[milters] Archive

Lists Index Date Thread Search

Article: 835
From: Anthony Howe
Date: 2010-07-01 05:15:02 -0400
Subject: Re: milter-spamc

On 25/06/2010 23:07, David L. Aldridge whispered from the shadows...:
> In this particular gateway cluster, I have 3 servers.  Two running
> FreeBSD 8 and one running 7.2.  It might be interesting to note that
> RES memory is more than twice as high on the v8 servers than on the
> v7.  Although higher, allocation, deallocation and or garbage
> collection were taking place because the number grew and declined as
> one might expect.

Another user has noted a problem in using milter-ahead on FreeBSD 8
under a VMware. Are you using FreeBSD 8 under a VM?

> I have not tried to examine the core dump because I don't know enough
> to do so effectively.

With a core dump, saved in work-dir, all you need to do is

# gdb milter-spamc $core_file_path
(gdb) where
(gdb) quit

Note the type of failure reported by gdb (seg. fault, bus, abort, etc)
when it starts. With the stack trace output it should point me in a
direction to look. This assumes you configured libsnert at build time
with --enable-debug option.

> I know that moving to FreeBSD 7 is the easy answer, but, since it is
> EOL, I would really like to use 8.

I agree, but it sounds really strange that this happens only because of
a switch to FreeBSD 8, when the milters have more or less been fine on
FreeBSD 4 through 7.

-- 
Anthony C Howe            Skype: SirWumpus                  SnertSoft
+33 6 11 89 73 78       Twitter: SirWumpus      BarricadeMX & Milters
http://snert.com/      http://nanozen.info/     http://snertsoft.com/

Lists Index Date Thread Search