corrupted filesystem (again!)

corrupted filesystem (again!)

[Paul L]

Hi all,

If my memory serves me right, this is the 4th time I've had corrupted
filesystem after using nilfs2 for almost 2 years. With help from the
list, I've managed to recover every time! A big thank you to the
everyone who helped! Let's see if this time I have any luck.

When I first noticed something wrong with the file system, command 'ls
-l' would list some files with outrageously big size, like 20 digits
or something. I don't remember doing any particular task before I
noticed this. So I thought I would just reboot and see if things come
back normal.

After a reboot, the disk failed to mount. That was almost two weeks
ago, and I was using 2.0.20. I didn't do anything further except
backing up the entire disk data.

Today I had a little more time and decided to upgrade to
nilfs2-2.0.21, and still the same. Here is what dmesg showed me when I
mount the file system (after I turned on CONFIG_NILFS_DEBUG and "echo
'-vvv recovery' > /proc/fs/nilfs2/debug_option"):


NILFS nilfs_fill_super: start(silent=0)
NILFS warning: broken superblock. using spare superblock.
NILFS warning: broken superblock. using spare superblock.
NILFS(recovery) nilfs_search_super_root: looking segment
(seg_start=1087488, seg_end=1089535, segnum=531, seg_seq=837256)
NILFS(recovery) load_segment_summary: checking segment
(pseg_start=1088825, full_check=1)
NILFS(recovery) load_segment_summary: done (ret=0)
NILFS(recovery) nilfs_search_super_root: found super root: segnum=531,
seq=837256, pseg_start=1088825, pseg_offset=1396
segctord starting. Construction interval = 5 seconds, CP frequency < 30 seconds
NILFS(dat) nilfs_dat_translate: failed (ret=-2)
NILFS warning (device mmcblk0p1): nilfs_ifile_get_inode_block: unable
to read inode: 2
NILFS: get root inode failed
NILFS(segment) nilfs_segctor_thread: segctord exiting.
NILFS nilfs_fill_super: aborted
NILFS put_nilfs: the_nilfs on bdev mmcblk0p1 was freed


Apparently, it was trying to use the spare superblock, but still
couldn't find the root inode.

I also tried the fsck0.niilfs2 utility from nilfs-utils-2.0.18. It
didn't help either. Here is the output:


Super-block:
    revision = 2.0
    blocksize = 4096
    write time = 2011-01-18 23:22:07
    indicated log: blocknr = 1088825
        segnum = 531, seq = 837256, cno=1787400

Clean FS.
A valid log is pointed to by superblock (No change needed): blocknr = 1088825
    segnum = 531, seq = 837256, cno=1787400
    creation time = 2011-01-18 23:22:07


I'm pretty clueless on what to do next. Any help would be greatly appreciated!

-- 
Regards,
Paul Liu
--
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

Re: corrupted filesystem (again!)

[Ryusuke Konishi]

Hi,
On Sat, 29 Jan 2011 15:06:03 -0800, Paul L wrote:
> Hi all,
> 
> If my memory serves me right, this is the 4th time I've had corrupted
> filesystem after using nilfs2 for almost 2 years. With help from the
> list, I've managed to recover every time! A big thank you to the
> everyone who helped! Let's see if this time I have any luck.
> 
> When I first noticed something wrong with the file system, command 'ls
> -l' would list some files with outrageously big size, like 20 digits
> or something. I don't remember doing any particular task before I
> noticed this. So I thought I would just reboot and see if things come
> back normal.
> 
> After a reboot, the disk failed to mount. That was almost two weeks
> ago, and I was using 2.0.20. I didn't do anything further except
> backing up the entire disk data.
> 
> Today I had a little more time and decided to upgrade to
> nilfs2-2.0.21, and still the same. Here is what dmesg showed me when I
> mount the file system (after I turned on CONFIG_NILFS_DEBUG and "echo
> '-vvv recovery' > /proc/fs/nilfs2/debug_option"):
> 
> 
> NILFS nilfs_fill_super: start(silent=0)
> NILFS warning: broken superblock. using spare superblock.
> NILFS warning: broken superblock. using spare superblock.
> NILFS(recovery) nilfs_search_super_root: looking segment
> (seg_start=1087488, seg_end=1089535, segnum=531, seg_seq=837256)
> NILFS(recovery) load_segment_summary: checking segment
> (pseg_start=1088825, full_check=1)
> NILFS(recovery) load_segment_summary: done (ret=0)
> NILFS(recovery) nilfs_search_super_root: found super root: segnum=531,
> seq=837256, pseg_start=1088825, pseg_offset=1396
> segctord starting. Construction interval = 5 seconds, CP frequency < 30 seconds
> NILFS(dat) nilfs_dat_translate: failed (ret=-2)
> NILFS warning (device mmcblk0p1): nilfs_ifile_get_inode_block: unable
> to read inode: 2
> NILFS: get root inode failed
> NILFS(segment) nilfs_segctor_thread: segctord exiting.
> NILFS nilfs_fill_super: aborted
> NILFS put_nilfs: the_nilfs on bdev mmcblk0p1 was freed
> 
> 
> Apparently, it was trying to use the spare superblock, but still
> couldn't find the root inode.

Sorry for inconvenience (again).

According to the log, DAT metadata file seems to be broken, but we
need more debug information to trace down what happened.

Can you try the following debug option?

 # echo "-vv dat" > /proc/fs/nilfs2/debug_option

Also, the summary information of the pointed log may be helpful:

 # dempseg /dev/mmcblk0p1 531
 # dempseg /dev/mmcblk0p1 530

And a few ragular questions:

- Do you think the garbage collector was working when it happened ?
- Was it near disk full ?


Regards,
Ryusuke Konishi

> I also tried the fsck0.niilfs2 utility from nilfs-utils-2.0.18. It
> didn't help either. Here is the output:
> 
> 
> Super-block:
>     revision = 2.0
>     blocksize = 4096
>     write time = 2011-01-18 23:22:07
>     indicated log: blocknr = 1088825
>         segnum = 531, seq = 837256, cno=1787400
> 
> Clean FS.
> A valid log is pointed to by superblock (No change needed): blocknr = 1088825
>     segnum = 531, seq = 837256, cno=1787400
>     creation time = 2011-01-18 23:22:07
> 
> 
> I'm pretty clueless on what to do next. Any help would be greatly appreciated!
--
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