[Mailman-Users] qrunners exiting

[Mark Sapiro]

George Barnett wrote:
>
>At this stage, mail starts backing up in the archive queue and isn't 
>delivered until mailman is restarted.  After having a dig through the 
>archive, I added MAX_RESTARTS = 10000 to the config file in the hope 
>that this would solve the problem, however it's continued to his this 
>limit of 10.
>
>Can anybody offer any help on why this may be happening or any debugs I 
>can enable?


MAX_RESTARTS = 10 is set locally in mailmanctl AFTER it imports mm_cfg
(and it doesn't reference mm_cfg.MAX_RESTARTS), so setting it in
mm_cfg.py won't do anything. You have to change it in mailmanctl.

Your log entries (e.g.)

>Mar 13 14:25:28 2006 (5122) Master qrunner detected subprocess exit
>(pid: 5132, sig: None, sts: 1, class: OutgoingRunner, slice: 2/2) 
>[restarting]


say the runner exited with exit status 1 and no signal caught. They
don't do this on their own without writing an 'error' log entry. In
this case, there "should be" something like

  child process leaked thru: OutgoingRunner

in the error log if OutgoingRunner called exit(1), but some runners
never call exit.

>Can anybody offer any help on why this may be happening or any debugs I 
>can enable?


There are no debugs you can turn on or off. You have to add your own.

I don't know what's happening here if there aren't any messages in the
'error' log. Just a guess, but possibly you are hitting some OS limit
on process memory or something like that. Have you checked your OS
logs?

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan