[Bug 3052] New: Using spamd with metalog seems to be broken

[bu...@bugzilla.spamassassin.org]

http://bugzilla.spamassassin.org/show_bug.cgi?id=3052

           Summary: Using spamd with metalog seems to be broken
           Product: Spamassassin
           Version: SVN Trunk (Latest Devel Version)
          Platform: PC
               URL: http://bugs.gentoo.org/show_bug.cgi?id=26694
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P5
         Component: spamc/spamd
        AssignedTo: spamassassin-dev@incubator.apache.org
        ReportedBy: spamassassin-contrib@msquadrat.de


This was reported via the Gentoo bug database and I see it too every day. See 
the given URL for the full, background. The summary: 
 
When spamd logs via metalog, the logs look like this (random snippet from my 
maillog): 
spamd[11109]: [info] setuid to mss succeeded_ 
spamd[11109]: [processing message <40...@wanadoo.fr> for mss] 
1000._ 
spamd[11109]: [clean message (-3.4/5.0) for mss] 1000 in 4.1 seconds, 2950 
bytes._ 
spamd[4275]: [info] setuid to mss succeeded_ 
spamd[4275]: [processing message <20...@comcast.net> 
for mss] 1000._ 
spamd[4275]: [clean message (-4.9/5.0) for mss] 1000 in 4.5 seconds, 3117 
bytes._ 
spamd[3503]: [info] setuid to mss succeeded_ 
spamd[3503]: [processing message <20...@comcast.net> 
for mss] 1000._ 
spamd[3503]: [clean message (-4.9/5.0) for mss] 1000 in 3.6 seconds, 3400 
bytes._ 
spamd[5855]: [info] setuid to mss succeeded_ 
spamd[5855]: [processing message <40...@ant.uni-bremen.de> for mss] 
1000._ 
spamd[5855]: [clean message (-4.9/5.0) for mss] 1000 in 5.4 seconds, 3214 
bytes._ 
spamd[281]: [info] setuid to mss succeeded_ 
spamd[281]: [processing message 
<Pi...@shrek.tuiasi.ro> for mss] 1000._ 
spamd[281]: [clean message (-4.9/5.0) for mss] 1000 in 5.0 seconds, 4374 
bytes._ 
 
I guess the reason why this happens is that spamd doesn't call openlog() but 
starts with syslog(). See also the FIXME comments by me and Justin in the 
sources :) 
 
I'll check if using openlog() before syslog() really helps. Even if it doesn't 
we should do this anyway (thats also what 'perldoc Sys::Syslog' says) but 
Justin said he remembered some Sys::Syslog related Perl crashs. Are there any 
more datails on this, was it maybe related to Perl <= 5.005 which we don't 
support anymore anyway?



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.