ss20 crashing and I'm totally stumped!
Greetings all,
I have a dual processor ss20 with Debian on it. It started out as
Potatoe about 2 years ago and has had many upgrades since then (custom
kernels, package upgrades, ..etc). This machine has been running solid
the day it was born.
Other than routine maintenance, which I do about every six months or
so, I don't really touch this box. I use it as a file share and a
mail/dns server (for home).
For some entirely unknown reason, it started crashing about a week
ago. It will stay up for 20 hours or so, then crash. I am racking my
brain trying to figure out why:
* I have not added or removed any hardware or software, nor changed
(knowingly) any configuration items on this server in at least 2
months.
* There is absolutely nothing in any log file (that I can find anyway)
that indicates a problem when it crashes. I even made a cron job to
append the date to a file every minute, so I would know exactly (to
the minute) when it crashed. Using this as a reference point - I go
through evey log file and most of them haven't had any entries for
hours. The ones that do have entries near the crash time are benign
(incoming email, dns queries, ..etc).
* I have updated all potential packages I could imagine might have
security holes (ssh, ftp, bind, exim, ...etc).
* I have closed all unneeded ports at my firewall. The remaining open
ports are: ftp, ssh, smtp, domain, and http).
* I moved off of my custom (2.2.22) kernel to a stock 2.2.20smp
kernel.
* I have removed from runlevel startup the vast majority of services
and daemons I was running in order to simplify things.
* All sun hardware diags check out fine.
* No core files or anything laying around.
This is killing me. I am not new to this and usually is cases like
this you at least have something to look at in some log somewhere.
This is a completely mute crash.
I REAAAALLY don't want to rebuild it. Can anyone suggest some ideas I
might try to get this resolved?
-Matt
Reply to: