last 541 kernel damages fileystem .. starting fresh FC3T2 candidate.

[Andy Green]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sunday 05 September 2004 09:19, Colin Charles wrote:
> On Sun, 2004-09-05 at 14:42, Jim Cornette wrote:
> > Since I booted up using the latest kernel and ended up with a lot of
> > file errors related to inodes, I'll be installing fresh.
>
> What errors are you noticing? Seems to be fine here, on my ppc box

"Seems to be fine" on machines here... but... I noticed that df -h filesystem 
usage does NOT match du -h -x / for the root fs.  Then I noticed that the 
boot partition df -h is completely broken...

[root at fastcat /]# du -h -x /
...
31G     /  <--- expected


[root at fastcat /]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/hdc2              36G   25G  9.2G  74% /
/dev/hdc1              99M  -16Z  101M 101% /boot
...

[root at fastcat /]# umount /boot
[root at fastcat /]# /sbin/e2fsck -f /dev/hdc1
e2fsck 1.35 (28-Feb-2004)
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
Free blocks count wrong for group #5 (37125, counted=7751).
Fix<y>? yes

Free blocks count wrong (107846, counted=78472).
Fix<y>? yes


/boot: ***** FILE SYSTEM WAS MODIFIED *****
/boot: 71/26104 files (18.3% non-contiguous), 25916/104388 blocks
[root at fastcat /]# /sbin/e2fsck -f /dev/hdc1
e2fsck 1.35 (28-Feb-2004)
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
/boot: 71/26104 files (18.3% non-contiguous), 25916/104388 blocks
[root at fastcat /]# mount /dev/hdc1 /boot
[root at fastcat /]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/hdc2              36G   25G  9.2G  73% /
/dev/hdc1              99M   23M   72M  24% /boot
...


I had a big blowup on the filesystem on my mail machine the other week, normal 
functioning until a machine freeze, on reboot into init=/bin/bash and doing 
a /sbin/e2fsck -f on it there were loads and loads of errors translating to 
random files being deleted.

The root filesystem stats above are after similarly booting this (so far, 
working) machine into init=/bin/bash and doing e2fsck twice to be certain 
there were no errors after fixes -- there was no errors reported the second 
time.  But df -h results are WRONG.

This is with selinux=0, although somewhere along the line this was lost from 
the kernel commandline for newer kernels.

VERY WORRYING!!!!!!

- -Andy

- -- 
Automatic actions for USB cameras, cardreaders, memory sticks, MP3 players
http://warmcat.com/usbautocam
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFBOtscjKeDCxMJCTIRAki4AJ9AisZT2XNX23rUDND2mrCyTSmYVwCff8k+
IH08utAHUHWXzNMoyURtpbQ=
=BeNh
-----END PGP SIGNATURE-----