Craig Green wrote:

Scott Russell wrote:
Greets.

Running SA 3.1.0 with spamc/spamd on RHEL4 i386. I have the following options for my spamd: -d --allow-tell --max-children=30 --min-children=10 --min-spare=3 --max-spare=6 -q -x -u spamd

Today I saw the following error in /var/log/maillog:

spamd[23150]: prefork: ordered child to accept, but child reported state '1'
at /usr/lib/perl5/site_perl/5.8.5/Mail/SpamAssassin/SpamdForkScaling.pm
line 390.

I'd guess you're being bitten by:
http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4594
Short answer: try --round-robin.

Thanks for pointing this out. FWIW, --round-robin had issues too. I ran with "-d --max-children=30 --allow-tell --round-robin -q -x -u spamd" and the problem was every time rules_du_jour updated and ran "/etc/init.d/spamd restart" the restart failed. It looks like spamd reports back immediately it has exited but in reality takes it's time exiting until all children are done processing. The result is that when spamd tries to start up again the socket is in use. I would come in every morning after rules_du_jour ran to find spamd not running :)

--
Scott Russell <[EMAIL PROTECTED]>
IBM Linux Technology Center

Reply via email to