You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Kevin Old <ke...@gmail.com> on 2004/09/11 00:11:59 UTC
SpamAssassin timed out errors
Hello,
Anyone have ideas why I'd be seeing these messages all of a sudden on
a 2.64 installation with no changes?
Sep 10 18:07:49 s15111287 MailScanner[28540]: SpamAssassin timed out
and was killed, failure 13 of 20
Sep 10 18:07:50 s15111287 MailScanner[28540]: Virus and Content
Scanning: Starting
Sep 10 18:07:51 s15111287 MailScanner[28540]: Uninfected: Delivered 1 messages
I'm using SURBL....are there problems with the SURBL servers? It
started happening last night and cleared up early this morning, but
has returned this afternoon.
Any ideas?
--
Kevin Old
kevinold@gmail.com
Re: SpamAssassin timed out errors
Posted by Kai Schaetzl <ma...@conactive.com>.
Kevin Old wrote on Fri, 10 Sep 2004 18:11:59 -0400:
> I'm using SURBL....are there problems with the SURBL servers?
>
Probably not. Run an "sa-learn -D --sync" to see if it's bayes-related. We
are getting the same and it's somehow related to the journal. I just saw a
170 MB journal which sounds quite ridiculous (= not possible) for a server
with only a few thousand mails arriving per day and a sync possibly
happening every 24 hours.
Also with MS, but with SA 3.0.
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
Re: SpamAssassin timed out errors
Posted by Matt Kettler <mk...@comcast.net>.
At 06:11 PM 9/10/2004 -0400, Kevin Old wrote:
>Anyone have ideas why I'd be seeing these messages all of a sudden on
>a 2.64 installation with no changes?
>
>Sep 10 18:07:49 s15111287 MailScanner[28540]: SpamAssassin timed out
>and was killed, failure 13 of 20
>Sep 10 18:07:50 s15111287 MailScanner[28540]: Virus and Content
>Scanning: Starting
>Sep 10 18:07:51 s15111287 MailScanner[28540]: Uninfected: Delivered 1 messages
>
>I'm using SURBL....are there problems with the SURBL servers?
Possible, but unlikely. The most common cause of timeouts I've seen when
using mailscanner is bayes expiry.
Check your bayes database directory to see if there's a bunch of bayes
"expire" files with PIDs as a part of their filename. If so, mailscanner's
been killing SA while it's trying to expire the bayes DB.
Options:
Disable bayes_auto_expire in your local.cf and run sa-learn
--force-expire from a daily cronjob
extend MailScanner's SA timeout to something on the order of 4 minutes in
MailScanner.conf