[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[leafnode-list] Starvation from large-binary groups


I checked the mailing list archive and didn't see anything answering this
question.  (Someone did post that the maxfetch= parameter should not be
used, because articles over the limit are discarded.)

We deployed LeafNode a short time ago, and as the queue is getting
populated I've noticed a pattern.  Someone will be the first person to
subscribe to a newsgroup, usually binaries or some other high-bandwidth
group.  The next time fetch runs, it tries to pull down all the messages
from that group, which can take a long, long time.  Meanwhile, people who
read the low-traffic groups won't see any updates, because fetch is busy
pulling down all those fat binaries.

I thought maxfetch= would be a perfect solution to this; we could set it
low and run fetch more often; each newsgroup would get a guaranteed number
of articles per pass.  But I can't use it if articles over the limit are
just discarded!  And the lower I set maxbytes=, the less useful the
binaries groups will be.

Does anyone else have a solution to this problem?


Brent J. Nordquist <brent-nordquist@xxxxxxxxxx>
W: +1 651 638-6136

leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list