[gentoo-user] [Gentoo install] Disk full at 35%?

[gentoo-user] [Gentoo install] Disk full at 35%?

[Hilco Wijbenga]

Hi all,

It's been quite a few years but I decided to try another Gentoo
install (on a VirtualBox instance). I wanted to try out some new
things...

I created a ton of partitions including /usr (I want to see if I can
get that to work), /portage, and /distfiles. The idea was to mount
/portage on top of /usr and /distfiles on top of /portage. This all
works fine.

However, when I try to extract the Portage snapshot, I get "No space
left on device" a long way into the untar process. According to df
/portage (i.e. /mnt/gentoo/usr/portage) is only 35% full. In fact, not
a single partition or mount is even close to full (except for
/mnt/static, the DVD).

If I untar directly to /usr (after unmounting /portage), everything
works fine. If I then try to copy or move to /portage, I get the "No
space left on device" again. And at the same place.

Does anyone know what's going on here? I didn't realize I was doing
such strange things. At least not this early on. :-)

Cheers,
Hilco

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[covici]

Hilco Wijbenga <hilco.wijbenga@gmail.com> wrote:

> Hi all,
> 
> It's been quite a few years but I decided to try another Gentoo
> install (on a VirtualBox instance). I wanted to try out some new
> things...
> 
> I created a ton of partitions including /usr (I want to see if I can
> get that to work), /portage, and /distfiles. The idea was to mount
> /portage on top of /usr and /distfiles on top of /portage. This all
> works fine.
> 
> However, when I try to extract the Portage snapshot, I get "No space
> left on device" a long way into the untar process. According to df
> /portage (i.e. /mnt/gentoo/usr/portage) is only 35% full. In fact, not
> a single partition or mount is even close to full (except for
> /mnt/static, the DVD).
> 
> If I untar directly to /usr (after unmounting /portage), everything
> works fine. If I then try to copy or move to /portage, I get the "No
> space left on device" again. And at the same place.
> 
> Does anyone know what's going on here? I didn't realize I was doing
> such strange things. At least not this early on. :-)

See if  you are out of inodes. The only way to get the inodes that I am
aware of is to  debugfs to the partition and do stat from within -- if
there is a better way please let me know.  But why not use 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] [Gentoo install] Disk full at 35%?

[Hilco Wijbenga]

On 20 August 2011 18:52,  <covici@ccs.covici.com> wrote:
> Hilco Wijbenga <hilco.wijbenga@gmail.com> wrote:
>
>> Hi all,
>>
>> It's been quite a few years but I decided to try another Gentoo
>> install (on a VirtualBox instance). I wanted to try out some new
>> things...
>>
>> I created a ton of partitions including /usr (I want to see if I can
>> get that to work), /portage, and /distfiles. The idea was to mount
>> /portage on top of /usr and /distfiles on top of /portage. This all
>> works fine.
>>
>> However, when I try to extract the Portage snapshot, I get "No space
>> left on device" a long way into the untar process. According to df
>> /portage (i.e. /mnt/gentoo/usr/portage) is only 35% full. In fact, not
>> a single partition or mount is even close to full (except for
>> /mnt/static, the DVD).
>>
>> If I untar directly to /usr (after unmounting /portage), everything
>> works fine. If I then try to copy or move to /portage, I get the "No
>> space left on device" again. And at the same place.
>>
>> Does anyone know what's going on here? I didn't realize I was doing
>> such strange things. At least not this early on. :-)
>
> See if  you are out of inodes. The only way to get the inodes that I am
> aware of is to  debugfs to the partition and do stat from within -- if
> there is a better way please let me know.  But why not use lvm?

Yes, df -i says /portage is out of inodes. I've never run into that
before. I reran mke2fs to increase the inode count and that fixed
things.

Would LVM somehow prevent these sort of things from happening? LVM
doesn't affect inode usage, does it? What exactly are the advantages
of LVM? Is it just that it's easier to resize LVM partitions after the
fact? (That would, of course, already be very useful.)

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Nilesh Govindarajan]

On 08/21/2011 09:00 AM, Hilco Wijbenga wrote:
> Yes, df -i says /portage is out of inodes. I've never run into that
> before. I reran mke2fs to increase the inode count and that fixed
> things.
> 

Sorry for the drop in, but I never knew that mke2fs can increase the
number of inodes!
I think I'll now place the portage tree on an ext2 disk image to speed
up things, / has got fragmented badly due to portage tree :-\
Thanks man!

-- 
Nilesh Govindarajan
http://nileshgr.com

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Hilco Wijbenga]

On 20 August 2011 21:21, Nilesh Govindarajan <contact@nileshgr.com> wrote:
> On 08/21/2011 09:00 AM, Hilco Wijbenga wrote:
>> Yes, df -i says /portage is out of inodes. I've never run into that
>> before. I reran mke2fs to increase the inode count and that fixed
>> things.
>
> Sorry for the drop in, but I never knew that mke2fs can increase the
> number of inodes!
> I think I'll now place the portage tree on an ext2 disk image to speed
> up things, / has got fragmented badly due to portage tree :-\

Well, for the record, I'm not using ext2 but ext3 (mke2fs -j).
Although, now that I think about it, I suppose there's not much point
in having the Portage tree on a journaled FS.

If you run man mke2fs, you should check out -N and -i. It was
trial-and-error (for me, anyway) to find the right number. Presumably,
-I fits in there somewhere as well. Do note that it only works when
creating the FS, you can't change the inode count dynamically.

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1267 bytes --]

On Sunday 21 Aug 2011 05:47:16 Hilco Wijbenga wrote:
> On 20 August 2011 21:21, Nilesh Govindarajan <contact@nileshgr.com> wrote:
> > On 08/21/2011 09:00 AM, Hilco Wijbenga wrote:
> >> Yes, df -i says /portage is out of inodes. I've never run into that
> >> before. I reran mke2fs to increase the inode count and that fixed
> >> things.
> > 
> > Sorry for the drop in, but I never knew that mke2fs can increase the
> > number of inodes!
> > I think I'll now place the portage tree on an ext2 disk image to speed
> > up things, / has got fragmented badly due to portage tree :-\
> 
> Well, for the record, I'm not using ext2 but ext3 (mke2fs -j).
> Although, now that I think about it, I suppose there's not much point
> in having the Portage tree on a journaled FS.
> 
> If you run man mke2fs, you should check out -N and -i. It was
> trial-and-error (for me, anyway) to find the right number. Presumably,
> -I fits in there somewhere as well. Do note that it only works when
> creating the FS, you can't change the inode count dynamically.

I've never run out of inodes, even on small partitions.  I just let ext4 make 
a fs with its default settings.  Is there a magic formula to determine how 
many inodes are optimal?
-- 
Regards,
Mick

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

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Adam Carter]

>> creating the FS, you can't change the inode count dynamically.
>
> I've never run out of inodes, even on small partitions.  I just let ext4 make
> a fs with its default settings.  Is there a magic formula to determine how
> many inodes are optimal?

Some FSes allocate inodes as required. I know btrfs does this and i
think reiser does it too.

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Alan McKinnon]

On Sun 21 August 2011 11:13:53 Mick did opine thusly:
> On Sunday 21 Aug 2011 05:47:16 Hilco Wijbenga wrote:
> > On 20 August 2011 21:21, Nilesh Govindarajan 
<contact@nileshgr.com> wrote:
> > > On 08/21/2011 09:00 AM, Hilco Wijbenga wrote:
> > >> Yes, df -i says /portage is out of inodes. I've never run
> > >> into that before. I reran mke2fs to increase the inode
> > >> count and that fixed things.
> > > 
> > > Sorry for the drop in, but I never knew that mke2fs can
> > > increase the number of inodes!
> > > I think I'll now place the portage tree on an ext2 disk
> > > image to speed up things, / has got fragmented badly due to
> > > portage tree :-\> 
> > Well, for the record, I'm not using ext2 but ext3 (mke2fs -j).
> > Although, now that I think about it, I suppose there's not much
> > point in having the Portage tree on a journaled FS.
> > 
> > If you run man mke2fs, you should check out -N and -i. It was
> > trial-and-error (for me, anyway) to find the right number.
> > Presumably, -I fits in there somewhere as well. Do note that it
> > only works when creating the FS, you can't change the inode
> > count dynamically.
> I've never run out of inodes, even on small partitions.  I just let
> ext4 make a fs with its default settings.  Is there a magic formula
> to determine how many inodes are optimal?

No, there's no such formula. 

The answer to "How many inodes do I need?" is always "How many do you 
need?"


-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Andrea Conti]

> If you run man mke2fs, you should check out -N and -i. It was
> trial-and-error (for me, anyway) to find the right number.

Consider using reiserfs for /usr/portage. No real performance advantage
over ext[234], but works well with lots of small files and there's no
inode count to worry about.

In my experience the main downside of reiserfs is that fsck.reiserfs is
almost never able to recover cleanly if the filesystem metadata does get
corrupted in a non-trivial way. But for the portage snapshot this isn't
really a problem...

andrea

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Hilco Wijbenga]

On 21 August 2011 03:46, Andrea Conti <alyf@alyf.net> wrote:
>> If you run man mke2fs, you should check out -N and -i. It was
>> trial-and-error (for me, anyway) to find the right number.
>
> Consider using reiserfs for /usr/portage. No real performance advantage
> over ext[234], but works well with lots of small files and there's no
> inode count to worry about.
>
> In my experience the main downside of reiserfs is that fsck.reiserfs is
> almost never able to recover cleanly if the filesystem metadata does get
> corrupted in a non-trivial way. But for the portage snapshot this isn't
> really a problem...

I have always used ReiserFS for everything but /boot. That explains
why I never ran into the inode issue, I guess.

I'm trying to install AMD64 and the handbook says that ([1]) "JFS and
ReiserFS may work but need more testing. If you're really adventurous
you can try the other filesystems.". That didn't sound too promising
so I went with ext3. :-)

The X86 handbook doesn't have this text. Is ReiserFS on AMD64 really
only for the adventurous? Or should this warning be removed?

[1] http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?part=1&chap=4#filesystemsdesc

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Peter Humphrey]

On Sunday 21 August 2011 19:14:53 Hilco Wijbenga wrote:

> The X86 handbook doesn't have this text. Is ReiserFS on AMD64 really
> only for the adventurous?

Certainly not. It's 100% stable as far as I know.

> Or should this warning be removed?

ASAP

-- 
Rgds
Peter		Linux Counter 5290, 1994-04-23

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Alex Schuster]

Hilco Wijbenga writes:

> Yes, df -i says /portage is out of inodes. I've never run into that
> before. I reran mke2fs to increase the inode count and that fixed
> things.
> 
> Would LVM somehow prevent these sort of things from happening? LVM
> doesn't affect inode usage, does it?

AFAIK you will gain more inodes when you increase the size.

> What exactly are the advantages
> of LVM? Is it just that it's easier to resize LVM partitions after the
> fact? (That would, of course, already be very useful.)

Mainly. But it also allows you to create snapshots, I use this often to make 
backups without needing to boot from a live-cd - I can even start emerges 
meanwhile, but the backup will have the state the file system was in when 
the snapshot was taken. 
And you can create logical volumes that reside on partitions on different 
physical drives. Or you could move logical volumes from one drive or 
location to another one, while being in use all the time, without the need 
to unmount the file system.
I also like the naming scheme (/dev/<volume group>/<logical volume>/ instead 
of /dev/sd<X><N>), although you can also use file system labels so this is 
not a big problem.

	Wonko

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Andrea Conti]

>> Would LVM somehow prevent these sort of things from happening? LVM
>> doesn't affect inode usage, does it?

LVM has nothing to do with inodes. Inodes are a filesystem concept, and
filesystems do not really care about the kind of block device they
reside on. Well, generally.

> AFAIK you will gain more inodes when you increase the size.

Only because by unless you specify a value mke2fs allocates a number of
inodes proportional to the size of the filesystem, with the default
being 1 inode every 16kB (see /etc/mke2fs.conf).

But for ext[234] the number of inodes is fixed at filesystem creation,
so even if you use LVM you can't increase it by -- say -- growing the
underlying LV and then using resize2fs.

andrea

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Alex Schuster]

Andrea Conti writes:

> > AFAIK you will gain more inodes when you increase the size.
> 
> Only because by unless you specify a value mke2fs allocates a number of
> inodes proportional to the size of the filesystem, with the default
> being 1 inode every 16kB (see /etc/mke2fs.conf).
> 
> But for ext[234] the number of inodes is fixed at filesystem creation,
> so even if you use LVM you can't increase it by -- say -- growing the
> underlying LV and then using resize2fs.

So I just tried that, create a small fs, filled it until no inodes were 
left. Resized, and gained more inodes:

weird ~ # lvcreate -L 4M -n inodetest weird
  Logical volume "inodetest" created

weird ~ # mke2fs -j /dev/weird/inodetest
mke2fs 1.41.14 (22-Dec-2010)
Filesystem label=
OS type: Linux
Block size=1024 (log=0)
Fragment size=1024 (log=0)
Stride=0 blocks, Stripe width=0 blocks
1024 inodes, 4096 blocks
[...]

weird ~ # mount /dev/weird/inodetest /mnt/

weird ~ # for (( i=1; ; i++ ))
> do
>   touch "/mnt/$( printf "file %06d" $i )" || break
> done                    
touch: cannot touch `/mnt/file 001014': No space left on device

weird ~ # df -i /mnt/
Filesystem                  Inodes IUsed IFree IUse% Mounted on
/dev/mapper/weird-inodetest   1024  1024     0  100% /mnt

weird ~ # lvresize -L 16M /dev/weird/inodetest 
  Extending logical volume inodetest to 16,00 MiB
  Logical volume inodetest successfully resized

weird ~ # resize2fs /dev/weird/inodetest 
resize2fs 1.41.14 (22-Dec-2010)
Filesystem at /dev/weird/inodetest is mounted on /mnt; on-line resizing 
required
old desc_blocks = 1, new_desc_blocks = 1
Performing an on-line resize of /dev/weird/inodetest to 16384 (1k) blocks.
The filesystem on /dev/weird/inodetest is now 16384 blocks long.

weird ~ # df -i /mnt/
Filesystem                  Inodes IUsed IFree IUse% Mounted on
/dev/mapper/weird-inodetest   2048  1024  1024   50% /mnt

	Wonko

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Andrea Conti]

> Filesystem                  Inodes IUsed IFree IUse% Mounted on
> /dev/mapper/weird-inodetest   1024  1024     0  100% /mnt

> /dev/mapper/weird-inodetest   2048  1024  1024   50% /mnt

Then I stand corrected. I guess that the man page for mke2fs saying that
the inode count of a filesystem cannot be changed does not take resizing
into account.

I also thought that if resize2fs had the ability to extend the inode
table, then it would have options to give the user some degree of
control over the process. Apparently that's not the case.

andrea

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Alan McKinnon]

On Sun 21 August 2011 21:23:15 Andrea Conti did opine thusly:
> > Filesystem                  Inodes IUsed IFree IUse% Mounted on
> > /dev/mapper/weird-inodetest   1024  1024     0  100% /mnt
> > 
> > /dev/mapper/weird-inodetest   2048  1024  1024   50% /mnt
> 
> Then I stand corrected. I guess that the man page for mke2fs saying
> that the inode count of a filesystem cannot be changed does not
> take resizing into account.

Correct. A resized fs is technically a re-formatted fs. So to the 
ultra-pedantic the man page is actually still correct.

> I also thought that if resize2fs had the ability to extend the inode
> table, then it would have options to give the user some degree of
> control over the process. Apparently that's not the case.

inodes are not dynamic, they are laid down at exact points on the disk 
and the info about their location is in the superblock(s). The 
simplistic explanation is something like this:

You have X number of inodes, spaced Y blocks apart on a disk of size Z 
bytes. A directory listing declares a file is at inode #M therefore 
it's physical position on the disk is guaranteed to be at 

M * (block size)

and the filesystem driver can seek directly to that spot. Two things 
are immediately self-evident:

1. Changing the density of inodes is not realistic (unless you want to 
invest the same effort that went into Window's defrag)

2. Fixed inodes are an ancient concept that provoke an "Eh? Say what? 
You still doing that????" response. ReiserFS was developed in part to 
address this and make an 

-- 
alan dot mckinnon at gmail dot com

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Paul Hartman]

On Sat, Aug 20, 2011 at 8:41 PM, Hilco Wijbenga
<hilco.wijbenga@gmail.com> wrote:
> However, when I try to extract the Portage snapshot, I get "No space
> left on device" a long way into the untar process. According to df
> /portage (i.e. /mnt/gentoo/usr/portage) is only 35% full. In fact, not
> a single partition or mount is even close to full (except for
> /mnt/static, the DVD).

Try "df -i" to check your inode usage.

Re: [gentoo-user] [Gentoo install] Disk full at 35%?

[Hilco Wijbenga]

On 20 August 2011 20:05, Paul Hartman <paul.hartman+gentoo@gmail.com> wrote:
> On Sat, Aug 20, 2011 at 8:41 PM, Hilco Wijbenga
> <hilco.wijbenga@gmail.com> wrote:
>> However, when I try to extract the Portage snapshot, I get "No space
>> left on device" a long way into the untar process. According to df
>> /portage (i.e. /mnt/gentoo/usr/portage) is only 35% full. In fact, not
>> a single partition or mount is even close to full (except for
>> /mnt/static, the DVD).
>
> Try "df -i" to check your inode usage.

Yes, thanks, I had just found out about "df -i" myself. :-)