You probably want to use the PRIME keyword to prevent the large lists
from becoming active during the time when other traffic is more
important, like allow the large lists to run only between 9 and 10 in
the evening. Of course, this kinda cranks down the potential for
"discussion" but it sounds like the larger ones you have are mostly
one-way type traffic.

Some of what you describe sounds a lot like the message was really just
too big and was being refused, or returned as non-deliverable, by some
sites - so LISTSERV was generating PROBE traffic to detect if the
failures were transient or persistent, and the PROBE items can be even
bigger than the original message (a somewhat self-perpetuating scenario
that can continue for days and days).  Check the daily error report for
the list to see if it has dozens or hundreds of subscribers newly
listed. Or do some investigating in the log details, but that might be
more trouble than you're willing to take.  Other than preventing huge
messages from going through your lists, I don't know any way to avoid
this headache.  Setting a SIZELIM keyword value quite a bit smaller than
the size of the recent bothersome message is probably a good idea.

If your newsletter publishers then complain that they can't get their
items distributed, suggest that they multiply publication frequency and
divide the number of articles per issue (instead of monthly, every two
weeks, for example).  Or leave out all the pretty pictures, supply links
instead.

On 10/19/2011 5:36 PM, Wauford, Melissa wrote:
> I’m running Listserv 15.5 under Solaris 10 on a standalone SPARC system.
> I have a lot of generally small lists, but I have a small handful quite
> large lists that we tolerate because they are used infrequently and the
> owners are generally good citizens and avoid sending out email during
> high traffic periods. However.
>
> Last week I had one of said list owners send a VERY large message to her
> quite large list of subscribers. Which, of course, backed up listserv
> right at the worst time, but that wasn’t the thing that really killed
> me. Apparently, Listserv decided to re-queue an individual copy of the
> message for every subscriber on the list which filled up the 32G or so I
> had free on the Listserv spool partition. Then all of my incoming mail
> queued up behind 18,000 copies of that message that had to be
> individually processed by Listserv. It took a while.
>
> I can’t say I’ve ever seen this before. And there is no indication in
> the Listserv log that it is reprocessing the message, so there’s no way
> to know how often it happens on other lists that I just haven’t noticed
> because the lists are small and the messages are small. But I need to
> figure out how to keep this from happening again without hanging the
> list owner up by her thumbs or buying HPO for the 3-4 lists I have that
> would benefit from it.
>
> I’m guessing that it may have something to do with the Auto-probe
> setting on the list, but I’m not sure. Can anyone shed any light?
>
> *Melissa Wauford
> *Systems Administrator, Office of Information Technology
> IT Systems: Communications & Collaboration
> The University of Tennessee

############################

To unsubscribe from the LSTSRV-L list:
write to: mailto:[log in to unmask]
or click the following link:
http://peach.ease.lsoft.com/scripts/wa.exe?SUBED1=LSTSRV-L&A=1