Return to BSD News archive
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!caen!usenet.cis.ufl.edu!usenet.eel.ufl.edu!MathWorks.Com!solaris.cc.vt.edu!uunet!zib-berlin.de!rs1-hrz.uni-duisburg.de!rrz.uni-koeln.de!RRZ.Uni-Koeln.DE!RRZ.Uni-Koeln.DE!news
From: se@FileServ1.MI.Uni-Koeln.DE (Stefan Esser)
Newsgroups: comp.os.386bsd.questions
Subject: Re: identical scsi drives, different performace
Date: 7 Oct 1994 22:01:44 GMT
Organization: Institute of Nuclear Physics, University of Cologne, Germany
Lines: 69
Distribution: world
Message-ID: <374gg8INN23k1@rs1.rrz.Uni-Koeln.DE>
References: <rcarterCx9unI.AwJ@netcom.com>
NNTP-Posting-Host: fileserv1.mi.uni-koeln.de
In article <rcarterCx9unI.AwJ@netcom.com>, rcarter@netcom.com (Russell Carter) writes:
|> I have two Seagate ST11200N drives, on an ASUS SP3G motherboard using
|> Wolf and Steffen's v2.1 pci ncr driver. The trouble is that the performance
|> differs substantially:
|>
|> Filesystem 512-blocks Used Avail Capacity Mounted on
|> /dev/sd0a 29184 16786 9478 64% /
|> proc 180 41 139 23% /proc
|> /dev/sd0e 1301644 700098 471380 60% /usr
|> /dev/sd0h 615568 445088 170480 72% /dos
|> /dev/sd1e 1899078 752348 956822 44% /mnt/sysb
|>
|> Output of bonnie on sd0e:
|>
|> -------Sequential Output-------- ---Sequential Input-- --Random--
|> -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
|> Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
|> 100 985 99.0 2061 18.0 557 9.0 1364 92.5 2379 25.5 52.7 5.0
|>
|> Output of bonnie on sd1e:
|>
|> -------Sequential Output-------- ---Sequential Input-- --Random--
|> -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
|> Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
|> 100 1044 99.1 1170 10.3 483 8.3 536 34.2 536 5.7 56.9 5.4
Looking at the write performance, I'd suggest
you unmount the sd1e partition and issue the
following command:
tunefs -a8 -d0
If any of the file system parameters change,
then it had not been set up for non interleaved
operation ...
(I'm suggesting this, since the block write
speed is half that of the other drive's and
the block read speed is only one block per
revolution of the disk. Looks like only each
other block is allocated to a file ...).
|> A dumpfs on sd1e starts off with:
|>
|> magic 11954 format dynamic time Thu Oct 6 13:39:46 1994
|> nbfree 71234 ndir 1826 nifree 201687 nffree 3493
|> ncg 61 ncyl 973 size 979324 blocks 949539
|> bsize 8192 shift 13 mask 0xffffe000
|> fsize 1024 shift 10 mask 0xfffffc00
|> frag 8 shift 3 fsbtodb 1
|> cpg 16 bpg 2013 fpg 16104 ipg 3776
|> minfree 10% optim time maxcontig 1 maxbpg 2048
^^^
|> rotdelay 4ms headswitch 0us trackseek 0us rps 60
^^^
Hmmm, seems to support my idea.
If your drive has been in use for some time,
I'd suggest doing a full dump/restore after
tuning the file system parameters ...
STefan
--
Stefan Esser Internet: <se@ZPR.Uni-Koeln.DE>
Zentrum fuer Paralleles Rechnen Tel: +49 221 4706010
Universitaet zu Koeln FAX: +49 221 4705160
Weyertal 80
50931 Koeln