Have a Linux problem

[scauffiel]

I've got a Linux driven file server (anyone remember the PugServer?) that has finally died after six years of running flawlessly. I'm hoping it's just one of the drives in the RAID 5 array (because I see a "drive 5 offline" statement in the startup blurb) and if I find which one to replace I'll be good. However, with the cover off of the box, I've got three 300GB drives on three SATA ports (obviously), but the ports the drives are attached to at the card are 2, 3 and 4 - so which drive is 'drive 5'? The whole thing eventually hangs with no way (that's evident to ME anyway) how to get to any sort of prompt that lets me type anything, with the error:

<0>Kernel panic: Aiee, killing interrupt handler!
In interrupt handler not synching

I have no idea how to do anything in Linux, so if you're gonna help, you're gonna have to speak newbie. I've got spare drives for this one, so replacement isn't a problem. The RAID is running off of a Promise SATA 150 TX4 card; I've got a GB of memory and there's an Intel Pentium IV 2GHz running the show. The only point I get to a screen where I can do anything is during boot at a screen with "Grub version 0.93 (634k lower / 10xxxxxx something)" at the top, and the ability to choose between Fedora Core (2.4.22-1.2199.nptl) (which is the what it defaults to) or Fedora Core (2.4.22-1.2115.nptl) (which errors out much faster than the default with a different error message).

I'd appreciate any help you can offer a newbie to get this machine back up and running while keeping the data intact in the RAID. If 'drive 5' can be found and replaced I'm hoping that will fix it and be done. I just don't know where to start.

Thanks,
Steve

 

[Nothinman]

If you've only got 3 drives then there shouldn't be a drive 5, if the card shows up as SCSI then they'll be /dev/sda, /dev/sdb and /dev/sdb no matter where they're plugged in on the card. What you may be seeing is the partition number like /dev/sda5 which is the 5th partition on the first disk.

But you should see more than just the last Kernel panic line, probably something like unable to mount root filesystem. But if only one drive is failed then the array should come up in degraded mode so you may have more than just a failed drive to fix.