[Mimedefang] High load problems & sendmail question

[Michael Genrich]

My mimedefang-multiplexor setup would occasionally spit out 4.7.1
rejections on large emails.  The solution was to increase the timeouts
for mimedefang-multiplexor and sendmail -- the SMTP sessions would time
out before MIMEDefang's scanning was complete.

For mimedefang-multiplexor, this can be done through the -b and -i flags
for busy and idle timeouts, respectively.

For sendmail, you'll want to change the values in your sendmail.mc file
in this line:

INPUT_MAIL_FILTER(`mimedefang',`S=unix:/var/spool/MIMEDefang/mimedefang.sock, 

F=T, T=S:5m;R:5m')

The timeouts are set as T=S:5m;R:5m (5 minutes in this case).  Then
regenerate your sendmail.cf, restart sendmail and the multiplexor, and
watch those 4.7.1's go away.

I use 5 minute timeouts here, but I can't tell you what the "optimum"
timeout value might be for your configuration.

-mg

-- 
Michael Genrich  | System Administrator  | The Jackson Laboratory
mpg at jax.org      | http://www.jax.org    | Bar Harbor, Maine  USA


Richardson M. wrote:
>  I used the initscript to start mimedefang and it indeed starts the
> multiplexor now.. The load is down to acceptable values but some weird
> rejects (4.7.1) are generated on outgoing mails but not all.. ?
> 
> I will look into that in a moment, first it's time to eat, but one question
> came up while quickly examining the initscript:
> Shouldn't sendmail be reconfigured to listen to
> /var/spool/MIMEdefang/mimedefang-multiplexor.sock ??
> 
> Hmm the rejects are kind of weird as they come from the internal domain