The mailing list for listmasters using Sympa

[Mark Sallee <address@concealed>]

About every 2 days, sympa will stop processing messages and postfix mail
log reports messages like this:

sympa postfix/smtp[718]: 03540138A9: to=<address@concealed>,
relay=mail.server.edu[192.116.xx.xx], *delay=1*, status=sent (250 Ok:
queued as BA6B1208128)

Previous process has died unexpectedly.
Date of crash : 22 Oct 2008  16:13
Errors:
Warning: Unrecognized variable: "tag"
Out of memory!

 and the "delay=#" will go very high as time goes on. No messages go out
at all until we stop and restart the sympa services (especially
sympa.pl). Then we get the message “Previous Process crashed” - “Out of
Memory” sent to the listmaster and then the queued messages are sent. 
Specs: Sympa 5.4.3, SLES10 server with 4 GB RAM, queries Oracle 10g,
Perl  v5.8.8, Postfix, some lists with ~25,000 subscribers.  We are
relaying our mail through our central mail server.

Is task_manager.pl supposed to be calling sympa.pl periodically but
perhaps not working?

Should we go back to an earlier version of Sympa that is more stable? 
Our sympa’s performance is so unpredictable right now.
Thank you

-- 
Mark Sallee