henry@zoo.toronto.edu (Henry Spencer) (02/28/91)
In article <EMV.91Feb27165108@poe.aa.ox.com> emv@ox.com (Ed Vielmetti) writes: >I would recommend strongly that the next C news patch have an >(optional) means of discarding news that is too old. If I recall the >"news need not be slow" paper, this function was in an early version >of C news, but was discarded for performance reasons. No, it's never been in C News that I recall. It is clearly necessary, and is in the works, although it may not make it into the next patch (which is imminent). Simply keeping a bit more history does noticeably reduce the problem, and may be useful as a stopgap. We haven't seen much old news here. -- "But this *is* the simplified version | Henry Spencer @ U of Toronto Zoology for the general public." -S. Harris | henry@zoo.toronto.edu utzoo!henry
emv@ox.com (Ed Vielmetti) (02/28/91)
In article <1991Feb27.193118.22487@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes:
HELP! My news partition is filling up fast. Could it be due to all of
those Jan 30th articles that have suddenly started coming in?
It looks as though cs.widener.edu is back on the net, now that their
router is working again. Unfortunately a lot of old queued articles
have gone out.
I would recommend strongly that the next C news patch have an
(optional) means of discarding news that is too old. If I recall the
"news need not be slow" paper, this function was in an early version
of C news, but was discarded for performance reasons.
--
Msen Edward Vielmetti
/|--- moderator, comp.archives
emv@msen.comkherron@ms.uky.edu (Kenneth Herron) (03/01/91)
emv@ox.com (Ed Vielmetti) writes: > HELP! My news partition is filling up fast. Could it be due to all of > those Jan 30th articles that have suddenly started coming in? >I would recommend strongly that the next C news patch have an >(optional) means of discarding news that is too old. We're running news on a 26-processor sequent. This machine is generally I/O bound; it's load average rarely gets above 0.8. I would gladly trade a bit of performance for improved old-article rejection, especially since it would be "only" CPU time. On the other hand, I'm keeping 38 days of message ID's, so I haven't seen many old articles either. -- Kenneth Herron kherron@ms.uky.edu University of Kentucky (606) 257-2975 Department of Mathematics "Never trust gimmicky gadgets" -- the Doctor
brtmac@maverick.ksu.ksu.edu (Brett McCoy) (03/01/91)
In <kherron.667759870@s.ms.uky.edu> kherron@ms.uky.edu (Kenneth Herron) writes: >emv@ox.com (Ed Vielmetti) writes: >> HELP! My news partition is filling up fast. Could it be due to all of >> those Jan 30th articles that have suddenly started coming in? >>I would recommend strongly that the next C news patch have an >>(optional) means of discarding news that is too old. >We're running news on a 26-processor sequent. This machine is generally >I/O bound; it's load average rarely gets above 0.8. I would gladly trade >a bit of performance for improved old-article rejection, especially since >it would be "only" CPU time. >On the other hand, I'm keeping 38 days of message ID's, so I haven't seen >many old articles either. I was only keeping about three weeks worth of old history lines, which I've now upped to a little over four weeks, and while that may make some difference, it won't fix the problem because a good share of these reposted articles have different message ID's than the original artical. It looks as if some site out there is modifying the message ID's and then reposting them. It is causing much distress to my server, and most of the sites that I feed. Does anyone know of a way to easily go through and expire any article that has a date older than X, not just expire articles which arrived more than X days ago? -- "I wrote a lisp program once...it wrote back to me." -- unknown Reality is for people who can't deal with drugs. Brett McCoy Computing and Telecommunications Activities brtmac@maverick.ksu.ksu.edu Kansas State University