[Mimedefang] "No free slaves" vs. "Queueing request"

[Andrzej Marecki]

>> Why are there *much more* messages of this kind:
>> mfconnect: No free slaves
>> ... than messages like this:
>> All slaves are busy: Queueing request (xx queued)
>> ... in my syslog?

> Because of the way queuing works.

> When a *new* SMTP session is opened, if there are no free slaves, the
> session is tempfailed and no queueing is done.

Hmmm. Unavailability of free slaves may be transient. Don't you think
a possibility of making a retry could be of any good in this context?
(I suspect such a possibility should be provided by Sendmail but apparently
isn't.) How about queuing *new* SMTP sessions via MD, anyway?

> If an *existing* SMTP session tries to do work that needs a slave, then
> the request is queued (if possible).  The idea is that existing SMTP
> sessions should get priority (because you've already expended resources
> on them.)  So we try hard to complete existing sessions, while making
> new ones retry.

All right, a kind of "priority flag" might solve the prioritization.
Alternatively, a "second order" queue might be created (which makes things 
even more complicated, I know. ;-).

All in all, tempfailing - IMVHO - is something bad and should avoided.
If possible, of course.

A.M.

-- 
-----------------------------------------------------------------------------  
Andrzej Marecki                | 
Torun Centre for Astronomy     |   e-mail: amr at astro.uni.torun.pl
N. Copernicus University       |   WWW:    http://www.astro.uni.torun.pl
ul. Gagarina 11                |   tel: +48 56 6113032
PL-87-100 Torun, POLAND        |   fax: +48 56 6113009
-----------------------------------------------------------------------------  
"If Bill Gates actually had to admin his own stuff, he'd shoot himself."
-----------------------------------------------------------------------------