[Mailman-Users] qrunner dying

[Richard Barrett]

At 08:26 29/04/2003, Kysh wrote:

>Ok, I'll try to keep this as short as possible. (I have skimmed the 
>archives, I
>have dug around for months on the internet trying to find a reference. I have
>asked on the IRC channel, all to no avail)
>
>I've been running mailman for years and years now. I've been upgrading along
>the way, and am currently running 2.1.1-5.
>
>Sometime back in December (May have been around the time of an upgrade, can't
>remember), my qrunner took a vacation. I didn't notice for a few days until
>dozens of users started emailing me frantically.
>
>I've upgraded quite a few times since then, but since then:
>
>When I run maimanctl start, qrunner only lasts about three seconds.
>
>Apr 29 00:21:05 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26362, sig: None, sts: 1, class: CommandRunner, slice: 1/1) [restarting]
>Apr 29 00:21:05 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26361, sig: None, sts: 1, class: BounceRunner, slice: 1/1) [restarting]
>
>[snip] ... [snip]
>
>Apr 29 00:21:08 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26432, sig: None, sts: 1, class: ArchRunner, slice: 1/1) [restarting]
>Apr 29 00:21:08 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26438, sig: None, sts: 1, class: ArchRunner, slice: 1/1) [restarting]
>Apr 29 00:21:08 2003 (26359) Qrunner ArchRunner reached maximum restart 
>limit of 10, not restarting.
>Apr 29 00:21:08 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26437, sig: None, sts: 1, class: VirginRunner, slice: 1/1) [restarting]
>Apr 29 00:21:08 2003 (26359) Qrunner VirginRunner reached maximum restart 
>limit of 10, not restarting.
>Apr 29 00:21:08 2003 (26359) Master qrunner detected subprocess exit
>(pid: 26436, sig: None, sts: 1, class: NewsRunner, slice: 1/1) [restarting]
>Apr 29 00:21:08 2003 (26359) Qrunner NewsRunner reached maximum restart 
>limit of 10, not restarting.
>
>I tried upping the maximum restart limit to an arbitrarily high number, but
>qrunner just didn't really work at all.
>
>However, if I run /var/lib/mailman/bin/qrunner -r All, everything works 
>great.
>
>(Except that after a while, it starts eating up more and more CPU time)
>
>I've checked all the easy stuff, and I'm stumped. I don't know python that 
>well,
>and this whole thing gives me a bit of a headache. :>
>
>Any thoughts? (reinstalling is DEFINIETLY not an option, with the number of
>users and lists that I have!)
>
>-Kysh
>--

This problem is not one I have seen myself but is anything being logged to 
$prefix/logs/error.

My reading of the mailmanctl and qrunner code is that they log problems to 
$prefix/logs/error with the string 'mailmanctl' and 'qrunner' respectively 
after the time stamp on the log entries.

If the qrunner processes started by mailmanctl are exiting abnormally I 
would expect to find some evidence in $prefix/logs/error.

See what you can see there and repost with it. I'm sure someone will be 
able to help with the additional information.