Hi,
2010/2/12 Jan de Kruyf <jan.de.kruyf@xxxxxxxxx>:
> Pliny the Elder (ad 23–79): ‘Semper aliquid novi Africam adferre
> [Africa always brings [us] something new]’
>
> Hallo,
> something is not quite right between the git repository and my git master copy
> somehow the patches dont update the local copy with git am
> I have to extract the patch and apply with 'patch -l' (Match patterns
> loosely, in case tabs or spaces have been munged in your files.)
> Which then has other complications like finding the wrong spot in
> super.c for one patch, so that hunk was then rejected and had to be
> done by hand.
>
>
> On the crash:
> My initial idea was quite wrong. Something went amiss somewhere in the
> middle of nowhere.
> It might be cleanerd connected, but I cannot say for sure.
>
> See the attached .tgz for details of my post mortem effort. If more
> data is needed please say so.
>
> Regards,
>
> Jan de Kruyf.
Thank you for the detail report.
According to your log, many tiny checkpoints were created
around 2010-02-09 20:30:01 ~ 2010-02-09 20:31:00.
I saw the dump data of the segment 343 to see the logs having checkpoints
on "20:30:01", but they looked normal to me.
The series of checkpoints after 1844624 seems to need care as you say.
1844624 2010-02-09 20:30:59 cp - 55 16621
1844625 2010-02-09 20:30:59 cp - 33 16621
1844626 2010-02-09 20:30:59 cp - 33 16621
1844627 2010-02-09 20:30:59 cp - 34 16621
1844628 2010-02-09 20:30:59 cp - 33 16621
1844629 2010-02-09 20:30:59 cp - 33 16621
1844630 2010-02-09 20:30:59 cp - 33 16621
But, it was not included in the segment 343 but in the segment 344.
Checkpoints can be created when application performs synchronous writes
(e.g. fsync or OSYNC writes).
So, we needs an additional dumpseg data to confirm if it's normal or not.
Can you still take out the dump ?
In addition, I think "ls -laiR" for the var directory would be helpful because
it prints inode numbers along with file names.
Thanks in advance,
Ryusuke Konishi
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
