Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!newshost.telstra.net!asstdc.scgt.oz.au!metro!metro!news.cs.su.oz.au!harbinger.cc.monash.edu.au!news.mira.net.au!Germany.EU.net!Dortmund.Germany.EU.net!nntp.gmd.de!news.ruhr-uni-bochum.de!news.dfn.de!prise.nz.dlr.de!zib-berlin.de!news.uni-ulm.de!rz.uni-karlsruhe.de!blackbush.xlink.net!tank.news.pipex.net!pipex!newsfeed.internetmci.com!uunet!not-for-mail From: peter@nmti.com (Peter da Silva) Newsgroups: comp.unix.bsd.freebsd.misc,news.admin.technical,news.software.nntp Subject: Re: Poor performance with INN and FreeBSD. Date: 29 Mar 1996 00:56:33 -0500 Organization: Network/development platform support, NMTI Lines: 36 Sender: zorch@ftp.UU.NET Approved: zorch@uunet.UU.NET Distribution: world Message-ID: <4jc0n1$t42@web.nmti.com> References: <311F8C62.4BC4@pluto.njcc.com> <4ipb5c$24om@usenetw1.news.prodigy.com> <4j6l4a$9ia@web.nmti.com> <4j9rce$210g@usenetw1.news.prodigy.com> NNTP-Posting-Host: ftp.uu.net Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:16268 news.admin.technical:1380 news.software.nntp:21160 In article <4j9rce$210g@usenetw1.news.prodigy.com>, bill davidsen <davidsen@tmr.com> wrote: > In article <4j6l4a$9ia@web.nmti.com>, Peter da Silva <peter@nmti.com> wrote: > | In article <4ipb5c$24om@usenetw1.news.prodigy.com>, > | bill davidsen <davidsen@tmr.com> wrote: > | > Having run on a log filesystem (JFS on AIX), I can tell you that the > | > log file is the bottleneck. To get performance up, first I put the > | > logfile on a dedicated disk (cut expire time by 3), and then on a > | > dedicated solid state disk (cut expire time by 3 again). > | This paragraph doesn't make sense to me. Maybe I'm missing something, > | but isn't the distinguishing feature of a log file system that all data > | is *only* stored in the log? I'm not familiar with JFS, but the acronym > | sounds like "Journalling File System"... which is to say a conventional > | file system with a small log containing recent updates only and primarily > | used for crash recovery rather than performance. > Changes to the directory structure are written to the log logical > volume (type jfslog), which is a circular queue. It's a separate > logical volume, and can be on any device. Once a directory change > has been written to the actual directory file the trailing pointer > of the queue is moved. Then it's not a log file system, it's a journalled file system using a log file for the journal. You can't generalize from the behaviour of JFS to the behaviour of the log file systems we're talking about. There is no separate write to the log... the write to the log is the *only* action performed, and it's performed in line with all the other writes (that is, the head just sits there at the end of the log going click...click...click track after track after track). -- Peter da Silva (NIC: PJD2) `-_-' 1601 Industrial Boulevard Bailey Network Management 'U` Sugar Land, TX 77487-5013 +1 713 274 5180 "Har du kramat din varg idag?" USA Bailey pays for my technical expertise. My opinions probably scare them