[Mailman-Users] run away process - cannot flock

[Richard Barrett]

On Saturday, July 19, 2003, at 05:21 PM, Paul H Byerly wrote:

>
>      Twice now my CPU has pegged after a message went through Mailman.  
>  Top shows a python2.2 process belonging to Mailman.  Shutting Mailman  
>  down stops it, but it comes right back when Mailman is restarted.   
> Killing the process ends the problem.
>      In digging through the logs I found similar entries in  
> mailmn/logs/smtp-failure:
>
> Jul 17 23:06:57 2003 (2643) delivery to foo at gloryworks.com failed with  
> code 451: 4.0.0 cannot flock(/etc/mail/access.db, fd=7, type=1,  
> omode=37777777777, euid=0): No locks available
>
> Jul 18 22:30:49 2003 (22481) delivery to foo2 at cookteam.com failed with  
> code 451: 4.0.0 cannot flock(/etc/mail/access.db, fd=7, type=1,  
> omode=37777777777, euid=0): No locks available
>
>      As this is the only two times I have such an entry and each  
> proceeds the runaway process to the minute, I assume it's related.   
> Google produced nothing that enlightened me.
>

The excessive CPU usage seems likely to result from MM's  
over-enthusiasm for retrying when the outgoing MTA reports temporary  
failures. This can manifest in a variety of circumstances.

Barry Warsaw has made changes in the CVS (see  
http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/mailman/mailman/Mailman/ 
Queue/RetryRunner.py) to curb this over-enthusiasm but I do not know if  
they are available as a separate patch or when MM 2.1.3, which will  
presumably incorporate the changes, will be available.

The primary trigger, from your log entries, is apparently a Sendmail  
problem causing a 451 SMTP response but I'll leave comment about that  
to an expert.

> Mailman 2.1.2
> Redhat 7.1
> Python 2.2.2
>
> TIA, Paul