[Rd] Re: Outbound e-mail slow from R servers?

Martin Maechler maechler at stat.math.ethz.ch
Thu Apr 3 19:36:12 MEST 2003


>>>>> "Marc" == Marc Schwartz <mschwartz at medanalytics.com>
>>>>>     on Thu, 03 Apr 2003 09:24:20 -0600 writes:

    Marc> Martin,

    Marc> Not sure if you are aware of this, but since yesterday
    Marc> there seems to be multi-hour delays in outbound e-mail
    Marc> from r-help and r-devel.  Inbound mail seems to show
    Marc> up on the archive OK.

    Marc> I am just now getting posts from late yesterday my
    Marc> time.  My post to r-devel on RH9 yesterday morning did
    Marc> not show up back to me until last night (my time).

    Marc> HTH,

Indeed that helped a bit, particularly
since I'm in a two day course today and tomorrow and haven't had
time to look closely.

What happened is that the mailman software has occasionally been producing
too long delays (of e.g., half an hour, even for R-core with 17 addresses!),
for quite a few weeks now -- which it didn't several weeks ago.
The first thing I was advised to try was to revert the mailman
and sendmail settings to their defaults -- which I had done
yesterday.  Additionally, yesterday a big part of ETH had severe
network problems during noon time -- hence it wasn't too easy to
find out which timeouts where caused by what.

I have now  re-reverted the mailman settings to what they
were before yesterday, and will start tackling the "only"
halfhour delays (which I think come from the different qrunner setup
that mailman 2.1.x is using -- which is active since about March 4).

Thanks for adverting me!

BTW: If anyone has time for writing a small perl or python script
that automatically retrieves the mailman delay from the two
"Received:" headers in each message that contain "hypatia" --
I'd be quite grateful..

Martin Maechler <maechler at stat.math.ethz.ch>	http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><



More information about the R-devel mailing list