*BSD News Article 81017


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.mira.net.au!news.vbc.net!garlic.com!news.he.net!news1.good.net!news.good.net!news.magicnet.net!iag.net!newspump.sol.net!howland.erols.net!EU.net!newsfeed.internetmci.com!in1.uu.net!gehenna.pell.chi.il.us!pell.pell.chi.il.us!there.is.no.cabal
From: orc@pell.chi.il.us (david parsons)
Newsgroups: comp.os.linux.development.system,comp.unix.bsd.freebsd.misc
Subject: Re: Linux vs FreeBSD disk io performance
Date: 16 Oct 1996 19:40:18 -0700
Organization: Every flame is sacred, every flame is great
Lines: 17
Message-ID: <5446ai$20e@pell.pell.chi.il.us>
References: <86g23ey1uc.fsf@romulus.ihosteng.priv.no>
NNTP-Posting-Host: pell.pell.chi.il.us
Xref: euryale.cc.adfa.oz.au comp.os.linux.development.system:33690 comp.unix.bsd.freebsd.misc:29495

In article <86g23ey1uc.fsf@romulus.ihosteng.priv.no>,
Ivar Hosteng  <ivarh@romulus.ihosteng.priv.no> wrote:
> Other tests I have done seems to comfirm
>that FreeBSD is handling heavy disk IO with more grace than linux
>(2.0.21) does. 

It may be a side-effect of sync vs async disk i/o; when you're doing an
expire, you're doing a whole bunch of metadata updates which will get
committed immediately on freeBSD but which will pool up on a Linux
machine and only commit when update kicks in or you do something that
drives data out to disk.

How many articles do you expire every night?

                ____
  david parsons \bi/ orc@pell.chi.il.us
                 \/