[Mimedefang] Mimedefang errors: What might be the cause?
Kris Deugau
kdeugau at vianet.ca
Mon Jan 16 11:30:03 EST 2006
Rick Aliwalas wrote:
> Are MX_MINIMUM and MX_MAXIMUM set to the defaults? It's says in my config
> file (/etc/sysconfig/mimedefang) that "The default value of 2 is probably
> too low". I set them as follows:
>
> MX_MINIMUM=4
> MX_MAXIMUM=30
>
> I got the same errors as you when my machines were under load and bumping
> these values made everything happy happy.
Watch that you don't run yourself out of memory by bumping these up.
Old estimates for "memory per slave" were ~16M IIRC, but I think I saw a
recent recommendation to allow ~40M per slave. If you're running a trim
mimedefang-filter with no SA or other addon Perl modules, 40M per
slave is probably quite pessimistic and you can probably reduce that to
~20-25M... but with SA3.x, it may even be too optimistic and you'll
need to allow 50M+.
Tuning sendmail's load-limiting options also helps keed MD from hitting
its limits; check through your sendmail.mc for these options:
define(`confQUEUE_LA', `2')dnl
define(`confREFUSE_LA', `7')dnl
define(`confCONNECTION_RATE_THROTTLE',`1')dnl
define(`confDELAY_LA', `2')dnl
define(`confBAD_RCPT_THROTTLE',`2')dnl
(These are numbers from a PII/400 192M server hosting ~30 domains, tuned
slowly over a week or so while watching the mail logs. I've only ever
had to take more drastic action once - a badly broken remote system
which initiated SMTP connections ~5 times per second without actually
doing anything with them. They got an iptables DROP entry.)
There may be other sendmail-level settings that can/should be tuned as
well; if you don't have a copy of the Bat Book, GET ONE! I've found it
invaluable in tracking down some of these options.
-kgd
More information about the MIMEDefang
mailing list