[Mailman-Users] Re: Mailman Using 70- 90% of CPU

[Will Yardley]

On Fri, May 02, 2003 at 10:50:58AM -0700, NOW Website Coordinator wrote:

> In answer to Will Yardley's question to my previous message:

>> What happens if you kill those processes? What about if you stop,
>> and restart mailman (via mailmanctl)?  You could also try "strace
>> -p[pid]" and see if you can see what (if anything) the process is
>> doing.  Usually, when I see really long running mailman procs like
>> that eating up so much CPU, I kill them.

> As soon as I restart Mailman, it uses up around 70-90% of my CPU
> with the outgoing Runner job.  It's not one run away job, I can
> recreate it in less than a minute.  And we haven't sent out any
> messages today, mailman has just sent some sub/unsub notices out.

Just out of curiosity, does 'strace -p[pid]' on that PID show anything
useful? It's possible that this output might be helpful in figuring
out what's going on, at least. I don't claim to be an expert on
figuring out this type of stuff, but often the text strings in there
will give you an idea of what the program is trying to do.

If you want to save the data to a file, you can do something like:
strace -p[pid] 2> strace.out

I imagine some of the developers might have a better idea of what
information would be useful, but possibly looking at the value of
'PWD' in /proc/[pid]/environ and the output of /proc/[pid]/status
might be helpful?

I've seen problems like this occasionally, but I've almost always been
able to "fix" them by just killing them or restarting Mailman, so it's
possible that there's an actual problem here.

-- 
"Since when is skepticism un-American?
Dissent's not treason but they talk like it's the same..."
(Sleater-Kinney - "Combat Rock")