All of lore.kernel.org
 help / color / mirror / Atom feed
* Stale NFS file handle after power loss
@ 2013-06-08 19:14 Thomas Koch
  2013-06-10 13:00 ` Josef Bacik
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Koch @ 2013-06-08 19:14 UTC (permalink / raw)
  To: linux-btrfs

Hi,

I had a power loss on my btrfs file system during package installation and now 
I found this:

<code>
% ls -l /var/lib/dpkg/tmp.ci
ls: cannot access /var/lib/dpkg/tmp.ci/conffiles: Stale NFS file handle
ls: cannot access /var/lib/dpkg/tmp.ci/md5sums: Stale NFS file handle
total 0
-????????? ? ? ? ?            ? conffiles
-????????? ? ? ? ?            ? md5sums
% sudo rm -rf /var/lib/dpkg/tmp.ci/conffiles
rm: cannot remove ‘/var/lib/dpkg/tmp.ci/conffiles’: Stale NFS file handle
</code>

I booted with a live system and ran btrfsck (without --repair!) but this did 
not change the situation. Unfortunately I wasn't clever enough to save the 
output of btrfsck somehow. Do you need it?

Could you help me to get rid of these bogus files, please?

Regards,

Thomas Koch, http://www.koch.ro

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: Stale NFS file handle after power loss
  2013-06-08 19:14 Stale NFS file handle after power loss Thomas Koch
@ 2013-06-10 13:00 ` Josef Bacik
  2013-06-10 18:18   ` Thomas Koch
  0 siblings, 1 reply; 6+ messages in thread
From: Josef Bacik @ 2013-06-10 13:00 UTC (permalink / raw)
  To: Thomas Koch; +Cc: linux-btrfs

On Sat, Jun 08, 2013 at 01:14:35PM -0600, Thomas Koch wrote:
> Hi,
> 
> I had a power loss on my btrfs file system during package installation and now 
> I found this:
> 
> <code>
> % ls -l /var/lib/dpkg/tmp.ci
> ls: cannot access /var/lib/dpkg/tmp.ci/conffiles: Stale NFS file handle
> ls: cannot access /var/lib/dpkg/tmp.ci/md5sums: Stale NFS file handle
> total 0
> -????????? ? ? ? ?            ? conffiles
> -????????? ? ? ? ?            ? md5sums
> % sudo rm -rf /var/lib/dpkg/tmp.ci/conffiles
> rm: cannot remove ‘/var/lib/dpkg/tmp.ci/conffiles’: Stale NFS file handle
> </code>
> 
> I booted with a live system and ran btrfsck (without --repair!) but this did 
> not change the situation. Unfortunately I wasn't clever enough to save the 
> output of btrfsck somehow. Do you need it?
> 
> Could you help me to get rid of these bogus files, please?

Don't run with --repair, but please capture the output so I can see what's going
wrong.  Thanks,

Josef

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: Stale NFS file handle after power loss
  2013-06-10 13:00 ` Josef Bacik
@ 2013-06-10 18:18   ` Thomas Koch
  2013-06-13 19:45     ` Thomas Koch
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Koch @ 2013-06-10 18:18 UTC (permalink / raw)
  To: Josef Bacik; +Cc: linux-btrfs

On Monday, June 10, 2013 09:00:47 AM Josef Bacik wrote:
> On Sat, Jun 08, 2013 at 01:14:35PM -0600, Thomas Koch wrote:
> > Hi,
> > 
> > I had a power loss on my btrfs file system during package installation and
> > now I found this:
> > 
> > <code>
> > % ls -l /var/lib/dpkg/tmp.ci
> > ls: cannot access /var/lib/dpkg/tmp.ci/conffiles: Stale NFS file handle
> > ls: cannot access /var/lib/dpkg/tmp.ci/md5sums: Stale NFS file handle
> > total 0
> > -????????? ? ? ? ?            ? conffiles
> > -????????? ? ? ? ?            ? md5sums
> > % sudo rm -rf /var/lib/dpkg/tmp.ci/conffiles
> > rm: cannot remove ‘/var/lib/dpkg/tmp.ci/conffiles’: Stale NFS file handle
> > </code>
> > 
> > I booted with a live system and ran btrfsck (without --repair!) but this
> > did not change the situation. Unfortunately I wasn't clever enough to
> > save the output of btrfsck somehow. Do you need it?
> > 
> > Could you help me to get rid of these bogus files, please?
> 
> Don't run with --repair, but please capture the output so I can see what's
> going wrong.  Thanks,
> 
> Josef

sudo btrfsck /dev/x121e/root
checking extents
checking fs roots
root 5 inode 9319255 errors 1040
root 5 inode 9698560 errors 1000
root 5 inode 11591532 errors 2001
        unresolved ref dir 11591531 index 2 namelen 9 name conffiles filetype 1 
error 4
root 5 inode 11591534 errors 2001
        unresolved ref dir 11591531 index 4 namelen 7 name md5sums filetype 1 
error 4
found 270172975104 bytes used err is 1
total csum bytes: 261797296
total tree bytes: 2063106048
total fs tree bytes: 1575575552
btree space waste bytes: 561459705
file data blocks allocated: 386128334848
 referenced 262981738496
Btrfs Btrfs v0.19

thx for looking into this!

Thomas Koch, http://www.koch.ro

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: Stale NFS file handle after power loss
  2013-06-10 18:18   ` Thomas Koch
@ 2013-06-13 19:45     ` Thomas Koch
  2013-06-14  1:27       ` Josef Bacik
  0 siblings, 1 reply; 6+ messages in thread
From: Thomas Koch @ 2013-06-13 19:45 UTC (permalink / raw)
  To: Josef Bacik; +Cc: linux-btrfs

Thomas Koch:
> On Monday, June 10, 2013 09:00:47 AM Josef Bacik wrote:
> > On Sat, Jun 08, 2013 at 01:14:35PM -0600, Thomas Koch wrote:
> > > Hi,
> > > 
> > > I had a power loss on my btrfs file system during package installation
> > > and now I found this:
> > > 
> > > <code>
> > > % ls -l /var/lib/dpkg/tmp.ci
> > > ls: cannot access /var/lib/dpkg/tmp.ci/conffiles: Stale NFS file handle
> > > ls: cannot access /var/lib/dpkg/tmp.ci/md5sums: Stale NFS file handle
> > > total 0
> > > -????????? ? ? ? ?            ? conffiles
> > > -????????? ? ? ? ?            ? md5sums
> > > % sudo rm -rf /var/lib/dpkg/tmp.ci/conffiles
> > > rm: cannot remove ‘/var/lib/dpkg/tmp.ci/conffiles’: Stale NFS file
> > > handle </code>
> > > 
> > > I booted with a live system and ran btrfsck (without --repair!) but
> > > this did not change the situation. Unfortunately I wasn't clever
> > > enough to save the output of btrfsck somehow. Do you need it?
> > > 
> > > Could you help me to get rid of these bogus files, please?
> > 
> > Don't run with --repair, but please capture the output so I can see
> > what's going wrong.  Thanks,
> > 
> > Josef
> 
> sudo btrfsck /dev/x121e/root
> checking extents
> checking fs roots
> root 5 inode 9319255 errors 1040
> root 5 inode 9698560 errors 1000
> root 5 inode 11591532 errors 2001
>         unresolved ref dir 11591531 index 2 namelen 9 name conffiles
> filetype 1 error 4
> root 5 inode 11591534 errors 2001
>         unresolved ref dir 11591531 index 4 namelen 7 name md5sums filetype
> 1 error 4
> found 270172975104 bytes used err is 1
> total csum bytes: 261797296
> total tree bytes: 2063106048
> total fs tree bytes: 1575575552
> btree space waste bytes: 561459705
> file data blocks allocated: 386128334848
>  referenced 262981738496
> Btrfs Btrfs v0.19

Hi Josef,

do you think that you might find some solution to this problem or should I plan 
to re-setup my machine?

Regards,

Thomas Koch, http://www.koch.ro

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: Stale NFS file handle after power loss
  2013-06-13 19:45     ` Thomas Koch
@ 2013-06-14  1:27       ` Josef Bacik
  2013-06-15 10:33         ` Russell Coker
  0 siblings, 1 reply; 6+ messages in thread
From: Josef Bacik @ 2013-06-14  1:27 UTC (permalink / raw)
  To: Thomas Koch; +Cc: Josef Bacik, linux-btrfs

On Thu, Jun 13, 2013 at 01:45:29PM -0600, Thomas Koch wrote:
> Thomas Koch:
> > On Monday, June 10, 2013 09:00:47 AM Josef Bacik wrote:
> > > On Sat, Jun 08, 2013 at 01:14:35PM -0600, Thomas Koch wrote:
> > > > Hi,
> > > > 
> > > > I had a power loss on my btrfs file system during package installation
> > > > and now I found this:
> > > > 
> > > > <code>
> > > > % ls -l /var/lib/dpkg/tmp.ci
> > > > ls: cannot access /var/lib/dpkg/tmp.ci/conffiles: Stale NFS file handle
> > > > ls: cannot access /var/lib/dpkg/tmp.ci/md5sums: Stale NFS file handle
> > > > total 0
> > > > -????????? ? ? ? ?            ? conffiles
> > > > -????????? ? ? ? ?            ? md5sums
> > > > % sudo rm -rf /var/lib/dpkg/tmp.ci/conffiles
> > > > rm: cannot remove ‘/var/lib/dpkg/tmp.ci/conffiles’: Stale NFS file
> > > > handle </code>
> > > > 
> > > > I booted with a live system and ran btrfsck (without --repair!) but
> > > > this did not change the situation. Unfortunately I wasn't clever
> > > > enough to save the output of btrfsck somehow. Do you need it?
> > > > 
> > > > Could you help me to get rid of these bogus files, please?
> > > 
> > > Don't run with --repair, but please capture the output so I can see
> > > what's going wrong.  Thanks,
> > > 
> > > Josef
> > 
> > sudo btrfsck /dev/x121e/root
> > checking extents
> > checking fs roots
> > root 5 inode 9319255 errors 1040
> > root 5 inode 9698560 errors 1000
> > root 5 inode 11591532 errors 2001
> >         unresolved ref dir 11591531 index 2 namelen 9 name conffiles
> > filetype 1 error 4
> > root 5 inode 11591534 errors 2001
> >         unresolved ref dir 11591531 index 4 namelen 7 name md5sums filetype
> > 1 error 4
> > found 270172975104 bytes used err is 1
> > total csum bytes: 261797296
> > total tree bytes: 2063106048
> > total fs tree bytes: 1575575552
> > btree space waste bytes: 561459705
> > file data blocks allocated: 386128334848
> >  referenced 262981738496
> > Btrfs Btrfs v0.19
> 
> Hi Josef,
> 
> do you think that you might find some solution to this problem or should I plan 
> to re-setup my machine?
> 

Yup sorry I just got sidetracked, I'll see what can be done tomorrow, but your
best bet it so pull down the latest btrfs-progs and build it and run

./btrfs-image /dev/whatever blah.img

and put the image up somewhere I can pull it down so I can fix btrfsck to fix
your file system and test it locally before I send you the patch.  Thanks,

Josef

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: Stale NFS file handle after power loss
  2013-06-14  1:27       ` Josef Bacik
@ 2013-06-15 10:33         ` Russell Coker
  0 siblings, 0 replies; 6+ messages in thread
From: Russell Coker @ 2013-06-15 10:33 UTC (permalink / raw)
  To: Josef Bacik; +Cc: Thomas Koch, linux-btrfs

On Fri, 14 Jun 2013, Josef Bacik <jbacik@fusionio.com> wrote:
> and put the image up somewhere I can pull it down so I can fix btrfsck to
> fix your file system and test it locally before I send you the
> patch.  Thanks,

Would it be possible to get this sort of thing fixed in the kernel by a scrub?

I've had a similar problem to Thomas (coincidentally also under /var/lib/dpkg) 
and I created a new filesystem and copied all the files across.  That wasn't 
particularly difficult as the system was a virtual machine, but it was still 
rather annoying.

Also if a file is corrupted then I think that rm -f should work to remove it. 
In the example cited the system would be in a good functional state after
"rm -f /var/lib/dpkg/tmp.ci/*" and it seems to me that file corruption 
shouldn't be a good reason to prevent that operation from completing.

-- 
My Main Blog         http://etbe.coker.com.au/
My Documents Blog    http://doc.coker.com.au/

^ permalink raw reply	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-06-15 10:33 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-06-08 19:14 Stale NFS file handle after power loss Thomas Koch
2013-06-10 13:00 ` Josef Bacik
2013-06-10 18:18   ` Thomas Koch
2013-06-13 19:45     ` Thomas Koch
2013-06-14  1:27       ` Josef Bacik
2013-06-15 10:33         ` Russell Coker

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.