[prev in list] [next in list] [prev in thread] [next in thread] 

List:       majordomo-workers
Subject:    Re: Mailing list round trip time (fwd)
From:       The Hermit Hacker <scrappy () hub ! org>
Date:       2000-03-31 4:22:33
[Download RAW message or body]

On Thu, 30 Mar 2000, Michael Yount wrote:

> On Thu, Mar 30, 2000 at 10:16:54PM -0400, The Hermit Hacker wrote:
> > >
> > > There are two possible problems...the first is that the MTA you're using tends
> > > to have delays during SMTP conversations due to DNS lookups.  The second
> > > is that sometimes the conversation between mj_queueserv and mj_queuerun
> > > will time out if the system load is high, and messages will remain
> > > in the Mj2 queue until another request arrives.  The second problem
> > > can be ameliorated by increasing the queue timeout in .mj_config
> > > and reinstalling.
> >
> > Okay, item one should be reduced by setting a delivery_rule of:
> >
> > ALL
> > sort, minseperate, maxaddrs 
> > hosts=(news.hub.org=(esmtp,pipelining,onex,timeout00))
> > backup=(hub.org=(esmtp,pipelining,onex,timeout00))
> >
> > to make for smaller packets, no?
>
> No.  Delivery is done serially, and if a particular address hangs
> during RCPT TO because of a slow DNS server, any adjustment has
> to be done in the MTA; there's nothing Mj2 can do about it.

ack ... anyway of parralelling this?  I'm laoth to look at the MTA, since
I'm actually using remote servers in delivery_rules to spread load, and I
don't necessarily have control over those servers (the servers are being
used *with* permission from their owners ... )

> > as for item two ... loadavg shouldn't be much of a concern right now, as
> > our "load intensive app" is currently offline until we throw a second CPU
> > into this thing ...
>
> Permissions are tight on your tmp/mj directory, or I'd check the logs.
> Look for lines in mj_queueserv.debug like:
>
> Queueserv: Runner did not respond.
> Queueserv: Out: Problems starting queue runner; queueing

Kinda like:

[63477].Queueserv: Out: Excessive load; queueing
[63477].Queueserv: Out: Excessive load; queueing
hub# grep queueing mj_queueserv.debug | wc -l
    8138

hub# grep Runner mj_queueserv.debug | grep -v Starting
[89415].Queueserv: Runner did not respond.
hub#

what does it consider to be 'excessive load'?

also, how much of this do I have to keep:

hub# ls | wc -l
   17358
hub# pwd
/usr/local/majordomo/tmp/mj

is it somethign that a simple 'find . -mtime +7 -exec rm {} \;' is safe to
use to clean up?

Marc G. Fournier                   ICQ#7615664               IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org

[prev in list] [next in list] [prev in thread] [next in thread]