Quoth Trevor Sky Garside on Tue, Jan 29, 2002 at 17:17:14 -0800
> Monday afternoon, the system began to die. I tried just about everything I
> could think of
But not running exim in debug mode and finding out what causes the
prblem?...
-- put as many limiting options into the config file I could,
> removed Perl from the binary (had to change the database a bit, but that's
> ok), even tried running the original binary I had built for the old server.
> Eventually, I even tried the original mail server. All with no luck. I
> would start up Exim, and within 2-3 minutes, I would have 5-10 processes
> taking anywhere from 80MB to 300MB of system RAM. Then, the system (with
> 500MB of RAM and 2GB of swap) would begin to utilize its swap space, and
> would slow down terribly.
*sigh* So manytimes I have answered that one...
> Does anybody have any idea where I should look?
exiwhat tells you which process does what. Look at your spool and see if
there are any large files. See if the processes are always happening on
one domain, user, whatever. Run mailq and see what that one gives you.
Have a look at the log files, what do they say? Not just the exim ones,
the others as well.
If all else fails, run exim in the forground and watch the debugging
output untill it stops -- cause the system gets overloaded. Then, work
back and see where exim starts to behave oddly.
If that fails, re-write your whole configuration testing it as you go
and see where the provlem appears.
If that fails, strace and debugging in full is your last chance.
--
yann@??? -=*=- www.kierun.org
PGP: www.kierun.org/pgp/key-kierun
IRC: nick kierun, server spod.uk.amiganet.org, channel #sanctus
