Friday 08 February 2008 12:53:15 Tony Finch napisał(a):
> On Fri, 8 Feb 2008, Marcin Krol wrote:
> >
> > I get very disturbing effect on one of my hosts (Debian Etch), from time
> > to time exim creates lots and lots of (child?) processes, they all bog down
> > the CPUs and they can't finish working on some task.
> Can you try running exim -d -qf to see if one of the messages on your
> queue is causing the crash?
It seems like Exim can't finalize handling the message:
da6 /var/spool/exim/input % ps auxwwww | grep 29154
mail 29154 26.1 0.0 10560 3680 ? Rs 13:07 1:35 /usr/sbin/exim -Mc 1JNS14-0007aD-L2
da6 /var/spool/exim/input % date
Fri Feb 8 13:15:33 CET 2008
That process is still bogging down CPU eight minutes after this message has already been
delivered to the remote host:
2008-02-08 13:07:48 1JNS14-0007aD-L2 Completed
exim -d -qf doesn't help. Simply speaking, It shouldn't, really, as this message is not in
queue anymore.
This problem is persistent on both v4.68 and v4.69 (on the same host, I upgraded Exim quickly).
Regards,
Marcin Krol
