[Mailman-Users] Problem with long term mailman setup.

Mark Sapiro msapiro at value.net
Thu Jul 13 16:10:00 CEST 2006


Paul Rubin wrote:
> 
>Yesterday at 9:40 in the morning mail just stopped flowing.  Messages came
>were stored in the archive and never sent out.  I tried restarting
>everything and looking for locks, etc.  There were no messages in the log to
>indicate anything out of the ordinary.  
> 
>The only thing I could find strange was the the qfiles/out directory had
>hundreds of file in it.


So apparently your OutgoingRunner either died or somehow quit
processing. What was in Mailman's 'qrunner' log?


>Finally with no other ideas I Upgraded from Mailman 2.1.6 to Mailman 2.1.8
>and mail started flowing.  When 2.1.8 was started there were a large number
>of errors:
> 
>Jul 12 17:46:08 2006 (22768) SHUNTING:
>1152740247.566874+dc88e771186bc6266faeda03363eb2b3a3076cd2
>Jul 12 17:46:08 2006 (22768) Uncaught runner exception: 'str' object has no
>attribute 'get_sender'
>Jul 12 17:46:08 2006 (22768) Traceback (most recent call last):
>  File "/var/mailman/Mailman/Queue/Runner.py", line 111, in _oneloop
>    self._onefile(msg, msgdata)
>  File "/var/mailman/Mailman/Queue/Runner.py", line 159, in _onefile
>    sender = msg.get_sender()
>AttributeError: 'str' object has no attribute 'get_sender'
>
>After that some mail went out and some was lost.


There was some problem with queue entries. Those that were shunted for
some reason have an unparsed (raw) message instead of a
Mailman.Message.Message object but don't have the _parsemsg flag set
in the message data to indicate this.

The 'lost' messages are probably in the shunt queue. You can examine
these with bin/show_qfiles or bin/dumpdb. and reprocess then with
bin/unshunt, however if they are defective, the error will probably
recur.

 
>Now this morning, after sending half a dozen messages, it appears to be
>happening again.  The seventh message showed up in the archive and has not
>apparently gone anywhere else.


Is it in the out queue. Is OutgoingRunner running?

Check all of Mailman's logs and check if OutgoingRunner is running. You
might try 'bin/mailmanctl stop' followed by 'bin/mailmanctl start'.


>Is there any information from the mailman daily status report that might
>provide any clue to what is happening?


Only the "Queue Directory Contents" listing, but that will just tell
you where the messages are stuck, not why.

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




More information about the Mailman-Users mailing list