On Friday, September 15, 2006 at 6:36:01 PM, Daryl confabulated: > Duane Hill wrote: >> Having an issue with spamd terminating abnormally. The server is >> running FreeBSD 6.0-Release, Perl 5.8.8, and SA 3.1.5.
> All the spamd processes actually terminate? Not end up like zombies or > stop responding? I do not believe so. Viewing processes with 'ps' and grep'ing for spamd yields nothing. Normally it shows something like this: [EMAIL PROTECTED] ~]$ ps auxw | grep spamd nobody 6899 18.6 5.5 57400 56716 ?? S 6:43PM 1:31.52 spamd child (perl5.8.8) nobody 7343 15.3 5.2 54220 53628 ?? S 6:51PM 0:15.46 spamd child (perl5.8.8) root 87623 0.0 4.7 49420 48852 ?? Ss 2:12PM 2:56.78 /usr/local/bin/spamd -m 15 -i 127.0.0.1 -p 783 -u nobody -g nobody - After spamd is no longer active, there isn't anything displayed for the process. I did forget to mention I was running spamd under the nobody user/group. However, I was having this same issue running without the -u and -g start parameters. >> spamd is started with the following: >> >> -d -r ${pidfile} -m 15 -i 127.0.0.1 -p 783 -s /var/log/spamd.log >> >> The ${pidfile} is set from within /usr/local/etc/rc.d/sa-spamd. >> >> I've checked through the spamd.log file created and can not find >> anything to indicate why this is happening. >> >> I've also made sure 'spamassassin --lint' comes out clean. >> >> I do use sa-update and had to remove the rules that I loaded into >> /usr/local/etc/mail/spamassassin as they were causing a conflict. >> That conflict wasn't visible until after I started logging. > What kind of conflict are you talking about here? >> If there is something I'm missing, please point out what. Thanks. > After how many messages does spamd exhibit the problem? How many > children are running? Are you using pre-forking or round-robin? > Does-rescanning the last message processed by each child, or at least > the very last message to be scanned, cause the problem to happen again > (after you've restarted spamd, of course). > Daryl -- "This message was sent using 100% recycled electrons."