[Mailman-Users] Delay in mailman passing messages
Brad Knowles
brad.knowles at skynet.be
Wed Jul 21 03:20:08 CEST 2004
At 7:52 PM -0400 2004-07-20, Spencer Yost wrote:
> Specifically my problem is there is a delay of anywhere from 15 minutes to
> several hours in Mailman's processing of a message. Mail receives the
> message but then takes a long time passing it to sendmail. I am not
> talking about delivery time - That can take an additional 5 hours but is
> a different issue. I am talking about the time it takes for Mailman to
> process the message. After my signature are "Received:" lines from example
> messages.
What version of Mailman are you running?
Have you looked at
<http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq04.041.htp>
and confirmed that you're not having any of these problems?
> 1.I have tried the "Run a sendmail receiver just for Mailman" suggestion
> found in the FAQs. Nothing changes in the delay.
If the delays are internal to Mailman, then this won't make a difference.
> 2.Have tried running multiple copies(4 qrunners) of the qrunners. No
> change.
This kind of thing works for sendmail servers, where you can run
multiple parallel sendmail queue runner processes. I am not
convinced that this strategy works for Mailman, however.
> 3.Virtually no change in behavior in the delay between a 450 MHZ server
> running 6.1 Linux with 1 meg memory to 3 GHZ server running 9.0 Linux with
> 1 meg memory.
1MB of memory?!? Surely you mean something else -- maybe 1GB?
Shortage of RAM can be a serious problem. If you don't have
enough RAM, you can pretty much forget about everything else.
> 4.Nothing out of the ordinary about the messages with the longest delays as
> best I can tell. They come form all over the globe from all kinds of
> email systems and encompass all types of emails. There are no errors in
> the error logs.
That sounds like a systemic problem.
> 6.A total of 8 lists that are small: Biggest is 500 subscribers with 30-50
> messages a day
Assuming the rest of the machine is configured correctly, that
should be a walk in the park, at least as far as Mailman is
concerned. I've run larger lists on smaller servers, and until
recently all the python.org mailing lists (including all the
mailman-related lists) were running on a smaller machine, and with
way more recipients.
Keep in mind that when we talk about mail performance tuning, a
lot of the same issues are applicable to Mailman as they are
sendmail, or postfix, or exim, or whatever. There's lots of good
stuff at
<http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq06.003.htp>
which you should take a look at, but in particular disk configuration
and filesystem layout can be two huge bottlenecks.
Your best reference here would be to get a copy of Nick
Christenson's book _sendmail Performance Tuning_ and apply the
lower-level techniques to your situation, and leave out just the
sendmail-specific stuff. You could apply the sendmail-specific stuff
at a later time, but the rest should be applicable to any related
process, such as Mailman.
--
Brad Knowles, <brad.knowles at skynet.be>
"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755
SAGE member since 1995. See <http://www.sage.org/> for more info.
More information about the Mailman-Users
mailing list