Autor: Graeme Fowler Data: A: exim-users@exim.org Assumpte: Re: [exim] High CPU
On Tue, 2009-08-04 at 10:00 +0100, Timothy Arnold wrote: > I'm at a loss really!
Luckily there's already something available to help you debug this, and
it's called:
exiwhat
It will very likely be in the same directory as Exim itself. Run it
without options and you should get something like this (obfuscated for
very obvious reasons):
[root@host ~]# exiwhat
1647 daemon: -q15m, listening for SMTP on port 25 (IPv4) port 587
(IPv4) and for SMTPS on port 465 (IPv4)
13455 delivering 1MYGEG-0003Uu-H2: waiting for a remote delivery
subprocess to finish
13457 delivering 1MYGEG-0003Uu-H2 to **** [*.*.*.*] (*@*.*)
13883 handling incoming connection from ([*.*.*.*]) [*.*.*.*]
I=[*.*.*.*]:25
14327 handling TLS incoming connection from ([*.*.*.*]) [*.*.*.*]
I=[*.*.*.*]:25
If you don't find the PIDs of your spinning processes in that list, then
it would appear that they have dissociated from the main process for
some reason (perhaps during a log rotation).
The only time I have seen something similar to this was with a clumsy
process restart during a daily logrotate which didn't use exicyclog;
under some conditions the running child threads would end up sitting
silently trying to write into an FD which had long since disappeared.
That was a long time ago, though.