EXT2-fs error with 2.4.4 (using CVS)

EXT2-fs error with 2.4.4 (using CVS)

[Frédéric L. W. Meunier]

Hi. I received the following error while updating my Mozilla
sources from MOZILLA_0_8_1_20010326_RELEASE to
MOZILLA_0_9_RELEASE via CVS:

==> /var/log/syslog <==
May  8 00:25:52 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
name_len=9
May  8 00:25:52 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
name_len=9

When CVS finished, I received the following error:

May  8 01:11:29 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
name_len=9
May  8 01:11:29 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
name_len=9

And the following CVS message:

cvs checkout: cannot remove
mozilla/xpinstall/wizard/windows/nszip: No such file or
directory

When I tried to access
/usr/local/src/CVS/X/mozilla/xpinstall/wizard/windows/ I got
the same EXT2-fs error messages.

The partition is /usr/local/src
(/dev/ide/host0/bus0/target0/lun0/part3)

Is this some sort of ext2 fs corruption or what ? I booted with
2.4.4 9 days ago, and there were no problems with fsck.

checkout start: Tue May 8 00:22:40 BRT 2001
checkout finish: Tue May 8 01:11:31 BRT 2001

My .config is at http://www.pervalidus.net/.config-2.4.4.txt
dmesg at http://www.pervalidus.net/dmesg-2.4.4.txt

-- 
0@pervalidus.{net, {dyndns.}org} Tel: 55-21-717-2399 (Niterói-RJ BR)

Re: EXT2-fs error with 2.4.4 (using CVS)

[Andreas Dilger]

Federic Meunier writes:
> ==> /var/log/syslog <==
> May  8 00:25:52 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
> directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
> name_len=9
> May  8 00:25:52 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
> directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
> name_len=9
> 
> When CVS finished, I received the following error:
> 
> May  8 01:11:29 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
> directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
> name_len=9
> May  8 01:11:29 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
> directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
> name_len=9

Since it is always the same inode, I would say it is corrupt.  You need to
run e2fsck to fix it.  It looks like it is a single-bit error on the disk.
The rec_len=16404=0x4014.  One (of many possible) valid rec_len would be
0x14=20.  To be valid we need name_len <= rec_len <= block size.

Chances are, when you run e2fsck, it will fix this dirent, but the rest of
the directory entries in that block will be moved to lost+found.

I would suspect a hardware problem, to create a single-bit error (if it
is such).

Cheers, Andreas
-- 
Andreas Dilger  \ "If a man ate a pound of pasta and a pound of antipasto,
                 \  would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/               -- Dogbert

Re: EXT2-fs error with 2.4.4 (using CVS)

[Frédéric L. W. Meunier]

On Mon, May 07, 2001 at 10:45:59PM -0600, Andreas Dilger wrote:

<skip>

> > May  8 01:11:29 pervalidus kernel: EXT2-fs error (device ide0(3,3)): ext2_readdir: bad entry in
> > directory #162813: directory entry across blocks - offset=92, inode=451111, rec_len=16404,
> > name_len=9

> Since it is always the same inode, I would say it is corrupt.  You need to
> run e2fsck to fix it.  It looks like it is a single-bit error on the disk.
> The rec_len=16404=0x4014.  One (of many possible) valid rec_len would be
> 0x14=20.  To be valid we need name_len <= rec_len <= block size.

OK, I ran.

# fsck -v /dev/ide/host0/bus0/target0/lun0/part3
Parallelizing fsck version 1.19 (13-Jul-2000)
e2fsck 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09
/dev/ide/host0/bus0/target0/lun0/part3 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information

  159689 inodes used (31%)
    1363 non-contiguous inodes (0.9%)
         # of inodes with ind/dind/tind blocks: 3884/1/0
  408910 blocks used (79%)
       0 bad blocks
			
  145848 regular files
   13624 directories
       0 character device files
       0 block device files
       0 fifos
       2 links
       208 symbolic links (208 fast symbolic links)
       0 sockets
--------
  159682 file

> Chances are, when you run e2fsck, it will fix this dirent, but the rest of
> the directory entries in that block will be moved to lost+found.

Nothing in /usr/local/src/lost+found

> I would suspect a hardware problem, to create a single-bit error (if it
> is such).

I hope not. I know at least my RAM is OK. I installed 256Mb
before booting with 2.4.4 and tested all with memtest86.

CVS couldn't remove mozilla/xpinstall/wizard/windows/nszip
after the process terminated, but
mozilla/xpinstall/wizard/windows/nszip/Entries and
mozilla/xpinstall/wizard/windows/nszip/Tag were updated after
the first EXT2-fs error messages.

Well, everything seems OK for now. Thanks.

-- 
0@pervalidus.{net, {dyndns.}org} Tel: 55-21-717-2399 (Niterói-RJ BR)