Totally sprious problem, suspecting samba as the culprit
I hope someone can help me as I am running flat out of ideas here.
I have a box running slackware 12.0, acting purely as a file server with samba. The hardware is all new, and it was running happily as larry for about 6 weeks. Recently at random times, the box just falls over. The syslog is just full of page faults on everything: httpd, syslogd, smbd, nmbd, everything.
So I took all the files off the server, and put them onto local machines, so the box was redundant, it was just sitting there, the only user on it was me logged in as root via SSH watching Top.
At 7.09 (GMT) it fell over again. No one was connected to it but me, as it crashed NMBD was being really silly:
1641 root 20 0 6040 1544 972 R 100 0.0 30:59.84 nmbd
with a load average of 9.00, 8.75. 7.07
Now as I understand it NMBD is part of samba, and this is where I get really flaky on my knowledge, I am no sysadmin by any means of the standard, but with samba I am at a loss.
I am trying to get the box back up again, but so far it will only respond to pings. If I can get it back up I can get at the logs again, otherwise I only have the logs from the previous crash. I can post these up if it is helpful.
After it crashed for the third time (before this mornings silly load average), I downloaded the latest kernel from kernel.org, and compiled a new kernel and put that in, the box was then happy for about 36 hours, then this turn of events started. I have run a memtest live disc to check the memory and that passes. I also left the box compiling kernels for about 4 hours to "stress test" it. It didnt falter once.
So I am now at a loss, and hope one of you has a bright idea to stop me seeing if the box is waterproof and buoyant!!
Many thanks in advance....
EDIT: Just got back on SSH, getting the logs now. Will post snippets up.....