EXIM and Spamassassin fails: "spam acl condition: cannot parse spamd output"!
Hello:
This is quite a strange error...
Exim does log:
[QUOTE]2013-11-29 14:29:26 1VmO8J-0000bI-MQ spam acl condition: cannot parse spamd output
2013-11-29 14:29:26 1VmO8J-0000bI-MQ H=gateway02.websitewelcome.com [x.x.x.x]:59424 I=[y.y.y.y]:25 Warning: "SpamAssassin as avilesyr detected message as NOT spam ()"
At the same time, Spamd does log: [QUOTE]Nov 29 14:29:18 hyperserver spamd[1094]: spamd: result: Y 67 - BAYES_99,DCC_CHECK,HTML_MESSAGE,LOTS_OF_MONEY,NORMAL_HTTP_TO_IP,SPF_PASS scantime=3.3,size=58542,user=ramondiz,uid=725,required_score=5.5,rhost=localhost,raddr=127.0.0.1,rport=58496,mid=,bayes=1.000000,autolearn=spam,shortcircuit=no Nov 29 14:29:19 hyperserver spamd[1005]: prefork: child states: II Nov 29 14:29:26 hyperserver spamd[1094]: spamd: connection from localhost [127.0.0.1] at port 58536 Nov 29 14:29:26 hyperserver spamd[1094]: spamd: setuid to avilesyr succeeded Nov 29 14:29:26 hyperserver spamd[1005]: prefork: child states: BI Nov 29 14:29:26 hyperserver spamd[1005]: spamd: handled cleanup of child pid [1094] due to SIGCHLD: exit 1 Nov 29 14:29:26 hyperserver spamd[1005]: spamd: server successfully spawned child process, pid 2323 Nov 29 14:29:26 hyperserver spamd[1005]: prefork: child states: II Nov 29 14:29:43 hyperserver spamd[1627]: spamd: connection from localhost [127.0.0.1] at port 58604 Nov 29 14:29:43 hyperserver spamd[1627]: spamd: setuid to cpanel succeeded Nov 29 14:29:43 hyperserver spamd[1627]: spamd: processing message for cpanel:32001
Have a look at spamd, first accepting the connection; then, the child die and the spamd parent proccess spawn a new one... Not a question of idle threads... other example: [QUOTE]Nov 29 14:56:02 hyperserver spamd[12524]: prefork: child states: IIIIII Nov 29 14:56:04 hyperserver spamd[12565]: spamd: connection from localhost [127.0.0.1] at port 34890 Nov 29 14:56:04 hyperserver spamd[12565]: spamd: setuid to cycloes succeeded Nov 29 14:56:04 hyperserver spamd[12565]: spamd: checking message for cycloes:603 Nov 29 14:56:04 hyperserver spamd[12524]: prefork: child states: BIIIII Nov 29 14:56:05 hyperserver spamd[12524]: spamd: handled cleanup of child pid [12565] due to SIGCHLD: exit 1 Nov 29 14:56:05 hyperserver spamd[12524]: spamd: server successfully spawned child process, pid 14342 Nov 29 14:56:05 hyperserver spamd[12524]: prefork: child states: IIIIII
This is happening not for all emails, but randomly... I have captured for myself (from the maildir) the email causing child death.... then: [QUOTE]/usr/local/cpanel/3rdparty/perl/514/bin/spamassassin -t < emailquefallabosco
And SA does right... just print out the scanning information... plenty... as usual. What do you think it could happen with SA? Thanks.
At the same time, Spamd does log: [QUOTE]Nov 29 14:29:18 hyperserver spamd[1094]: spamd: result: Y 67 - BAYES_99,DCC_CHECK,HTML_MESSAGE,LOTS_OF_MONEY,NORMAL_HTTP_TO_IP,SPF_PASS scantime=3.3,size=58542,user=ramondiz,uid=725,required_score=5.5,rhost=localhost,raddr=127.0.0.1,rport=58496,mid=,bayes=1.000000,autolearn=spam,shortcircuit=no Nov 29 14:29:19 hyperserver spamd[1005]: prefork: child states: II Nov 29 14:29:26 hyperserver spamd[1094]: spamd: connection from localhost [127.0.0.1] at port 58536 Nov 29 14:29:26 hyperserver spamd[1094]: spamd: setuid to avilesyr succeeded Nov 29 14:29:26 hyperserver spamd[1005]: prefork: child states: BI Nov 29 14:29:26 hyperserver spamd[1005]: spamd: handled cleanup of child pid [1094] due to SIGCHLD: exit 1 Nov 29 14:29:26 hyperserver spamd[1005]: spamd: server successfully spawned child process, pid 2323 Nov 29 14:29:26 hyperserver spamd[1005]: prefork: child states: II Nov 29 14:29:43 hyperserver spamd[1627]: spamd: connection from localhost [127.0.0.1] at port 58604 Nov 29 14:29:43 hyperserver spamd[1627]: spamd: setuid to cpanel succeeded Nov 29 14:29:43 hyperserver spamd[1627]: spamd: processing message for cpanel:32001
Have a look at spamd, first accepting the connection; then, the child die and the spamd parent proccess spawn a new one... Not a question of idle threads... other example: [QUOTE]Nov 29 14:56:02 hyperserver spamd[12524]: prefork: child states: IIIIII Nov 29 14:56:04 hyperserver spamd[12565]: spamd: connection from localhost [127.0.0.1] at port 34890 Nov 29 14:56:04 hyperserver spamd[12565]: spamd: setuid to cycloes succeeded Nov 29 14:56:04 hyperserver spamd[12565]: spamd: checking message for cycloes:603 Nov 29 14:56:04 hyperserver spamd[12524]: prefork: child states: BIIIII Nov 29 14:56:05 hyperserver spamd[12524]: spamd: handled cleanup of child pid [12565] due to SIGCHLD: exit 1 Nov 29 14:56:05 hyperserver spamd[12524]: spamd: server successfully spawned child process, pid 14342 Nov 29 14:56:05 hyperserver spamd[12524]: prefork: child states: IIIIII
This is happening not for all emails, but randomly... I have captured for myself (from the maildir) the email causing child death.... then: [QUOTE]/usr/local/cpanel/3rdparty/perl/514/bin/spamassassin -t < emailquefallabosco
And SA does right... just print out the scanning information... plenty... as usual. What do you think it could happen with SA? Thanks.
-
Helo: Discovered: [QUOTE]loadplugin Mail::SpamAssassin::Plugin::ResourceLimits resource_limit_mem 26843545
SA proccesses were dying because of this memory limitation at local.cf0 -
Hello :) I am happy to hear you were able to resolve the issue. Thank you for sharing the solution you found with us. Thank you. 0 -
This happened to me too, I'm trying to solve this increasing memory limit in /etc/mail/spamassassin/local.cf
file:loadplugin Mail::SpamAssassin::Plugin::ResourceLimits # resource_limit_mem 536870912 (default) resource_limit_mem 1073741824
But, will this file be overwritten by a future cPanel update ? Regards!0 -
@mlopez - at the top of that file, you'll see this entry: # This is the right place to customize your installation of SpamAssassin.
so it's fine to make edits there. Just to be extra sure, I edited the resource_limit_mem value on my personal machine and then ran "/scripts/upcp --force" and the value was not reverted back to the default.0
Please sign in to leave a comment.
Comments
4 comments