After nine days another deadlock

From: Gary Mills <mills_at_cc.umanitoba.ca>
Date: Wed, 18 May 2011 18:11:47 -0500

We got another deadlock with opendkim-2.3.2 after nine days running
normally. It was up to almost 4000 threads before I noticed it. This
is from my monitoring script:

  Wed May 18 16:08:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 37692 34544 48 opendkim -x /etc/mail/opendkim.conf
  Wed May 18 16:18:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 40988 37840 191 opendkim -x /etc/mail/opendkim.conf
  Wed May 18 16:28:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 71932 68784 1371 opendkim -x /etc/mail/opendkim.conf
  Wed May 18 16:38:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 124012 120864 2612 opendkim -x /etc/mail/opendkim.conf
  Wed May 18 16:48:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 171560 168412 3788 opendkim -x /etc/mail/opendkim.conf
  Wed May 18 16:58:00 CDT 2011
      USER PID PPID VSZ RSS NLWP COMMAND
    daemon 16473 1 176100 172952 3893 opendkim -x /etc/mail/opendkim.conf

I did get another core file before I restarted the service. The
frequency of deadlocks seems to depend on the load on our e-mail
system. Lately our load has been lower than it was a month ago.

Any luck on finding the cause? Anything I can do to help?

-- 
-Gary Mills-        -Unix Group-        -Computer and Network Services-
Received on Wed May 18 2011 - 23:11:55 PST

This archive was generated by hypermail 2.3.0 : Mon Oct 29 2012 - 23:33:10 PST