Re: nilfs_btree_propagate problem

[Ryusuke Konishi <ryusuke-sG5X7nlA6pw@public.gmane.org>]

Hi,
On Thu, 7 Oct 2010 15:23:45 +0200, Andrea Gelmini wrote:
> Hi all,
>   and thanks a lot for your work on Nilfs.
>   I'm using it for my /home, with
>   Ubuntu 10.10
>   Vanilla kernel from Linus git repository (at this commit)¹
>   nilfs_cleanerd v2.0.18 (from ubuntu deb nilfs-tools 2.0.18-2).
> 
>   Well, everything's working fine, but since the initial clone of my
> data, I have problem
>   with ~/dev/prg/wine.
>   I mean, if I try to compile it, after a while, I had process stucked
> (it seems they are not
>   able to write, but they can read from the disk).
>   Meanwhile my log are flooded with this:
> Oct  7 11:53:39 vaio kernel: [ 3408.688565] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:53:44 vaio kernel: [ 3413.696833] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:53:49 vaio kernel: [ 3418.704543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:53:54 vaio kernel: [ 3423.712541] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:53:59 vaio kernel: [ 3428.720545] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:04 vaio kernel: [ 3433.728548] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:09 vaio kernel: [ 3438.736543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:14 vaio kernel: [ 3443.744543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:19 vaio kernel: [ 3448.752543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:24 vaio kernel: [ 3453.760542] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:29 vaio kernel: [ 3458.768554] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:34 vaio kernel: [ 3463.776543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:39 vaio kernel: [ 3468.784458] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:44 vaio kernel: [ 3473.792545] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:49 vaio kernel: [ 3478.800543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:54 vaio kernel: [ 3483.808542] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:54:59 vaio kernel: [ 3488.816546] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:55:04 vaio kernel: [ 3493.824543] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:55:09 vaio kernel: [ 3498.832545] nilfs_btree_propagate: key
> = 2823, level == 0
> Oct  7 11:55:14 vaio kernel: [ 3503.840554] nilfs_btree_propagate: key
> = 2823, level == 0
> 
>    I guess it's important to note the 5 seconds of delay between log
> entries. Yes, it's my
>    snapshot interval.
>    The only way to solve it is a crude switch off/on.
>    At reboot everything is looking fine.
> 
>    Reading access gives no problems:
>     find -type f  -exec cat {} > /dev/null \;
> 
> Any clues?
> 
> Thanks a lot for your time and work,
> Andrea

Do you still have this problem?

The messages indicate that nilfs met a wrong btree node block during
making a log for writing.

The 5 seconds interval is the default writeback time.  So, It seems to
be generated by the corrupted block left on memory.
 
> n.b.: well, I almost use my notebook with hibernation/resume. Does
> nilfs_cleanerd suffer long jumps in time? Also, sometimes
>         the resume doesn't go well. I force a sync before hibernation,
> but does the fs can be screwed in this way?

nilfs_cleanred may see a long jump after resuming from hibernation,
but I haven't seen any issues so far.

If you meet the issue next time, try dmesg command to print kernel log
buffered on memory. It may give us some clue.

Thanks,
Ryusuke Konishi
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html