[seul-edu] RAID help... I have a big mess... I think

[Ryan Booz <ryanbooz@alumni.psu.edu>]

Hello everyone...

Let me just say off the bat that I am sorry for making this so long.  If
this is better sent to another mailing list or newsgroup that someone
knows of... I'll be happy to direct it there.  I know this is a general
educational linux users board, not a tech support hotline.  Any
direction, either directly or by refering me somewhere else is great!

I have been working on and off for a week with a few other people trying
to figure out how remedy this situation I'm having with RAID.  Haven't
made much progress at the moment and wanted to get some insight from
others.

I have a RH 6.2 Samba file/logon server that I converted to a
RAID system (for backup purposes) about two months ago.  Deciding that
RAID-1 was what I wanted, I figured, upon some suggestions from people,
that using the graphical install with RH 6.2 was the quickest way to go
about it.  I had done all the reading (HOW-TO and some background
papers) but honestly didn't have the time to back everything off, create
the system, then the RAID layout and do all the testing.  OK, I was
wrong there and should have.

Anyway, so I have two 13G hard drives with four partitions (/, /boot,
/pchome, swap).  The three useable directories were then put into three
RAID-1 disks.  RH allows you to boot off of RAID now - althought my
lilo.conf still said /dev/hda1 as the boot disk.

Things seemed to work fine for a time.  But then last week while fixing
another server that crashed (my fault for not getting it backed up
yet...) this thing went down.  It wouldn't boot.  Each try would give
one of two messages... either that /dev/hda5 had a bad superblock, or
that it couldn't get a shell.  So I got in the one time it gave me a
basic shell, and could mount all the drives (both hda and hdc), but
I did have some problems on /dev/hda5.  So, scanned it, only to realize
that it was the swap drive and shouldn't be mountable so what was the
deal.  Well, as I played around with mounting the drives, I noticed that
the drives in the /etc/fstab didn't match the drives at all.  So I
edited that (figured I had nothing to lose) and tried a reboot.  Still
nothing.

Then I tried to pass to the kernel a different boot disk... actually
just went through all the drives.  Well, when I gave it /dev/hdc1 (since
it should have been the copy of the boot on /dev/hda1) it booted up
fine... except for RAID of course.  So now, I have the thing running on
one disk, booted off of /dev/hdc1, but /dev/hda mounted for the rest of
the drives.  I can mount the other partitions on /dev/hdc but can't get
RAID working again.  I thought that I was supposed to be able to change
over to another drive, if one failed, and then stick a new one in and
have it start a new mirror again.  Well, that didn't work and so now
nothing from the last week has backed up, since I got it running again,
I haven't messed with it because I'm afraid that it's not going to start
up again.  If I can make it to Christmas break, I'm willing to put the
time into it to do this "right".

All of that being said, my questions are:  How do I get back to a
working RAID system?  Do I need to set up RAID the old fasioned way with
two others disks in this thing and just copy over the whold system to
them?  Should I give up on software RAID-1?

Thanks for any help.  As usually, it really is appreciated.  That seems
like a feeble cop out to a list that's been so helpful, but it's the
best I can offer.

Sincerely,
Ryan Booz
Tech Coordinator
Belleville Mennonite School