bayes expire tokens
maillists at CONACTIVE.COM
Wed Mar 9 22:31:33 GMT 2005
[ The following text is in the "iso-8859-1" character set. ]
[ Your display is set for the "US-ASCII" character set. ]
[ Some characters may be displayed incorrectly. ]
Brady Tucker wrote on Wed, 9 Mar 2005 06:55:50 -0600:
> You may be right in that Julian may not be able to do anything... However,
> if the problem is that the expiry process is getting interrupted --- and he
> can keep MS from trying add to the db or interrupt the expiry process then
> it would finish properly even if it took an excessive time, right ?
No. It seems when the timeout is reached the process gets called off. Finito.
> The problem is that I'm not clear on what is stopping the expiry process in
> the first place...
Ok, if I misinterpret the cause for that stopping my arguments are wrong,
> Is SA is timing out because it thinks it's taking to long
if "it" is "MS", then yes.
> -- or if MS is 'no longer stopping to wait during expiry' because it was
> taking to long and interrupting a long expiry process....
I always interpreted that setting that it means that MS will obey this if a
rebuild is in progress. F.i. via sa-learn or MS thru the SA lib. However, the
SA run gets stopped when MS reaches its internal SA timeout. There's no
separate SA process which could keep running.
> The other solution presented here to set SA to timeout at 120 seconds didn't
> work for me...
because it was too short!
> In the meantime... Forcing expire 3 times a day from a cron job fixes the
Yes, and if you set the timeout to ten minutes it will work as well (NO, don't
do it!). However, forcing an expire regularly is much cleaner and preferrable
in my eyes :-)
That you have to expire three days a time just proves that you have so many
mail running thru your system that any fewer runs would make it take too long
for the 120 sec timeout. You might consider upping the token count in your
bayes dbs as well. Not much sense in slashing half of your db each day.
Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com
IE-Center: http://ie5.de & http://msie.winware.org
------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).
Support MailScanner development - buy the book off the website!
More information about the MailScanner