[Mailman-Users] Mailman sending duplicate msgs

[Marc MERLIN]

On Wed, Aug 08, 2001 at 03:32:18PM -0700, Amanda wrote:
> In particular I'm noticing it happen with people with very large
> (several hundred to several thousand) users. This is raising some red
> flags for me, as one of our current lists slated to be migrated to
> Mailman (if we ever get it functional) runs a permanent subscriber
> base of ~ twenty-nine thousand people. Now, it is an announce-only

Just for the record, I run about 13,000 lists on one server, with a total of
261,000 subscribed  Emails (137,000 unique)  and haven't yet had  reports of
such problems.
However, I just checked and my biggest list only has 4611 members, but I'm
not sure that list size would really  make that big a difference, while exim
is fast, it can't  always deliver 4,000+ emails in less  than one minute, so
every so often, one of my queue runner runs past its minute timeslot, and
I've never seen problems.
Aug 07 08:54:03 2001 usw-sf-list1 (8720) Could not acquire qrunner lock
Aug 07 08:55:03 2001 usw-sf-list1 (9134) Could not acquire qrunner lock
Aug 07 08:56:03 2001 usw-sf-list1 (9605) Could not acquire qrunner lock

My guess is that  it may be a problem with qrunner  locks, but having played
with  that code  a bit  (and debugged  qrunners that  were running  too long
mostly due to pipermail html archiving), I've not seen any problems with the
locking.

The one thing that does worry me a bit, is seeing these every so often:

qrunner log:
Aug 08 06:31:11 2001 usw-sf-list1 (17279) Unlinking orphaned .db file: /var/loca
l/mailman/qfiles/cd455f72612f82876d68c875f9b1eb2951e6dc71.db

error log:
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279): Traceback (innermost last):
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279):   File "/var/local/mailman/cro
n/qrunner", line 282, in ?
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279):      kids = main(lock)
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279):   File "/var/local/mailman/cro
n/qrunner", line 202, in main
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279):      os.unlink(root+'.db')
Aug 08 06:31:11 2001 usw-sf-list1 qrunner(17279): OSError :  [Errno 2] No such f
ile or directory: '/var/local/mailman/qfiles/cd455f72612f82876d68c875f9b1eb2951e
6dc71.db'


Marc
-- 
Microsoft is to operating systems & security ....
                                      .... what McDonalds is to gourmet cooking
  
Home page: http://marc.merlins.org/   |   Finger marc_f at merlins.org for PGP key