day, if my customers allow ...

Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Monday, May 05, 2014 08:10:55 AM Stefan G. Weichinger wrote:
> Am 04.05.2014 21:22, schrieb J. Roeleveld:
> > On Sunday, May 04, 2014 09:03:08 PM Stefan G. Weichinger wrote:
> >> People run KVM-hosts with storage on ZFS ... nice with the
> >> snapshots etc ...
> > 
> > Does KVM support running snapshots? Eg. also storing the memory and
> > registers? I have not found any indication that KVM supports that.
> > 
> > Without that, KVM is useless to me.
> 
> KVM != ZFS ... you could use ZFS with other hypervisors.

I know, comment was about KVM. ZFS is on my list, but I don't want to use files 
for the VMs.

> AFAI understand you can take snapshots using libvirt (which in turn
> controls the underlying qemu-kvm etc).

And I found no evidence on memory taken along with the snapshot. Only disk. 
(unfortunately)
I will start a new thread about this with the KVM mailing lists later today.

> > I thought the MBR info that's possible with GPT would make it
> > possible with any BIOS? Provided the /boot partition is early enough
> > on the disk.
> 
> I had no separate /boot ... just sda1 as "BIOS boot partition",
> unformatted and with partition type ef02 (and sda2 as my gentoo-root).

I have 8300 for my /boot. If I would have that on " / ", I would give that 
8300 as type.

> >>> booting works from BIOS/MBR.
> >> 
> >> ... if your BIOS isn't crappy ;-)
> > 
> > Try updating? :)
> 
> Never found an update for this box.
> 
> -> DMI: Hewlett-Packard HP Elite 7300 Series MT/2AB5, BIOS 7.12 10/12/2011
> 
> hints welcome ;-)

I use seperate components for the desktops and a " local " (30 - 40 minute 
drive) supplier for server hardware.

I find that cheaper then using the big boys like HP or Dell.

> > Also, the /boot partition needs to have the mbr-boot flag (or
> > whatever it's called) enabled.
> 
> Thanks, but I had that as well ... I will skip that stuff and try going
> the UEFI-way later this day, if my customers allow ...

Good luck (with your customers)

--
Joost

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Monday, May 05, 2014 08:52:08 AM J. Roeleveld wrote:
> On Monday, May 05, 2014 08:10:55 AM Stefan G. Weichinger wrote:
> > Am 04.05.2014 21:22, schrieb J. Roeleveld:
> > > On Sunday, May 04, 2014 09:03:08 PM Stefan G. Weichinger wrote:
> > >> People run KVM-hosts with storage on ZFS ... nice with the
> > >> snapshots etc ...
> > > 
> > > Does KVM support running snapshots? Eg. also storing the memory and
> > > registers? I have not found any indication that KVM supports that.
> > > 
> > > Without that, KVM is useless to me.
> > 
> > KVM != ZFS ... you could use ZFS with other hypervisors.
> 
> I know, comment was about KVM. ZFS is on my list, but I don't want to use
> files for the VMs.
> 
> > AFAI understand you can take snapshots using libvirt (which in turn
> > controls the underlying qemu-kvm etc).
> 
> And I found no evidence on memory taken along with the snapshot. Only disk.
> (unfortunately)
> I will start a new thread about this with the KVM mailing lists later today.

Update:
I just found the following page (didn't show up last year using Google):
http://redes-privadas-virtuales.blogspot.nl/2011/03/taking-snapshots-on-kvm-with-libvirt.html

Memory-snapshot appears possible as well. Time to start testing this.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 05.05.2014 09:09, schrieb J. Roeleveld:

> Update:
> I just found the following page (didn't show up last year using Google):
> http://redes-privadas-virtuales.blogspot.nl/2011/03/taking-snapshots-on-kvm-with-libvirt.html
> 
> Memory-snapshot appears possible as well. Time to start testing this.

Yes, I also found that link this morning. Let us know what you find out.

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Monday, May 05, 2014 09:48:50 AM Stefan G. Weichinger wrote:
> Am 05.05.2014 09:09, schrieb J. Roeleveld:
> > Update:
> > I just found the following page (didn't show up last year using Google):
> > http://redes-privadas-virtuales.blogspot.nl/2011/03/taking-snapshots-on-kv
> > m-with-libvirt.html
> > 
> > Memory-snapshot appears possible as well. Time to start testing this.
> 
> Yes, I also found that link this morning. Let us know what you find out.

That will take some time. Earliest moment I might have the time would be June 
as I need to find a decent howto on setting up KVM along with libvirt on a test 
system.

--
Joost

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 05.05.2014 11:14, schrieb J. Roeleveld:

> That will take some time. Earliest moment I might have the time would be June 
> as I need to find a decent howto on setting up KVM along with libvirt on a test 
> system.

I do that for a living :-P

just kidding ...

S

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Monday, May 05, 2014 01:25:52 PM Stefan G. Weichinger wrote:
> Am 05.05.2014 11:14, schrieb J. Roeleveld:
> > That will take some time. Earliest moment I might have the time would be
> > June as I need to find a decent howto on setting up KVM along with
> > libvirt on a test system.
> 
> I do that for a living :-P

In that case, got a decent howto?

I use virtualization on the server (don't like having one instance do 
everything) and for testing software I need for my paying job.
Currently my lab machine runs XCP, but I want to move that to either native 
Xen or KVM.

> just kidding ...

Hehe :)

--
Joost

[gentoo-user] Re: Qemu/KVM

[Stefan G. Weichinger]

Am 05.05.2014 14:27, schrieb J. Roeleveld:
> On Monday, May 05, 2014 01:25:52 PM Stefan G. Weichinger wrote:
>> Am 05.05.2014 11:14, schrieb J. Roeleveld:
>>> That will take some time. Earliest moment I might have the time would be
>>> June as I need to find a decent howto on setting up KVM along with
>>> libvirt on a test system.
>>
>> I do that for a living :-P
> 
> In that case, got a decent howto?

not really ...

https://wiki.gentoo.org/wiki/QEMU

http://www.linux-kvm.org/page/KvmOnGentoo

for a start ... but I pull my howto together from various sources.

basically compile KVM support into your kernel, set QEMU-variables in
make.conf ... emerge qemu ... then libvirt with USE="qemu" ... it isn't
that hard to make it work (at least now that I did it 10 times or so).

Stefan

Re: [gentoo-user] Re: Qemu/KVM

[J. Roeleveld]

On 5 May 2014 16:12:43 CEST, "Stefan G. Weichinger" <lists@xunil.at> wrote:
>Am 05.05.2014 14:27, schrieb J. Roeleveld:
>> On Monday, May 05, 2014 01:25:52 PM Stefan G. Weichinger wrote:
>>> Am 05.05.2014 11:14, schrieb J. Roeleveld:
>>>> That will take some time. Earliest moment I might have the time
>would be
>>>> June as I need to find a decent howto on setting up KVM along with
>>>> libvirt on a test system.
>>>
>>> I do that for a living :-P
>> 
>> In that case, got a decent howto?
>
>not really ...
>
>https://wiki.gentoo.org/wiki/QEMU
>
>http://www.linux-kvm.org/page/KvmOnGentoo
>
>for a start ... but I pull my howto together from various sources.
>
>basically compile KVM support into your kernel, set QEMU-variables in
>make.conf ... emerge qemu ... then libvirt with USE="qemu" ... it isn't
>that hard to make it work (at least now that I did it 10 times or so).
>
>Stefan

Thanks.

I find Xen easy to set up as well. For the same reason (done it several times).

What I really need is a VM server with a decent frontend (multiplatform and where I can specify which user can access which VM and have some restrictions on network settings for new VMs)
And where I can specify fast (SSD) stprage for the memory snapshots and normal (Spinning) storage for the disks (Using LVM or ZFS snapshots)

I might end up writing my own frontend, but libvirt would then be a usefull API.

--
Joost
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 05.05.2014 08:52, schrieb J. Roeleveld:

>> Thanks, but I had that as well ... I will skip that stuff and try going
>> the UEFI-way later this day, if my customers allow ...
> 
> Good luck (with your customers)

customers quiet so far ... so I fired up a live cd and re-did all this.

long story short: system boots from GPT-partitioned SSD via UEFI ...

I still have to test installing and upgrading kernels in this new setup
but so far it looks promising.

Stefan

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 09:53:35 AM Peter Humphrey wrote:
> On Saturday 03 May 2014 23:04:49 Stefan G. Weichinger wrote:
> > I spent nearly the whole day digging around this issue ...
> 
> You did better than I did recently: I spent four days at it.

For mission-critical systems, I would have done a clean re-install already 
with data copied back from a backup. More then 24 hours is a deadline.
For non-critical, I am willing to invest more time.

> > I wonder if I speak for more users when I say that all this is kind of
> > confusing sometimes ...
> 
> I'm with you there, Stefan. I find the whole RAID and LVM area deeply
> mysterious, and the docs I've seen only say what to do, not why. I'd still
> like to find a proper explanation of how it all works.

I used to have a howto bookmarked that gave more detail then the current step-
by-step examples. Unfortunately, that whole website disappeared about 5 or 6 
years ago.

Maybe check the old thread where Dale started with LVM. There is a lot of 
detail in there.

> > I am not so far to skip the initramfs -> I don't *know* that, I just
> > tested removing the line from grub2 and it failed finding the root-fs.
> 
> I've never had an initramfs, seeing no need in my case to keep /usr on its
> own partition.

Same here, until that whole mess started and I ended up using an initramfs.
At the same time, I moved everything except /boot onto RAID-0 for the 
desktops.

> > For booting from a plain partition on an SSD I think I shouldn't need an
> > initramfs? Does it have to do with MBR/GPT as well (the SSD is
> > still/again MBR, as UEFI booting broke badly for me back then) ?
> 
> As far as I know, the only thing that /requires/ an initramfs is having a
> separate /usr. And I can't help you with GPT or UEFI - sorry.

A seperate " /usr " or " / " on LVM.

> > Maybe I learn more soon ;-)
> 
> I sometimes say that life is just one long journey of discovery  :-)

It is. And that's what makes life interesting.

--
Joost

Re: [gentoo-user] boot problems

[Peter Humphrey]

On Sunday 04 May 2014 12:37:02 J. Roeleveld wrote:
> On Sunday, May 04, 2014 09:53:35 AM Peter Humphrey wrote:
> > On Saturday 03 May 2014 23:04:49 Stefan G. Weichinger wrote:
> > > I spent nearly the whole day digging around this issue ...
> > 
> > You did better than I did recently: I spent four days at it.
> 
> For mission-critical systems, I would have done a clean re-install already
> with data copied back from a backup. More then 24 hours is a deadline.

That's what I was doing when I discovered the IRQ16 misbehaviour. After that 
it was a matter of finding the cause. As it happens, it went away again by 
itself. Makes me wonder how much more life to expect from this motherboard.

--->8

> I used to have a howto bookmarked that gave more detail then the current
> step- by-step examples. Unfortunately, that whole website disappeared about
> 5 or 6 years ago.
> 
> Maybe check the old thread where Dale started with LVM. There is a lot of
> detail in there.

Hmm...don't remember that. I'll see if I can find it - thanks.

--->8

> > As far as I know, the only thing that /requires/ an initramfs is having a
> > separate /usr. And I can't help you with GPT or UEFI - sorry.
> 
> A seperate " /usr " or " / " on LVM.

Oh yes, of course. I missed that one. I shouldn't have, because that's the 
reason my / is on RAID-1 but not LVM - to avoid needing an initramfs.

-- 
Regards
Peter

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 02:22:16 PM Peter Humphrey wrote:
> On Sunday 04 May 2014 12:37:02 J. Roeleveld wrote:
> > On Sunday, May 04, 2014 09:53:35 AM Peter Humphrey wrote:
> > > On Saturday 03 May 2014 23:04:49 Stefan G. Weichinger wrote:
> > > > I spent nearly the whole day digging around this issue ...
> > > 
> > > You did better than I did recently: I spent four days at it.
> > 
> > For mission-critical systems, I would have done a clean re-install already
> > with data copied back from a backup. More then 24 hours is a deadline.
> 
> That's what I was doing when I discovered the IRQ16 misbehaviour. After that
> it was a matter of finding the cause. As it happens, it went away again by
> itself. Makes me wonder how much more life to expect from this motherboard.
> 
> --->8

Not necessarily a dead motherboard. Not gotten round to replying on your email 
in that thread yet, will do so in a few moments.

> > I used to have a howto bookmarked that gave more detail then the current
> > step- by-step examples. Unfortunately, that whole website disappeared
> > about
> > 5 or 6 years ago.
> > 
> > Maybe check the old thread where Dale started with LVM. There is a lot of
> > detail in there.
> 
> Hmm...don't remember that. I'll see if I can find it - thanks.
> 
> --->8

Let me know if you have difficulty locating it, I might be able to find the 
actual date/time of the first post for you. (Or even forward the whole thread 
in a zip-file directly to your email if requested)

> > > As far as I know, the only thing that /requires/ an initramfs is having
> > > a
> > > separate /usr. And I can't help you with GPT or UEFI - sorry.
> > 
> > A seperate " /usr " or " / " on LVM.
> 
> Oh yes, of course. I missed that one. I shouldn't have, because that's the
> reason my / is on RAID-1 but not LVM - to avoid needing an initramfs.

I figured that as I am forced to redo my machine with an initramfs anyway, I 
might as well put "/" on RAID as well.
That's why I had "/usr" seperately in the past, so I had that part on RAID+LVM 
and still able to boot.

That's a use-case that is no longer accepted as "normal"

--
Joost

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 945 bytes --]

On Sat, 03 May 2014 20:40:47 +0200, J. Roeleveld wrote:

> > * SystemRescueCD and the Gentoo minimal installation CD both start
> > any raid arrays they find and apply their own names to them. It is
> > then impossible, or so I thought, to resume an interrupted
> > installation process. Of course, all I had to do was "mdadm
> > --stop /dev/md127" etc.  
> 
> Yes, I noticed that annoyance myself. I would much prefer it to default
> to more logical names.

ISTR that's because of the hostname stored in the RAID, so when you plug
the array into another computer, it doesn't clobber any existing array
names. I had this happen when transplanting an array to a new system.
There is a sequence of commands to reset the names but it was a while ago
and all I remember is that the sequence started with "man lvm".


-- 
Neil Bothwick

Q   How many screws are there in a lesbians coffin?
A   None. It's all tongue and groove.

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

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Sunday, May 04, 2014 01:15:51 PM Neil Bothwick wrote:
> On Sat, 03 May 2014 20:40:47 +0200, J. Roeleveld wrote:
> > > * SystemRescueCD and the Gentoo minimal installation CD both start
> > > any raid arrays they find and apply their own names to them. It is
> > > then impossible, or so I thought, to resume an interrupted
> > > installation process. Of course, all I had to do was "mdadm
> > > --stop /dev/md127" etc.
> > 
> > Yes, I noticed that annoyance myself. I would much prefer it to default
> > to more logical names.
> 
> ISTR that's because of the hostname stored in the RAID, so when you plug
> the array into another computer, it doesn't clobber any existing array
> names. I had this happen when transplanting an array to a new system.
> There is a sequence of commands to reset the names but it was a while ago
> and all I remember is that the sequence started with "man lvm".

Actually, the steps are:
1) invalidate the RAID
2) create a new, broken, RAID using the invalidatd disk
3) copy data to new RAID
4) delete md127
5) add other disk from md127 to md1

Or something to that effect.
A "rename" option or even a "I_KNOW_WHAT_I_AM_DOING_JUST_KEEP_THE_SAME_NAME" 
option, which would be enabled by default when booting with sysresccd, would 
be nice.

--
Joost

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 1264 bytes --]

On Sun, 04 May 2014 20:48:05 +0200, J. Roeleveld wrote:

> > ISTR that's because of the hostname stored in the RAID, so when you
> > plug the array into another computer, it doesn't clobber any existing
> > array names. I had this happen when transplanting an array to a new
> > system. There is a sequence of commands to reset the names but it was
> > a while ago and all I remember is that the sequence started with "man
> > lvm".  
> 
> Actually, the steps are:
> 1) invalidate the RAID
> 2) create a new, broken, RAID using the invalidatd disk
> 3) copy data to new RAID
> 4) delete md127
> 5) add other disk from md127 to md1

It was far simpler than that and did not involve copying data. You can
change the hostname using one of the mdadm management commands AFAIR (I
meant man mdadm not man lvm of course).

> A "rename" option or even a
> "I_KNOW_WHAT_I_AM_DOING_JUST_KEEP_THE_SAME_NAME" option, which would be
> enabled by default when booting with sysresccd, would be nice.

You can specify device names for md devices on the kernel command line,
but an option to ignore the hostname would be handy and just keep the
current naming system, at your own risk, would be handy.


-- 
Neil Bothwick

.sig a .sog of sixpence.

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

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

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

Am 04.05.2014 22:09, schrieb Neil Bothwick:

> It was far simpler than that and did not involve copying data. You
> can change the hostname using one of the mdadm management commands
> AFAIR (I meant man mdadm not man lvm of course).

option "--update=homehost" ?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJTZywvAAoJEClcuD1V0Pzmui0P/34bjDa4m3CFHaeuJUkSLct/
ruIrm3smXulKXk/Vx95/ks0NFRBluD5Zb6+AUPS8j47nBhg4g8dxKQcPFxYdj/E3
b+SKcOFn7bwA+GVBJuh82SnNH7pRaQ8FSJ1P76VFJuHY2Gu1R1n8H1y2jB/GP5b3
vnlQNDmmls2Ubv4xqkuwae9FprH1X60PrU/dC7oDMej2afy2tmFPG7BVoSgFd8ZL
yD6UmB7sYelYgcF3znK6eeLRtBRvf1nl4y7DHJInvwuCc11sQF2IMxIRGyFnx7Mr
O+Flzx1Aa1pI+TXwOgBVMurmMgY+IjzA08p0GzGnj89u1+8R0QmYEV4RJZthhnBX
agSaaReZZSYvC+J+r1cbsW5TPj58ixydnkTP/MfGPJpEPsYivzvKhAIdT8JazDZ2
Nwv/hiMpQjM3o0iqFCt5uYjlmG3lC1LXYdm6DK4HZJz0V7mee/5SAqMQDC//ESwS
JNxfMPN5zZoPR5zOnq29fVSuilMjnXdCcJdDk/nWiP1Jb6DWw1z+pya3CvIo3vrb
DH34ppI7DVqKIq+MfTKVUH4Xx8ywtw+JziPnO+7YaulGR8RAQkWdh2TQo8AOR+Md
XK97Cw9zPH6DU1rVmlVU295eLD4f4+ykED1JH/ZVtSmYURaNc+WjX1pJIekGCnf3
tOkY3BBHCSci8Y5ynlDh
=VsTT
-----END PGP SIGNATURE-----

Re: [gentoo-user] boot problems

[J. Roeleveld]

On Monday, May 05, 2014 08:14:07 AM Stefan G. Weichinger wrote:
> Am 04.05.2014 22:09, schrieb Neil Bothwick:
> > It was far simpler than that and did not involve copying data. You
> > can change the hostname using one of the mdadm management commands
> > AFAIR (I meant man mdadm not man lvm of course).
> 
> option "--update=homehost" ?

Hmm... never noticed that one myself.

Thank you.

Next step, find a way to avoid auto-build of the mdadm by the kernel.
When I find the time...

--
Joost

Re: [gentoo-user] boot problems

[Neil Bothwick]

[-- Attachment #1: Type: text/plain, Size: 556 bytes --]

On Mon, 05 May 2014 08:14:07 +0200, Stefan G. Weichinger wrote:

> > It was far simpler than that and did not involve copying data. You
> > can change the hostname using one of the mdadm management commands
> > AFAIR (I meant man mdadm not man lvm of course).  
> 
> option "--update=homehost" ?

That's the one, I did a quick search  of the mdadm man page before but
was looking for "hostname", which didn't help.


-- 
Neil Bothwick

The trouble with doing something right the first time is that nobody
appreciates how difficult it was.

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

Re: [gentoo-user] boot problems

[Tom H]

On Sun, May 4, 2014 at 2:48 PM, J. Roeleveld <joost@antarean.org> wrote:
> On Sunday, May 04, 2014 01:15:51 PM Neil Bothwick wrote:
>> On Sat, 03 May 2014 20:40:47 +0200, J. Roeleveld wrote:
>>>>
>>>> * SystemRescueCD and the Gentoo minimal installation CD both start
>>>> any raid arrays they find and apply their own names to them. It is
>>>> then impossible, or so I thought, to resume an interrupted
>>>> installation process. Of course, all I had to do was "mdadm
>>>> --stop /dev/md127" etc.
>>>
>>> Yes, I noticed that annoyance myself. I would much prefer it to default
>>> to more logical names.
>>
>> ISTR that's because of the hostname stored in the RAID, so when you plug
>> the array into another computer, it doesn't clobber any existing array
>> names. I had this happen when transplanting an array to a new system.
>> There is a sequence of commands to reset the names but it was a while ago
>> and all I remember is that the sequence started with "man lvm".
>
> Actually, the steps are:
> 1) invalidate the RAID
> 2) create a new, broken, RAID using the invalidatd disk
> 3) copy data to new RAID
> 4) delete md127
> 5) add other disk from md127 to md1
>
> Or something to that effect.
> A "rename" option or even a "I_KNOW_WHAT_I_AM_DOING_JUST_KEEP_THE_SAME_NAME"
> option, which would be enabled by default when booting with sysresccd, would
> be nice.

There's an mdadm option to change the hostname ("--homehost=<hostname>
--update=homehost") but in the case of 0.9 metadata it'll change the
uuid.

You can also set "HOMEHOST=none" in mdadm.conf before creating an
array, but I'm not sure how the md device will be numbered if you then
assemble it on a box with "HOMEHOST=system" set.

Re: [gentoo-user] boot problems

[Tom H]

On Sat, May 3, 2014 at 8:08 AM, Stefan G. Weichinger <lists@xunil.at> wrote:
>
> Box boots now with kernel 3.14.1 and with commented LVs in fstab.
>
> When I login and check there are no mdadm-raids assembled.
>
> When I "mdadm -A --scan" they get correctly assembled:
>
> # cat /proc/mdstat
> Personalities : [raid1]
> md4 : active raid1 sdb3[0] sdc6[2]
>       52395904 blocks super 1.2 [2/2] [UU]
>
> md1 : active raid1 sdb6[0] sdc3[1]
>       623963072 blocks [2/2] [UU]
>
> (Don't ask for the strange setup with sdb3/sdc6 and sdb6/sdc3,
> historically grown somehow ...)
>
> "vgchange -ay" then gets me all my LVs. Great so far.
>
> So the question is, what part of the whole setup should now assemble the
> arrays?
>
> I think, dracut, right?
>
> So I will now test booting with these funny "rd.auto" kernel line
> parameters ...
>
> Has mdadm.service to be enabled as well? Or is that redundant in a way?

Try "rd.md.uuid=<...>" for all the md arrays that you want to bring up.

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 03.05.2014 14:52, schrieb Tom H:

>> Has mdadm.service to be enabled as well? Or is that redundant in a way?
> 
> Try "rd.md.uuid=<...>" for all the md arrays that you want to bring up.

Tested that right now, no mdadm-array assembled!

Thanks anyway, Stefan

Re: [gentoo-user] boot problems

[Mark Pariente]

I had a similar issue after upgrading to sys-apps/systemd-212-r3. Turned
out it was due to udev rules.d directory location. I still had a bunch
of udev rules under /usr/lib64/udev/rules.d and it looks like the new
systemd/udev is only looking at rules under /lib/udev/.

The particular issue was that the RAID device (/dev/md127 in my case)
was no longer auto-assembled because the udev rules installed by the
mdadm package were still in /usr/lib64/udev/rules.d instead
of /lib/udev. This required me to re-emerge mdadm which then installed
the udev rules into /lib/udev.

Unfortunately there were many other such packages affected. I just ran 
'equery belongs' on all files under /usr/lib64/udev/rules.d and
re-emerged all respective packages which fixed all issues.

--Mark


On Sat, 2014-05-03 at 12:12 +0200, Stefan G. Weichinger wrote:
> I see boot problems on two of my machines ... very likely related to
> lvm2 as far as I can tell so far.
> 
> Downgraded and even disabled systemd, re-emerged stuff ... I am still at
> trying to find the reason. Maybe it is with the mdadm raids also ...
> 
> I just want to ask you other gentoo users if anyone else hit this issue
> since around yesterday? Maybe someone else also scratches his head
> already ...
> 
> Stefan
> 

Re: [gentoo-user] boot problems

[Stefan G. Weichinger]

Am 04.05.2014 03:56, schrieb Mark Pariente:
> I had a similar issue after upgrading to sys-apps/systemd-212-r3. Turned
> out it was due to udev rules.d directory location. I still had a bunch
> of udev rules under /usr/lib64/udev/rules.d and it looks like the new
> systemd/udev is only looking at rules under /lib/udev/.
> 
> The particular issue was that the RAID device (/dev/md127 in my case)
> was no longer auto-assembled because the udev rules installed by the
> mdadm package were still in /usr/lib64/udev/rules.d instead
> of /lib/udev. This required me to re-emerge mdadm which then installed
> the udev rules into /lib/udev.
> 
> Unfortunately there were many other such packages affected. I just ran 
> 'equery belongs' on all files under /usr/lib64/udev/rules.d and
> re-emerged all respective packages which fixed all issues.

There is a nice command for finding and rebuilding the affected packages in

https://bugs.gentoo.org/show_bug.cgi?id=509492

(in Comment 6)

I just did that one of my thinkpads as I wasn't able to properly login
anymore (/home encrypted via cryptsetup).

Stefan