On Fri, 5 Mar 2004, Sasha Pachev wrote:
Being hacked is a possibility, and it could have happened in conjuction with something else, but I doubt that hacking alone with give you the panic. Most hackers are not very interested in destroying your data per say unless you tick them off - they just want to control the box. And even if they feel destructive, rm -rf / or dd if=/dev/zero of=/dev/hda makes a lot more sense and requires a lot less work than introducing a weird problem into the kernel.
The logs he quoted indicate a breakin almost certainly. Kernel errors could be a side-effect of the original hack or later, poorly designed measures.
I really like that he wants to track down what the original break was, but I'm not coming up with too many ways of doing that after the fact. Seems like the best we could do would be to do forensics on the current state of the machine - what binaries have changed and whatnot. But that doesn't necessarily have any bearing on the original attack.
I don't suppose you ever ran tripwire on that machine?
-J
No, it was only a development machine, so no serious hardening was ever done. It wasn't set up by me directly anyway, so I can't speak to any security measures beyond basic firewalling.
As to the RAM and disk suggestions, it could be, but it has run perfectly fine for a couple of months now, and e2fsck didn't indicate anything wrong, nor did memtest86. I know that is not conclusive as Sasha said, but for a consistently repeatable (and repeating) error, you would think it would show up under a good load test like memtest.
No panics when booted off of the rescue CD. No errors at all, that I can find.
I will be heading down there in a few minutes. I will try booting off a Knoppix CD and see if I can cause anything to fail that way. That should at least minimize the likelihood that it was hardware failure. I will also look at the rest of the logs and see if I can find anything interesting.
Just in case I didn't make it clear before, I can still boot and get a shell, I just can't do much with it :-). The EIP output repeats every few seconds, reportedly. Segfaults on most commands (ls and the like).
It may not be worth the effort. I just figured someone somewhere had a way of interpreting the repeating kernel failure messages, outside of ksymoops on the local machine.
Maybe when I see it for myself I will be able to get a better handle on what happened / is happening.
Oh well.
I do appreciate the help,
Adam Augustine
____________________
BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________
List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
