SA Unexpected Crash on Centos 4.4

[Rob Lloyd <ro...@hang-wire.com>]

I recently moved Spamassassin from RH9 to Centos 4.4. On RH9 it worked great, but now SA is crashing unexpectedly on Centos 4.4. I have resorted to running a cronjob which restarts SA every hour. This has helped minimize the downtime when and if there is a crash.

I would appreciate any guidance as to what I can do to prevent this from happening. Thanks in advance and here's some more info.

running Spamassassin 3.1.7 on Centos 4.4 with Sendmail
this is how I start SA as a daemon: 
/usr/bin/spamd --daemonize --username nobody --pidfile /var/run/spam.pid --socketpath /etc/mail/spamassassin/spam.sock --max-children=10 -r /etc/mail/spamassassin/spamd.pid

Procmail:
# Spamassassin
:0fw
* <300000
| /usr/bin/spamc -U /etc/mail/spamassassin/spam.sock -t 600

# if spam send to spam box
:0
* ^X-Spam-Flag: YES
/var/spool/mail/spam

Sendmail Logs around crash time:
########## problem 2 ##########
Nov 14 07:12:19 emate spamd[9898]: prefork: cannot ping 9900, file handle not defined, child likely to still be processing SIGCHLD handler after killing itself
Nov 14 07:12:19 emate spamd[9898]: prefork: killing failed child 9900 fd=undefined at /usr/lib/perl5/site_perl/5.8.5/Mail/SpamAssassin/SpamdForkScaling.pm line 137.
Nov 14 07:12:19 emate spamd[9898]: prefork: killed child 9900
Nov 14 07:51:57 emate spamc[13982]: connect(AF_UNIX) to spamd /etc/mail/spamassassin/spam.sock failed: Interrupted system call
Nov 14 07:51:57 emate sendmail[13977]: kAEFfq4u013976: to=<em...@address.com>, delay=00:10:02, xdelay=00:10:00, mailer=local, pri=31796, dsn=2.0.0, stat=Sent
Nov 14 07:52:00 emate sendmail[14358]: kAEFq0wd014358: from=<em...@address.com>, proto=ESMTP, daemon=MTA, relay=msa1-gh.centurytel.net [209.206.160.251]
Nov 14 07:52:01 emate sendmail[14275]: kAEFoxo8014274: timeout waiting for input from local during Draining Input
Nov 14 07:52:02 emate spamc[13989]: connect(AF_UNIX) to spamd /etc/mail/spamassassin/spam.sock failed: Interrupted system call
Nov 14 07:52:02 emate sendmail[13984]: kAEFfwYT013983: to=<em...@address.com>, delay=00:10:00, xdelay=00:10:00, mailer=local, pri=31787, dsn=2.0.0, stat=Sent
Nov 14 07:52:08 emate spamc[13996]: connect(AF_UNIX) to spamd /etc/mail/spamassassin/spam.sock failed: Interrupted system call
Nov 14 07:52:08 emate sendmail[13991]: kAEFg68W013990: to=<em...@address.com>, delay=00:10:01, xdelay=00:10:00, mailer=local, pri=51369, dsn=2.0.0, stat=Sent
Nov 14 07:52:09 emate sendmail[14283]: kAEFp7fT014282: timeout waiting for input from local during Draining Input
Nov 14 07:52:12 emate sendmail[14365]: kAEFqBQE014365: from=<em...@address.com>, bodytype=8BITMIME, proto=ESMTP, daemon=MTA, relay=lists.tnr.com [209.212.93.45]
Nov 14 07:52:13 emate spamc[14003]: connect(AF_UNIX) to spamd /etc/mail/spamassassin/spam.sock failed: Interrupted system call
Nov 14 07:52:13 emate sendmail[13998]: kAEFgCOF013997: to=<em...@address.com>, delay=00:10:00, xdelay=00:10:00, mailer=local, pri=30708, dsn=2.0.0, stat=Sent
Nov 14 07:52:19 emate sendmail[14372]: kAEFqHM0014372: from=<em...@address.com>, proto=SMTP, daemon=MTA, relay=p54840545.dip0.t-ipconnect.de [84.132.5.69]
Nov 14 07:52:20 emate spamc[14012]: connect(AF_UNIX) to spamd /etc/mail/spamassassin/spam.sock failed: Interrupted system call
Nov 14 07:52:20 emate sendmail[14007]: kAEFgJnY014005: to=<em...@address.com>, delay=00:10:00, xdelay=00:10:00, mailer=local, pri=108912, dsn=2.0.0, stat=Sent








-rob lloyd
Hang Wire LLC
W. 206.722.3802
partners@hang-wire.com
www.hang-wire.com