[Mailman-Users] Mailman/qrunner dies after 50 weeks

Tom Limoncelli tal at whatexit.org
Thu Sep 1 14:17:31 CEST 2005

I run a number of mailing lists on a server with a few virtual domains 
(http://www.groupname.org/mailman/listinfo & 
http://www.plts.org/mailman/listinfo) on a Solaris server.  I was always 
impressed at how mailman needed so little maintenance (I previously ran 
majordomo since around 1993, and moved to mailman in 2000).  It is one 
of the few bits of software that I run that "runs itself" after 
installation.  I'm currently using Mailman version 2.1.2.

Last week I got complaints from users that messages weren't getting 
through.  After some investigation it turns out they were right.  I also 
noticed that the mailman processes had been running for over 50 weeks.  
I did a "mailmanctl restart" and suddenly the entire backlog of email 
had been processed and everything was back to normal.

You may be wondering what my uptime was.  So I'll tell you:

newjoisey.whatexit.org% uptime
  8:03am  up 755 day(s), 11:58,  7 users,  load average: 0.86, 0.73, 0.48
newjoisey.whatexit.org% uname -a
SunOS joisey 5.9 Generic_112233-06 sun4u sparc SUNW,Ultra-2

Would it be worth it to do a "mailmanctl" every 49 weeks as a preventive 
measure?  Is there a recommended way to automate this?  I'm concerned 
that I might not remember to do it in July 2006.  I might be traveling then.

Is this a common problem with mailman?  Should I be concerned?

Tom Limoncelli
Author, The Practice of System and Network Administration

P.S.  Thanks for the excellent software.  50 weeks!  You guys rule! :-)

More information about the Mailman-Users mailing list