[TriLUG] Help Please - Strange Filesystem Error on Boot

[Scott Chilcote]

Hi Folks,

I've dealt with several Linux filesystem problems but this one has me 
stumped.

I have a notebook computer with an upgraded hard drive that has worked 
fine for a couple of months.  It has Redhat 9 and a slightly newer 
kernel (2.4.20-20.9).

I started up the machine today and found that it's failing to initialize 
the filesystem at bootup.  Some of the messages are:

> loading default keymap(us):
> /etc/rc.d/rc.sysinit: line 16222:/dev/null: read-only filesystem [failed]
> 
> setting hostname localhost.localdomain: dup2: bad file descriptor [failed]
> /etc/rc.sysinit:line 173: /dev/null:read only filesystem
> initializing USB controller [failed]
> 
> awk: cmd.. line:2: fatal: cannot open file "/proc/mount for reading(no such file of directory)
> 
> checking root filesystem
> dup2: bad file descriptor [failed]
> 
> *** dropping you to a shell
> *** give root password for maintenance
> (or type control-d to continue):
> repair filesystem 1#:

It looks like one of the partitions is bad, however, when I run e2fsck 
on both of them it returns and says they are clean.  It reports no errors.

The drive has two ext3 partitions: /dev/hda3 (boot) and /dev/hda4 (/). 
/dev/hda2 is swap.

I've done the following so far, and they have not worked:

1. Used e2fsck to check both partitions for bad blocks (-c), none found
2. Booted using the RHL9 CD-ROM using "linux rescue" and looked at all 
of the recently modified files in /etc to see if anything was hosed 
(looks clean)
3. Rebooted and used the old kernel (.20-8) from grub, no improvement.
4. Used the RH9 CD-ROMs to reinstall the RPMs, kernel, modules, and boot 
configuration.  This ran without errors but did not have any effect.

My first thought when I have a problem like this is "What did I mess up 
the last time I had it running?"  but scout's honor, the worst I've done 
to the machine lately is try a PCMCIA modem, which runs fine.  It also 
boots and runs fine from /dev/hda1, which is a Windows Me partition 
(yes, for games).  I've had it that way for several months.

I miss the first couple of lines of errors reported during the boot 
process.  I've run dmesg and they aren't being captured there.  If 
anyone has some ideas to try for zeroing in on the problem I'd be most 
grateful.

How can this computer fail to boot and yet e2fsck say it's clean?

Thanks for any help!

--
Scott C.