*BSD News Article 81792


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.rmit.EDU.AU!news.unimelb.EDU.AU!munnari.OZ.AU!news.Hawaii.Edu!news.uoregon.edu!hammer.uoregon.edu!csulb.edu!news.sgi.com!news.sprintlink.net!news-peer.sprintlink.net!howland.erols.net!news-peer.gsl.net!news.gsl.net!portc01.blue.aol.com!newsstand.cit.cornell.edu!news.acsu.buffalo.edu!dsinc!newsfeed.pitt.edu!bb3.andrew.cmu.edu!cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!cmcl2!newsserv.cs.sunysb.edu!news.
cc.sunysb.edu!cfanning
From: Chris Fanning <cfanning@jingoro.prevmed.sunysb.edu>
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: FreeBSD as news-server??
Date: 23 Oct 1996 23:09:01 GMT
Organization: State University of New York at Stony Brook
Lines: 71
Message-ID: <54m8id$p70@abel.ic.sunysb.edu>
References: <537ddl$3cc@amd40.wecs.org> <53ott7$579@adv.IAEhv.nl> <53pm5c$5ks@twwells.com> <53u1ic$61i@flash.noc.best.net> <53ucuj$8qh@twwells.com> <540dos$asu@jingoro.prevmed.sunysb.edu> <54398m$img@twwells.com>
NNTP-Posting-Host: jingoro.prevmed.sunysb.edu
X-Newsreader: TIN [UNIX 1.3 unoff BETA release 961020]

Well, this teaches me to ever make a post about news.  I just spent the past
3 days cleaning up a huge mess after I posted this article.  (I typically
get 100MB of news a day, suddenly I got 600!  Egads!)

Thanks to everyone who pointed out that I could remount the disks.  I haven't
tried it yet, because part of my recovery was upgrading to 1.5b1.  I'll
probably try that tonight.

T. William Wells <bill@twwells.com> wrote:
[my stuff snipped]
: One thing you have to be aware of is that the time to store an
: article is (now) dominated by the time it takes to search and
: update a directory. And that's proportional to the size of the
: directory.

: What this means is that the time to *store* an article grows
: faster than the newsfeed size. For a rule of thumb, consider it
: proportional to the newsfeed size times your spool size.

Well, in my case, the spool size is the same size today as it was years ago.
I have cut down on my expire times and what I carry considerably to
compensate.

: : Doing roughly the same thing today on a P133 with 64MB RAM with FreeBSD I
: : get a peak of 10 art/sec.

: (What disks? What disk controller? With disk I/O being the
: bottleneck, these are critical.)

I used to run with:
Linux 1.something, 486/66, 16MB
NCR 810 (5MB/s async mode only supported):
  2GB Barracuda for spool
  2GB Barracuda for system, history, homes, etc.

I now run with:
FreeBSD-stable, P5-133, 64MB
NCR 810
  2GB Barracuda for spool
NCR 810
  4GB Quantum Grand Prix for system, history, etc.
  2GB Barracuda for homes

So, as far as I/O is concerned, there's been nothing but a huge increase in
that area.

: mount -u -o async

: I think will do it; it's in the man page anyway. But I don't know
: if it makes much difference, though I do run that way. Most of the
: time is spent in *reading* the directory -- one has to do that in
: order to determine if the file already exists. Async writes don't
: help there.

Yup, I'm going to give this a crack.  The metadata writes are primarily of
concern when you create/write to a file and I think I have enough locality
to properly cache most reads.

As an aside, after cleaning up my huge mess, ytalk stopped working for me.
I get this in syslog:

Oct 23 11:55:35 jingoro talkd[341]: sendto: Address family not supported by
protocol family

Anyone have a clue?  I did recompile the kernel, my auto-sup might have
changed a few things around which broke it.  Recompiled ytalk and ntalkd but
I get the same thing.

Now to see if I can still post.. :)  (Argh!  More things to fix!)

Chris