On Wed, 27 Sep 2000, Michael Stevens wrote:
> If you *do* want to keep an eye on what's going on, I can recommend the
> eximon tool.
I do two things.
(1) I have a daily cron job that mails me if it finds the panic log not
to be empty.
(2) I run an eximon when I'm at work, to keep an eye on the queue
length. When I'm not around, I let it run on its own.
Actually, (2) is no longer true. I used to do that when we had 3
servers that I was looking after. When it got to 4 I got fed up with 4
eximons. I wrote a Perl/Tk script that shows a single stripchart with
the queue total for all 4 servers, each strip using different colours
for each server. It also indicates whether there are any frozen messages
on any server, and which servers are down. The script works by using rsh
to fetch data from the servers every 2 minutes by running "exim -bp".
However, it has a small memory leak (about 40-50K per day). This I
traced to happening somewhere inside Perl/Tk, but I couldn't track it
right down. Since I never run this thing for more than a working day, I
haven't bothered.
I haven't polished the script into anything like a product, and there's
no documentation, but if anybody wants a copy they are welcome to it.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
