errors in ext4_mb_generate_buddy and ext4_remount

errors in ext4_mb_generate_buddy and ext4_remount

[Christoph Anton Mitterer]

Hi.

I've had several crashes recently, which lead to things like these:
May  6 00:23:53 heisenberg kernel: [   31.941036] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016253
May  6 00:23:53 heisenberg kernel: [   31.941049] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016250
May  6 00:23:53 heisenberg kernel: [   31.941073] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016249
May  6 00:23:53 heisenberg kernel: [   31.941106] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016248
May  6 00:23:53 heisenberg kernel: [   31.941120] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016247
May  6 00:23:53 heisenberg kernel: [   31.941135] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016240
May  6 00:23:53 heisenberg kernel: [   31.941158] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016239
May  6 00:23:53 heisenberg kernel: [   31.941168] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016238
May  6 00:23:53 heisenberg kernel: [   31.941174] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016237
May  6 00:23:53 heisenberg kernel: [   31.941180] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016236
May  6 00:23:53 heisenberg kernel: [   31.941190] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016235
May  6 00:23:53 heisenberg kernel: [   31.941200] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016234
May  6 00:23:53 heisenberg kernel: [   31.941212] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016233
May  6 00:23:53 heisenberg kernel: [   31.941217] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016232
May  6 00:23:53 heisenberg kernel: [   31.941228] EXT4-fs (dm-2):
ext4_orphan_cleanup: deleting unreferenced inode 2016231

... and some (at least to me) worrying questions during fsck.

However, I constantly get these errors:
[  331.744025] EXT4-fs (dm-2): error count: 10
[  331.744030] EXT4-fs (dm-2): initial error at 1304673122:
ext4_mb_generate_buddy:731
[  331.744034] EXT4-fs (dm-2): last error at 1304673124: ext4_remount:4250
... (probably) when reading some files. I guess it's reading files, at
least it happens during a debsums run (checks the package sums of more or
less all Debian packages installed).
Also the numbers in the errors above, are (at least until know, and as far
as I looked after it) the same.

It also happens (as far as I can see) only once, i.e. if the error has
already occurred, and I do e.g. my debsums again, it's not printed again.

fsck runs through fine (but quite quick; yes - of course - I've used -f)

Any ideas? (Please CC me as I'm off list).


Cheers,
Chris.


btw:
This (occuring in the host's kernel log and not the VM's)
[  325.655192] EXT4-fs (dm-2): Unaligned AIO/DIO on inode 1060986 by
VirtualBox; performance will be poor.
means that the blocks of the the guest are not aligned to the blocks of
the host, right?
But why is it only printed once?

Re: errors in ext4_mb_generate_buddy and ext4_remount

[Ted Ts'o]

On Sun, May 08, 2011 at 12:56:19AM +0000, Christoph Anton Mitterer wrote:
> 
> However, I constantly get these errors:
> [  331.744025] EXT4-fs (dm-2): error count: 10
> [  331.744030] EXT4-fs (dm-2): initial error at 1304673122:
> ext4_mb_generate_buddy:731
> [  331.744034] EXT4-fs (dm-2): last error at 1304673124: ext4_remount:4250
> ... (probably) when reading some files. I guess it's reading files, at
> least it happens during a debsums run (checks the package sums of more or
> less all Debian packages installed).
> Also the numbers in the errors above, are (at least until know, and as far
> as I looked after it) the same.
> 
> It also happens (as far as I can see) only once, i.e. if the error has
> already occurred, and I do e.g. my debsums again, it's not printed again.
> 
> fsck runs through fine (but quite quick; yes - of course - I've used -f)
> 
> Any ideas? (Please CC me as I'm off list).

Kernels 2.6.35 and newer will set superblock fields that will result
in those messages.  They're not error messages per se; it's just a
notification that there was an file system error in the past that
hasn't been fixed yet.  In your case, it *has* been fixed, but you
don't have a new enough version of e2fsprogs so that it clears the
relevant parts of the superblock.

The debian packages of e2fsprogs version 1.41.14-1 in experimental, or
1.41.12-4 in unstable, has the necessary new code in e2fsck.  I'm
working on getting the patch pushed to Debian obsolete^H^H^H^H^H^H^H
stable; hopefully that will happen in the next week or two.

		       	    	      	       - Ted

Re: errors in ext4_mb_generate_buddy and ext4_remount

[Christoph Anton Mitterer]

Hi Ted.

On Sun, 2011-05-08 at 21:01 -0400, Ted Ts'o wrote:
> Kernels 2.6.35 and newer will set superblock fields that will result
> in those messages.  They're not error messages per se; it's just a
> notification that there was an file system error in the past that
> hasn't been fixed yet.  In your case, it *has* been fixed, but you
> don't have a new enough version of e2fsprogs so that it clears the
> relevant parts of the superblock.
Thanks for that explanation.

> The debian packages of e2fsprogs version 1.41.14-1 in experimental, or
> 1.41.12-4 in unstable
Ah, I haven't tried it with the one that was uploaded to unstable
yet.... will give it a try later.


Another thing. Perhaps you've seen my post where I've unfortunately
overridden an (important ;) ) ext3 fs with mkfs, instead of doing fsck.

Of course this is my fault but do you see any chance to add checks to
mke2fs whether any known fs/partition type/other containter like
luks/mdadm/etc. is on the device? (And of course require a
--force=I-really-know-what-I-do™ parameter to the evil.)

Cheers,
Chris.
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: errors in ext4_mb_generate_buddy and ext4_remount

[Christoph Anton Mitterer]

[-- Attachment #1: Type: text/plain, Size: 419 bytes --]

Hi Ted.

On Sun, 2011-05-08 at 21:01 -0400, Ted Ts'o wrote:
> The debian packages of e2fsprogs version 1.41.14-1 in experimental, or
> 1.41.12-4 in unstable, has the necessary new code in e2fsck.  I'm
> working on getting the patch pushed to Debian obsolete^H^H^H^H^H^H^H
> stable; hopefully that will happen in the next week or two.
The problem at least persists in the 1.41.12-4 version.


Cheers,
Chris.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5677 bytes --]

Re: errors in ext4_mb_generate_buddy and ext4_remount

[Ted Ts'o]

On Wed, Jun 01, 2011 at 12:38:13AM +0200, Christoph Anton Mitterer wrote:
> Hi Ted.
> 
> On Sun, 2011-05-08 at 21:01 -0400, Ted Ts'o wrote:
> > The debian packages of e2fsprogs version 1.41.14-1 in experimental, or
> > 1.41.12-4 in unstable, has the necessary new code in e2fsck.  I'm
> > working on getting the patch pushed to Debian obsolete^H^H^H^H^H^H^H
> > stable; hopefully that will happen in the next week or two.
> The problem at least persists in the 1.41.12-4 version.

Um, are you sure?  Try forcing an fsck run on the file systems in question.

    	    	       	       	       - Ted

Re: errors in ext4_mb_generate_buddy and ext4_remount

[Christoph Anton Mitterer]

[-- Attachment #1: Type: text/plain, Size: 301 bytes --]

On Wed, 2011-06-01 at 00:15 -0400, Ted Ts'o wrote:
> Um, are you sure?  Try forcing an fsck run on the file systems in question.
Of course I did ;-)  (fsck.ext4 -f -D /dev/bla)...

And as I had some other fs troubles with that ext4 yesterday, I ran it
also several times ;)


Cheers,
Chris.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5677 bytes --]