I'm trying to figure out why auto-expiration of bayes tokens is not
working here.... local.cf contains

[EMAIL PROTECTED] spamassassin]# cat /etc/mail/spamassassin/local.cf | grep
bayes
# bayes
use_bayes                               1
bayes_auto_learn                        1
bayes_auto_learn_threshold_nonspam      0.5
bayes_auto_learn_threshold_spam         7.0
bayes_min_ham_num                       200
bayes_min_spam_num                      200
bayes_path                              /var/spool/bayes/
bayes_auto_expire                       1
bayes_expiry_max_db_size                150000
bayes_journal_max_size                  102400
bayes_learn_to_journal                  0


sa-learn --dump shows me when the last expiry time is...

[EMAIL PROTECTED] spamassassin]# sa-learn --dump magic | grep expir
0.000          0 1065632593          0  non-token data: last expiry
atime
0.000          0      43200          0  non-token data: last expire
atime delta
0.000          0     107897          0  non-token data: last expire
reduction count

[EMAIL PROTECTED] spamassassin]# echo "select FROM_UNIXTIME(1065632593)" |
mysql
FROM_UNIXTIME(1065632593)
2003-10-08 12:03:13

I run a cron job to --force-expiry every 4 hours because I cant get it
to do it automatically, that is why it last expired at Noon.  If I look
at my spamd debug log, I see this....

[EMAIL PROTECTED] spamassassin]# cat /var/log/spamd/spamdlog | grep 'expiring
old tokens'
2003-10-08 12:29:10.620576500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:29:11.054307500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:34:19.384717500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:39:30.096086500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:44:41.260125500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:49:51.749513500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 12:55:03.858093500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:00:16.542512500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:05:33.742631500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:10:52.135959500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:15:56.256611500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:15:56.260592500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:21:13.807579500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:21:14.547465500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:21:14.767032500 debug: Syncing Bayes journal and expiring
old tokens...
2003-10-08 13:26:30.078650500 debug: Syncing Bayes journal and expiring
old tokens...

so about every 5 minutes or so, it tries to expire... but I have
bayes_expiry_max_db_size set to 150,000 and the number of toks is
currently 582,057

[EMAIL PROTECTED] spamassassin]# sa-learn --dump magic | grep ntokens
0.000          0     582057          0  non-token data: ntokens

This causes my bayes db to grow 20 to 30mb in size and puts a
performance hit on the server.

Does "non-token data: last expire atime delta = 43200" prevent the
expiry from occuring unless it's 43,200 seconds later than the last
expiry?  If this is the case, is that setting modifiable?  Or do i need
to --force-expire manually more often than every 4hours..  this is a
high traffic spamd only box... doing 12-16 concurrent spamd children at
a time.

ideas welcome!

thx,
dallas





-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
SourceForge.net hosts over 70,000 Open Source Projects.
See the people who have HELPED US provide better services:
Click here: http://sourceforge.net/supporters.php
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to