[Mailman-Developers] Outgoing qrunner dies on 'None' object has no attribute...

[Martin Maechler]

This has happened with increasing intensity here, only since
running mailman 2.1.1 (since March 4 or so).
Note that the following is an informal bug report of what I
think at least two different problems.

-------------------------------------------------------------------------------
AttributeError: 'None' object has no attribute 'has_key'
Traceback (most recent call last):
  File "/usr/local/app/mailman-sfs/2.1.1/bin/qrunner", line 270, in ?
    main()
  File "/usr/local/app/mailman-sfs/2.1.1/bin/qrunner", line 230, in main
    qrunner.run()
  File "/usr/local/app/mailman-sfs/2.1.1/Mailman/Queue/Runner.py", line 59, in run
    filecnt = self._oneloop()
  File "/usr/local/app/mailman-sfs/2.1.1/Mailman/Queue/Runner.py", line 88, in _oneloop
    msg, msgdata = self._switchboard.dequeue(filebase)
  File "/usr/local/app/mailman-sfs/2.1.1/Mailman/Queue/Switchboard.py", line 151, in dequeue
    if data.has_key('rejection-notice'):
-------------------------------------------------------------------------------

Qrunner dies and is restarted -- but only 10 times when it
really dies with 

------ logs/qrunner -------------------------------------
Apr 14 11:06:24 2003 (1230) Qrunner OutgoingRunner reached maximum restart limit of 10, not restarting.
---------------------------------------------------------

I have just now set this (MAX_RESTARTS in bin/mailmanctl) to
10000 instead of 10 
(BTW, shouldn't this be configurable via mm_cfg.py as well?)

>From then on,  qfiles/out/*  starts growing and growing and
mailing list delivery becomes very slow at least in the case,
where I have two (or 4) slices Outgoing Runners.

--- Bug # 2 ---

BTW: When I do have more than one runner, things get faster,
*BUT* I sometimes get the same message delivered by more than
one of the outgoing runners.
This seems to happen (almost?) ony  in the case where the queue is
almost empty.  But it is quite embarrassing to this twice
delivered to everyone on the list.

Thank you for mailman,
and thanks for feedback / suggestions / fixes ...

Martin Maechler <maechler at stat.math.ethz.ch>	http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><