Redhat 7.3 hosed - can I just reinstall?
 

Up until this morning my Redhat 7.3 install was working great.

I problem started when I tried to do 'ls' on directories and was getting a Segmentation Fault. So I restarted the machine.

Now the machine seg faults during the boot up process, at some point goes into showing weird asci characters and eventually freezes. I have fsck the disk and there are no errors. If I use the Redhat Rescue I can see the files are on the drive but if I try to chroot to /mnt/sysimage, ls stops working again.

I was thinking that I could just reinstall or use the upgrade function to fix it.

I'd like to salvage the current configuration if possible.

On a side note, I wonder if the server was hacked or what would cause this?

Any advice would be very appreciated.

TIA!
Carl

red hat 7.3 is my favorite, so I feel your pain.

actually one way you can retrieve your data in the future is by creating a separate primary partition for your /home directory, that way if you lose everything then you can just replace it and use the same /home partition - you just have to mount it once you re-install.
I would just say to backup any info you want via a cd recorder if you have one by using cdrecord from the command line. If not then start over. I don't know what could have caused that - I don't run a server and my 7.3 has been super stable since day 1.

Actually I had various partitions. The most important one being /www which was on dev/sda8.

Anyone know if I can just do an upgrade?

upgrading will save as much of your configuration
stuff as it can. i have a few times backed up stuff
and upgraded to see how it went, and it was generally
ok. most of the settings stay the same. the upgrade
will usually even make backups of the files it needs
to overwrite.

Not to get you unduly worried, but when 'ls' started segfaulting on me, it was the first symptom of bad RAM. If this is the case then no amount of re-installs will solve it. http://www.memtest86.com have a really good RAM checker that you add to your boot menu, then just let it run for a while.

As Obi mentioned, "Segmentation fault" is pretty much HW related issue.

Is it just the "ls" command, or anything you typed gave the same error. try "cat" or "telnet" or anything else, and see if the error keep appearing.

Another factor indicative to bad RAM is the compiler segfaulting on compiling bigger programs - sometimes even small ones. If you get internal compiler segmentation faults during compilation, then type make again and you get a bit further and it happens again, very likely to be bad RAM.

I had considered that it might be hardware related but why would ls and such work fine using linux rescue? (Using a different kernel, etc)?

ok, I ran a short memory test and no errors came up. I ran the upgrade, didn't fix anything.

What's happening during bootup is that I get a bunch of Seg faults on lines in rc.sysinit(rm -r {filename} type lines) (Also more times further into the boot process)

Hardware wise, does this look like memory, hard drive controller or?

When I am in rescue mode, everything seems to work just fine.

maybe the HD is bad.

try to boot into single user and do fsck on the fs.

Upgrading to RH9 seems to have done the trick. I tried new memory, a new SCSI card, 'upgrading' to RH7.3. None of that worked. I had previously tried fsck. No errors were reported.

Still not sure what was up but it's working for now..

now to get my httpd.conf file migrated from Apache1.3 to Apache2.x :)