[parisc-linux] ext3 failure with 2.6.16-rc2-pa1?

[Joel Soete]

Joel Soete wrote:
> Hello *,
> 
> Just to mentioned that I observe an ext3 failure on my d380 runing a
> 2.6.16-rc2 32bit up:
> [162962.520000] EXT3-fs error (device sdb6): ext3_readdir: bad entry in
> directory #40163: rec_len is smaller than mi
> nimal - offset=0, inode=0, rec_len=0, name_len=0
> [162962.700000] Aborting journal on device sdb6.
> [162962.780000] ext3_abort called.
> [162962.820000] EXT3-fs error (device sdb6): ext3_journal_start_sb: Detected
> aborted journal
> [162962.920000] Remounting filesystem read-only
> 
> /dev/sdb6               248895    117456    118589  50%
> /var/lib/vservers/Ubuntu/var
> 
> (to avoid it's not the vserver kernel so the guest sever is not runing on this
> disk and even not simply chrooted, though).
> 
> Fwiw, I didn't noticed on b2k runing same kernel level (i.e. 2.6.16-rc2-pa +
> latest kyle's patches, compiled with gcc-4.0 I get rid of willy's patch).
> 
> Any body else encountered same pb?
> 
Grr, on this same d380 & same kernel, pb re-occured but this time on root fs. I have so to reboot:
Checking root file system.../ contains a file system with errors, check forced.

Inode 500617, i_blocks is 64, should be 16.  FIXED.

/: Inode 500618, i_blocks is 8, should be 2.  FIXED.

/: Inode 500620, i_blocks is 32, should be 8.  FIXED.

/: Inode 500621, i_blocks is 8, should be 2.  FIXED.

/: Inode 500622, i_blocks is 16, should be 4.  FIXED.

/: Inode 500623, i_blocks is 64, should be 16.  FIXED.

/: Inode 500624, i_blocks is 48, should be 12.  FIXED.

Duplicate or bad block in use!

/: Multiply-claimed block(s) in inode 122474: 249729 249730 249731 249732 249749

/: Multiply-claimed block(s) in inode 500617: 249729 249730 249731 249732

/: Multiply-claimed block(s) in inode 500623: 249749

/: (There are 3 inodes containing multiply-claimed blocks.)



/: File /root/.bash_history.archive (inode #122474, mod time Tue Jan  3 13:43:56 2006)

   has 5 multiply-claimed block(s), shared with 2 file(s):

/:      /usr/bin/Mail (inode #500623, mod time Sat Feb  4 05:52:07 2006)

/:      /usr/share/man/man8/ipcs.8.gz (inode #500617, mod time Tue Jan 17 15:52:09 2006)

/:



/: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.

         (i.e., without -a or -p options)

  failed!



An automatic fsck upon the root filesystem failed.

A manual fsck must be performed, and the system rebooted.

This fsck can be performed in maintenance mode.



Please note, the root filesystem is currently mounted read-only.

The fsck should only be performed while the root filesystem is

mounted read-only. However, the command to remount it read-write

is:


INIT: Id "T0" respawning too fast: disabled for 5 minutes

INIT: Id "TB0" respawning too fast: disabled for 5 minutes

INIT: no more processes left in this runlevel



   # mount -n -o remount,rw /



In order to exit from the maintenance shell, press CONTROL-D

and the system will REBOOT.

Give root password for maintenance

(or type Control-D to continue):

Any idea?

Tx,
	Joel