[Mailman-Users] Re: qrunner running away

[Mark Symonds]

Hello, 

Previously posted problem below has been resolved.  Scrutiny 
of the logs revealed that qrunner was complaining about the
nonexistence of a file in /var/spool/mail/qfiles.  This apparently 
caused qrunner to run in circles, pegging the CPU at 100%. 

There were only a few files in there, so we rm -fr'd
/var/spool/mail/qfiles and voila, all is back to normal.  
Yay!  This was on a RH 7.1 system. 

-- 
Mark Symonds
mark at symonds.net



----- Original Message ----- 
From: "Mark Symonds" <mark at symonds.net>
To: <mailman-users at python.org>
Sent: Saturday, June 21, 2003 3:29 AM


> 
> Hello, 
> 
> Had mailman working wonderfully here until three days ago, 
> when a larger list (just over 2500 users) ceased functioning 
> and I noticed the load average was through the roof (usually 
> it's very low): 
> 
> [root at tx exim]# w
>   1:24pm  up 387 days,  2:26,  1 user,  load average: 4.21, 4.29, 4.24
> USER     TTY      FROM              LOGIN@   IDLE   JCPU   PCPU  WHAT
> mark     pts/0    ca1.symonds.net   1:19pm  0.00s  0.28s  0.01s  w
> [root at tx exim]# ps auxw |grep python
> mailman   6751 23.9  2.4 66712 12372 ?       R    Jun19 534:48 /usr/bin/python -S /var/mailman/cron/qrunner
> mailman  21895 24.8  2.6 62840 13608 ?       R    Jun19 404:09 /usr/bin/python -S /var/mailman/cron/qrunner
> mailman   3114 25.1  9.2 61756 46716 ?       R    Jun19 258:03 /usr/bin/python -S /var/mailman/cron/qrunner
> mailman  16755 24.5 11.2 61592 56844 ?       R    06:18 104:49 /usr/bin/python -S /var/mailman/cron/qrunner
> root     25531  0.0  0.1  1716  596 pts/0    S    13:24   0:00 grep python
> [root at tx exim]#
> 
> ...read FAQ 4.19, killed the qrunner processes and removed the 
> lockfiles.  Still no joy: 
> 
> [root at tx Mailman]# w
>   4:44am  up 387 days, 17:46,  1 user,  load average: 1.08, 1.07, 1.26
> USER     TTY      FROM              LOGIN@   IDLE   JCPU   PCPU  WHAT
> mark     pts/1    ca1.symonds.net   4:37am  0.00s  0.24s  0.02s  w
> [root at tx Mailman]# ps auxw |grep python
> root     11804 96.9 11.3 59256 57492 ?       R    03:59  43:54 python ./qrunner
> 
> In the logs:
> 
> [root at tx mailman]# tail qrunner
> Jun 21 06:08:03 2003 (15197) Could not acquire qrunner lock
> Jun 21 06:09:02 2003 (15209) Could not acquire qrunner lock
> Jun 21 06:10:01 2003 (15217) Could not acquire qrunner lock
> Jun 21 06:11:03 2003 (15240) Could not acquire qrunner lock
> Jun 21 06:12:02 2003 (15252) Could not acquire qrunner lock
> Jun 21 06:13:02 2003 (15265) Could not acquire qrunner lock
> Jun 21 06:14:02 2003 (15274) Could not acquire qrunner lock
> Jun 21 06:15:02 2003 (15294) Could not acquire qrunner lock
> Jun 21 06:16:02 2003 (15307) Could not acquire qrunner lock
> Jun 21 06:17:01 2003 (15319) Could not acquire qrunner lock
> [...ad infinitum...]
> 
> Also during this time: 
> 
> * Trying to subscribe via the mailing list web page, upon 
>   clicking the "subscribe" button it hangs at the page 
>   until the browser times out. 
> 
> * Same thing when trying to auth via the admin web page. 
> 
> * This is the only broken list on that machine.  The rest
>   are functioning normally in all respects. 
> 
> * I thought perhaps someone had changed the domain preferences
>   for the list?  Checked it with config_list and it is correct.  
>   Also ran check_db and check_perms, both report all's OK. 
> 
> Exim 4, RedHat 7.0, Mailman 2.0.13.  
> 
> Ideas?
> 
> Sincerely, 
> 
> -- 
> Mark Symonds
> mark at symonds.net
> 
> 
>