[Wine] Mailing list going beserk (Done!?!) [Search newsgroup, NNTP]

[Jeremy White]

John Drescher wrote:
>> There may be a few stragglers yet, but the bulk of the queued
>> mail is now gone, so hopefully we're finally done.  My apologies
>> for not understanding where this mail was queued sooner; I could
>> have stopped this last night if I had understood it.
>>
> The flow appears to be done for me as well. Thanks for fixing that.

Good to hear.  I'm writing this email so that it can go in the
mailing list archive, so that if something like this happens in
the future, I can find this thread, and use it to more quickly address
the problem.

The thing I should have done faster, but did not, was flush
the exim outbound queue.  It turns out that these emails were mostly
queued there; in hind sight, that is obvious.  (That's tricky;
mail queues are not exactly databases with easy 'click here to
delete all the emails people don't want' buttons, but it is doable.
Main commands are mailq; exim4 -Mvh, exim4 -Mvb, and exim4 -Mm).

The main thing I still do not know is how to see what the mailman
queue looks like.  Further, I was misled by an odd apparent queuing
behavior; I'd see files appear and disappear in /var/lib/mailmain/qfiles/news,
and assume that I needed to figure out why the heck that was happening
in order to solve the problem.  (I still need to figure out exactly
why that happens, but I should have gone to the exim queue faster).

The two key scripts are /var/lib/mailman/cron/gate_news, which is
what brings email in from the newsgroup, and
/usr/lib/mailman/Mailman/Queue/NewsRunner.py, which sends the news out.
/var/log/mailman/fromusenet is errors from the former, and
/var/log/mailman/error reports the latter.

The admin interface for the list does have the essential elements, though.

Also, this apparently happened once last fall at some point, but only
300 messages or so were caught up, so folks didn't scream at me then.

Alright, that's enough for posterity.

Cheers,

Jeremy