Matt Kettler wrote on Tue, 07 Oct 2003 10:57:08 -0400: > I was refering to the eating memory problem. And in general I'd expect > bayes to cause some kind of slow increase too.
No, I really can't see any increase over the time. It stays almost exactly at the same level over days and weeks. > > However, even with it being a sudden jump from 23m to 800m, unless someone > else has a better idea, I'd still try turning off various features one at a > time. This way you can at least start eliminating subsystems as the source > of the problem and start narrowing in on the cause. Yes, I understand what you mean, that's a standard procedure. The problem just is that the problem is very erratic and I don't know if or when it could happen again. Should I wait one week, two, three, four before I switch on a feature again? I really don't want to run it four weeks without Bayes. Also, as you see from my local.cf, I don't have many features activated, so there isn't too much to deactivate. I keep it now running with Bayes on and wait if it happens again. Then I switch off Bayes at least as long as it took to repro. > > At this point, it looks like some kind of wild memory leak that occurs > while processing a message of some unusual sort, and that could be almost > anywhere in the SA code. It could be some part of bayes, the AWL, the HTML > parsing, base64 decoding, or pretty much anything that gets called on a > per-message basis. > I definitely agree :-) That's why I asked here, I didn't really see a good path of narrowing it down quickly. Thanks for your suggestions! Kai -- Kai Schätzl, Berlin, Germany Get your web at Conactive Internet Services: http://www.conactive.com IE-Center: http://ie5.de & http://msie.winware.org ------------------------------------------------------- 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
