bl4:
> If there were any write operations by other processes during this time,
> they should be listed in dmesg, but there were none.
Ok.
> Then I did:
> echo 1 > /proc/sys/vm/block_dump
> mount -o remount,rw /ro
> rsync --exclude=".wh.*" --exclude=lost+found -aHSx --devices --specials
> /rw/ /ro/
> mount -o remount,ro /ro
> echo 0 > /proc/sys/vm/block_dump
>
> There were some writes to hda1 (/ro), such as:
> [ 269.275225] rsync(2435): dirtied inode 2 (/) on hda1
> [ 269.275326] rsync(2435): dirtied inode 313873 (etc) on hda1
> [ 269.275445] rsync(2435): dirtied inode 43755 (cache) on hda1
> etc. Among them were 3 writes to tmpfs (/rw):
> [ 269.276621] rsync(2434): dirtied inode 5933 (.cache) on tmpfs
> [ 269.303232] rsync(2434): dirtied inode 5349 (motd) on tmpfs
> [ 269.303276] rsync(2434): dirtied inode 4965 (sm-notify.pid) on tmpfs
>
> I don't know if it means anything. I have attached full dmesg log.
>
> My knowledge in this area is very limited so please tell me if you know
> other ways to investigate this situation.
These "dirtied inode" messages are correct and showed those write are
done expectedly.
Accroding to your full log, I can see several "?" as a filename. They
may be replaced files by rsync. For instance,
> [ 269.279952] rsync(2436): dirtied inode 320026 (?) on hda1
> [ 269.280040] rsync(2436): dirtied inode 320029 (blkid.tab) on hda1
> [ 269.280114] rsync(2436): dirtied inode 320031 (?) on hda1
And i#320026 had appeared in your previous mail too.
----------------------------------------------------------------------
> Pass 1: Checking inodes, blocks, and sizes
:::
> Deleted inode 320026 has zero dtime. Fix? no
>
> Pass 2: Checking directory structure
:::
> Entry 'blkid.tab' in /etc (313873) has deleted/unused inode 320026.
> Clear? no
----------------------------------------------------------------------
As for as I know, /etc/blkid.tab is not updated usually. On your system,
is it updated frequently?
More interestingly, rsync and block_dump tells us i#320029 is blkid.tab
and i#320026 has no name. But fsck says blkid.tab is i#320026. The pair
of name and inode number is stored in a directry. So it seems /etc dir
on hda1 is not written-back correctly.
And your log shows
> [ 241.753731] EXT2-fs warning: mounting unchecked fs, running e2fsck is
> recommended
I am afraid your "mount -o remount,ro /ro" didn't work.
can you check it by /proc/mounts or something?
Of course, the first "fsck /dev/hda1" has to make sure the fs is really
clean.
J. R. Okajima
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
