[Mailman-Users] delate all bounces....

[Mark Sapiro]

kalin mintchev wrote:
>
>from the qrunner log:
>Nov 15 17:00:14 2005 (2584) Qrunner BounceRunner reached maximum restart
>limit of 10, not restarting.
>
>what's this thing about maximum restart limit?!


What's in the log prior to this for the 10 deaths that required restart?

The idea is the master mailmanctl daemon watches the qrunner processes
and if one dies (for some, but not all reasons), the master restarts
it, but once it reaches the limit of 10 (coded in mailmanctl), it
stops the restarting on the theory that this is a "hard error" and we
don't want to loop forever on it.


>that's the last record before the entry of the new restart. right now it's
>running..
>
>now there is this huge and growing bounce-events-.pck file thought. i
>guess at some point the bouncer will process it?


Bounce runner processes the queued bounces, but rather than processing
them as they come and going through the list lock, load and save
process for each bounce, they are stored in the bounce-events-.pck and
that in turn is processed once every (default, controlled by
REGISTER_BOUNCES_EVERY) 15 minutes in list order so the list is only
locked, loaded and saved once.

I still don't know what the underlying issue is, possibly you have so
many bounces that the bounce-events-.pck reaches some size limitation.
What does the 'qrunner' log say about the deaths? What's in the
'error' log.

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