A couple of weeks back my freebsd system was moved from one host
machine to another.
Ever since the move i've had some unexplainable exim failure that
I'll try to describe here. I
was hoping someone in exim-user land would recognize the issue, or at
least be able to
tell me how to better troubleshoot it.
First off, most of the time exim is up and running fine. But after
the move users started to
notice that mail from the outside wasn't arriving. Mail from users on
the same system was
mysteriously still working. Without time to troubleshoot, it appeared
as though a simple
exim restart fixed the issue. But of course it reoccurs a few days or
hours later.
My first step was to increase my reliability. So I did two things,
upgraded exim to the latest
in the ports tree, then installed monit -- a service manager/
watchdog. So now, instead of
my services going down until I realize and restart them, they are
monitored and tested and
restarted automatically.
So currently, I have monit checking the pid and smtp port
connectivity and if exim isn't
running or the connection is ever refused or fails it restarts exim.
Exim doesn't seem to
quit or crash, but the connection test eventually fails. This happens
erratically from as
short as 8 hours to as long as maybe 2 days between failures. Monit
bounces the app and
everything is fine for awhile.
mainlog, paniclog, and messages all show nothing interesting. It
appears as though exim
continues to run but simply stops allowing mail from the outside (and
doesn't even log a
rejection).
FYI: i'm running exim 4.53 with clamav and spamassassin (all
controlled by vexim).
Any thoughts? I'd be happy to post chunks of my config and/or logs,
but I don't even know
what would be pertinent at this point.
Oh, and i do know about exim -bh but since the problem is erratic and
it normally is working fine it's difficult for me to know when to do
this. I have an idea on how to catch it when it's pants are down, but
it unfortunately will disturb my service a bit so I haven't been
acting on it yet.
