*BSD News Article 58480


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!spool.mu.edu!howland.reston.ans.net!newsfeed.internetmci.com!netnews.nwnet.net!nwnet.net!not-for-mail
From: aad@nwnet.net (Anthony D'Atri)
Newsgroups: comp.unix.bsd.bsdi.misc,comp.periphs.scsi,comp.arch.storage
Subject: Re: SCSI disk geometries:  weird probe defaults are faster than custom!
Date: 8 Jan 1996 11:47:36 -0800
Organization: NorthWestNet, Bellevue, WA
Lines: 42
Message-ID: <4crsco$4o2@olympus.nwnet.net>
References: <4cf25e$m82@olympus.nwnet.net> <4cjtb8$m8j@gap.cco.caltech.edu>
Reply-To: aad@nwnet.net
NNTP-Posting-Host: olympus.nwnet.net
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.bsdi.misc:1880 comp.periphs.scsi:44068 comp.arch.storage:7637

>You are talking about SCSI disks, right?

Yep.

>Then the concept of # of
>tracks, sectors/track, etc. are essentially meaningless. The SCSI
>addressing scheme is a linear collection of blocks numbered 0 to N-1.
>What's more, virtually all modern are ZCAV

....

Yes, I'm well acquainted with the fact that any geometry isn't going to
accurately describe the physical disk.  What I'm asking here is for advice
on which sort of fake geometry is best.

>What's probably going wrong is that either you reformatted the disk
>with an interleave

I haven't tried formatting a SCSI disk in years.  Most of the utilities I
tried to use for the purpose failed miserably.  It's interesting to not that
BSDI's man page for scsiformat says that it only works on the HP300 and SPARC
ports -- one of many "What the @#$!"'s in this OS.

>or more likely that in doing the newfs you
>included a rotdelay or set maxcontig to be something like 1

Defaults -- dumpfs shows maxcontig to be 8, rotdelay to be 4ms.

>little or no caching/readahead, etc.

I've never seen anything that described how to make sure that such features
are enabled on a disk, but have seen a few people claiming that they're
often disabled by default.

>Bottom line: carefully check the man pages for newfs, mkfs, and even
>mount to see that you haven't instructed it to limit consecutively
>placed blocks.

Defaults all the way, except -b 4096 -f 512 on the two news spools, which
restricted c/g to 8 for no apparent reason.