From owner-freebsd-scsi Wed Apr 19 8:11:38 2000 Delivered-To: freebsd-scsi@freebsd.org Received: from shimi.swcp.com (shimi.swcp.com [198.59.115.14]) by hub.freebsd.org (Postfix) with ESMTP id BFFD737BCBE for ; Wed, 19 Apr 2000 08:11:35 -0700 (PDT) (envelope-from cheeks@swcp.com) Received: (from cheeks@localhost) by shimi.swcp.com (8.8.8/8.8.7) id JAA18259; Wed, 19 Apr 2000 09:11:01 -0600 (MDT) Date: Wed, 19 Apr 2000 09:11:01 -0600 (MDT) Message-Id: <200004191511.JAA18259@shimi.swcp.com> From: Mark Costlow To: grog@lemis.com Cc: freebsd-scsi@FreeBSD.ORG In-reply-to: <20000419103225.B92373@freebie.lemis.com> (message from Greg Lehey on Wed, 19 Apr 2000 10:32:26 +0930) Subject: Re: vinum write performance Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Greg> I don't know where to start with this one, but how about trying rawio Greg> (/usr/ports/benchmarks/rawio) and telling me the results of that. Here's the output of rawio: # rawio -v 1 -r -w -R -W /dev/vinum/rvtest Test ID K/sec /sec %User %Sys %Total RR anon 1789.4 111 0.1 0.5 0.5 16384 SR anon 14159.2 864 0.2 3.4 3.6 16384 RW anon 1470.7 91 0.1 0.4 0.4 16384 SW anon 1939.9 118 0.0 0.5 0.5 16384 # rawio -v 1 -r -w -R -W /dev/rda1s1e Test ID K/sec /sec %User %Sys %Total RR anon 2685.5 167 0.1 0.5 0.6 16384 SR anon 24518.1 1496 0.2 4.1 4.4 16384 RW anon 1942.9 121 0.0 0.4 0.4 16384 SW anon 2565.7 157 0.0 0.5 0.5 16384 That's the same disk, using the two different device nodes. I'm suprised at the pronounced difference in SR, and the smaller-than-I-expected difference in SW. Greg> Also, for the sake of reference, the info I ask for at Greg> http://www.lemis.com/vinum/how-to-debug.html. Here you go (sorry, I should have done that the first time): What problems are you having? Slow sequential writes on a vinum volume, compared to the same disk without vinum. Which version of FreeBSD are you running? FreeBSD 3.4 Have you made any changes to the system sources, including Vinum? Rebuilt kernel to include soft-updates. Problem is not sensitive to soft updates being enabled or not enabled. Supply the output of the vinum list command. Configuration summary Drives: 1 (4 configured) Volumes: 1 (4 configured) Plexes: 1 (8 configured) Subdisks: 1 (16 configured) D d0 State: up Device /dev/da1s1e Avail: 0/17501 MB (0%) V vtest State: up Plexes: 1 Size: 17 GB P vtest.p0 C State: up Subdisks: 1 Size: 17 GB S vtest.p0.s0 State: up PO: 0 B Size: 17 GB Supply an extract of the Vinum history file. 19 Apr 2000 00:59:53.484281 *** vinum started *** 19 Apr 2000 00:59:53.484839 create -v /usr/tmp/vtest.conf # Sample vinum configuration file # # Our drives drive d0 device /dev/da1s1e volume vtest plex org concat sd length 0 drive d0 19 Apr 2000 00:59:53.529276 *** Created devices *** 19 Apr 2000 01:00:00.834817 *** vinum started *** 19 Apr 2000 01:00:00.835383 list Supply an extract of the file /var/log/messages. Apr 19 00:59:53 proaccess /kernel: vinum: drive d0 is up Apr 19 00:59:53 proaccess /kernel: vinum: vtest.p0.s0 is up Apr 19 00:59:53 proaccess /kernel: vinum: vtest.p0 is up Apr 19 00:59:53 proaccess /kernel: vinum: vtest is up Apr 19 00:59:53 proaccess /kernel: vinum: vtest.p0 is up Thanks for your help, Mark -- Mark Costlow | Southwest Cyberport | Fax: +1-505-232-7975 cheeks@swcp.com | Web: www.swcp.com | Voice: +1-505-232-7992 "Education is never a waste" - Viscount du Valmont To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message