While I've not (as yet) established exactly what is wrong, I have established that there is a problem. When RulesDuJour updated my SARE rule-sets it caused spamd to fail to process all subsequent emails until it was restarted with "/etc/init.d/spamd restart". This is a (sanitised) extract from syslog showing when this problem arose (N.B. spamc continued to fail for several hours until I restarted spamd manually.) : -- Nov 15 03:20:00 svr fcron[5328]: process already running: root's /usr/bin/test -x /usr/sbin/run-crons && /usr/sbin/run-crons Nov 15 03:20:14 svr postfix/pickup[11065]: ...: uid=0 from=<root> Nov 15 03:20:14 svr postfix/cleanup[11232]: ...: message-id=... Nov 15 03:20:15 svr spamd[7808]: spamd: connection from localhost [127.0.0.1] at port 1125 Nov 15 03:20:15 svr spamd[7808]: spamd: setuid to foouser succeeded Nov 15 03:20:15 svr spamd[7808]: spamd: processing message .. for foouser:1000 Nov 15 03:20:18 svr spamd[7808]: spamd: clean message (-2.9/5.0) for foouser:1000 in 3.1 seconds, 647 bytes. Nov 15 03:20:18 svr spamd[7808]: spamd: result: . -2 - AWL,BAYES_00 scantime=3.1,size=647,user=foouser,... Nov 15 03:20:18 svr postfix/local[11237]: ... Nov 15 03:20:18 svr postfix/qmgr[5607]: ...: removed Nov 15 03:20:19 svr spamd[5462]: prefork: child states: II Nov 15 03:20:26 svr postfix/pickup[11065]: ...: uid=0 from=<root> Nov 15 03:20:26 svr postfix/cleanup[11232]: ... Nov 15 03:20:27 svr spamd[7808]: spamd: setuid to foouser succeeded Nov 15 03:20:27 svr spamd[7808]: spamd: processing message ... for foouser:1000 Nov 15 03:20:29 svr spamd[7808]: spamd: clean message (-2.2/5.0) for foouser:1000 in 2.7 seconds, 612 bytes. Nov 15 03:20:29 svr spamd[7808]: spamd: result: . -2 - AWL,BAYES_05 scantime=2.7,size=612,user=foouser,uid=1000,... Nov 15 03:20:29 svr postfix/local[11237]: EEA5F3B945: to=<foouser@shic.lan>, orig_to=<root>, relay=local, delay=3, status=sent (delivered to command: /usr/bin/proc Nov 15 03:20:29 svr postfix/qmgr[5607]: EEA5F3B945: removed Nov 15 03:20:30 svr spamd[5462]: prefork: child states: II Nov 15 03:21:05 svr spamd[5462]: spamd: server killed by SIGTERM, shutting down Nov 15 03:21:11 svr rc-scripts: Failed to stop spamd Nov 15 03:30:00 svr fcron[5328]: process already running: root's /usr/bin/test -x /usr/sbin/run-crons && /usr/sbin/run-crons Nov 15 03:40:00 svr fcron[11746]: Job /usr/bin/test -x /usr/sbin/run-crons && /usr/sbin/run-crons started for user root (pid 11747) Nov 15 03:50:00 svr fcron[11759]: Job /usr/bin/test -x /usr/sbin/run-crons && /usr/sbin/run-crons started for user root (pid 11760) Nov 15 03:50:24 svr postfix/smtpd[11772]: connect from localhost[127.0.0.1] Nov 15 03:50:24 svr postfix/smtpd[11772]: ...: client=localhost[127.0.0.1] Nov 15 03:50:24 svr postfix/cleanup[11775]: ...: message-id=... Nov 15 03:50:24 svr postfix/qmgr[5607]: 73FAA3B4FB: from=... Nov 15 03:50:24 svr postfix/smtpd[11772]: disconnect from localhost[127.0.0.1] Nov 15 03:50:24 svr spamc[11779]: connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#1 of 3): Connection refused Nov 15 03:50:25 svr spamc[11779]: connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#2 of 3): Connection refused -- Issuing '/etc/init.d/spamd restart' as root resolved the situation... but I don't want to have to do this every time a rule-set is automatically updated overnight. I enquired about this on both the gentoo user list and the spamassassin user list. On the latter I got this response: http://marc.theaimsgroup.com/?l=spamassassin-users&m=116377918014476&w=2 My spamassassin is running on humble hardware when compared with state of the art - however, with the exception of this glitch (which looks, on the surface, like a bug in waiting a static time for a shutdown) my setup performs perfectly for my purposes...
(In reply to comment #0) > Issuing '/etc/init.d/spamd restart' as root resolved the situation... > but I don't want to have to do this every time a rule-set is > automatically updated overnight. > You have to restart your spamd ! Look in your /etc/rulesdujour/config at the bottom of the file : SA_RESTART="/etc/init.d/spamd restart" and rulesdujour will auto reload your spamd process. This is a feature.
something is broken in your spamd. rules_du_jour validates the new rules and then restarts the spamd if the rules pass the spamassassin --lint check.
I experience this problem too.. I do have 'SA_RESTART="/etc/init.d/spamd restart"' on my rulesdujour config file, but still, almost every night (probably when rules are updated) spamd is killed without being reloaded..
(In reply to comment #3) > I experience this problem too.. > > I do have 'SA_RESTART="/etc/init.d/spamd restart"' on my rulesdujour config > file, but still, almost every night (probably when rules are updated) spamd is > killed without being reloaded.. > -- # grep SA_RESTART /etc/rulesdujour/config SA_RESTART="/etc/init.d/spamd restart" # spamassassin --lint # -- So, I do have /etc/init.d/spamd restart in my /etc/rulesdujour/config. My configuration does pass spamassassin --lint. I originally reported this problem and it remains a problem. I've not messed about with the configuration of rulesdujour or spamassassin - it is all vanilla portage stuff. The only thing that might be unusual about me is that I run my gentoo server on a very low-spec PC - hence if this is some kind of timeout issue, this might exhibit itself more frequently on my configuration.
spamassasin-ruledujour will be removed at the start of September. See my email to gentoo-dev ML.
