Sendmail: Number of queue runners?

[DAve]

John Goggan wrote:
> I have been using MailScanner with a sendmail configuration for years now.  
> Today, I noticed that I had about 30 messages in my /var/spool/mqueue directory 
> that had not been delivered -- some of which were over a month old.  After some 
> research, it appears that almost all of them were to server that do 
> greylisting -- so all of these were the deferred messages.
> 
> Basically, I don't see anything that would be watching mqueue for these 
> messages and attempting to redeliver them.  Should I have another queue runner 
> for mqueue?  We've always run like this:
> 
> One sendmail for incoming email that goes to mqueue.in:
> 
> sendmail -bd -L sm-mta -OPrivacyOptions=noetrn -ODeliveryMode=queueonly -
> OQueueDirectory=/var/spool/mqueue.in
> 
> And one sendmail queue runner for local/clientmqueue:
> 
> sendmail -Ac -q15m -L sm-cm
> 
> We use Mailscanner with the "Delivery Method = batch" setting, so my 
> understanding is that that means that MailScanner will take care of the 
> messages after they get to mqueue.
> 
> This works fine -- except where there is a problem.  It appears that nothing 
> ever retries them.  Is that correct?
> 
> As a solution, I've added another sendmail queue runner for mqueue:
> 
> sendmail -q15m -L sm-mq
> 
> So, I now have three sendmails: one daemon listening on 25 and putting mail in 
> mqueue.in, one queue runner for local mail in clientmqueue, and one queue 
> runner for mail that might hang out in mqueue if it doesn't go the first time 
> MailScanner tells it to go.
> 
> Is that how it should be?  Almost all of the documentation or guides that I 
> could find for setting up MailScanner with sendmail don't seem to mention that 
> second queue runner for mqueue.  And, as I said, I ran for years without it 
> (and likely missed sending some delayed emails now and then that were never 
> retried).  I just want to make sure that I'm not missing something.

Yes, we use a third queue runner as well. We have several clients we do 
"scrubbing" for acting as a front end for their in house Exchange 
servers. Several of those clients have <unamed poorly functioning 
security devices> which they cannot seem to figure out how to whitelist 
my MailScanner servers through. So our MailScanner servers become 
greylisted and the messages do not go through.

Currently we run a cron task that pulls them out once and hour and drops 
them into a slow queue we process once every four hours.

DAve

-- 
Don't tell me I'm driving the cart!