[Mailman-Users] Bounce-events pck files...

Mark Sapiro msapiro at value.net
Wed Jun 8 22:19:19 CEST 2005


Glenn Sieb wrote:
>
>I was just looking in my mailman/data directory and noticed that I have 
>155 of these bounce-events-#####.pck files taking up about 24 meg of 
>space there.
>
>They date back as far as January 18th of this year. Are these garbage 
>files? Unprocessed bounce messages? What does one do with these--I 
>didn't see anything on the FAQ or FAQ-wizard...

Yes, they are mostly if not all garbage and yes, they contain
unprocessed bounce messages.

The ##### in bounce-events-#####.pck is the pid of the bounce runner
that wrote the file. Here are comments from BounceRunner.py describing
the file:

# Registering a bounce means acquiring the list lock, and it would be
# too expensive to do this for each message.  Instead, each bounce
# runner maintains an event log which is essentially a file with
# multiple pickles.  Each bounce we receive gets appended to this file
# as a 4-tuple record: (listname, addr, today, msg)
#
# today is itself a 3-tuple of (year, month, day)
#
# Every once in a while (see _doperiodic()), the bounce runner cracks
# open the file, reads all the records and registers all the bounces.
# Then it truncates the file and continues on.  We don't need to lock
# the bounce event file because bounce qrunners are single threaded
# and each creates a uniquely named file to contain the events.

If the pid is not that of a running bounce runner, the file is dead and
can be deleted. Presumably, these dead files can come about if the
bounce runner dies or mailman is stopped or restarted when a non-empty
file exists.

The contents of these files can be examined with bin/dumpdb.

--
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