*BSD News Article 42423


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!spool.mu.edu!howland.reston.ans.net!Germany.EU.net!zib-berlin.de!news.cs.uni-magdeburg.de!RRZ.Uni-Koeln.DE!usenet
From: se@parc03.MI.Uni-Koeln.DE (Stefan Esser)
Newsgroups: comp.os.386bsd.misc
Subject: Re: Speccing a FreeBSD box
Date: 13 Feb 1995 01:33:50 GMT
Organization: Institute for Mathematics, University of Cologne, Germany
Lines: 40
Distribution: world
Message-ID: <3hmctu$h3o@news.rrz.uni-koeln.de>
References: <D3q71E.64v@bonkers.taronga.com> <3hh7bl$9qs@idiom.com> <3hhpqd$ae8@deep.rsoft.bc.ca> <3hm10n$76d@idiom.com>
NNTP-Posting-Host: parc03.mi.uni-koeln.de

In article <3hm10n$76d@idiom.com>, muir@idiom.com (David Muir Sharnoff) writes:
|> I'm sure 1.1 will correct all, but NetBSD 1.0's driver for the
|> 53c810 is both slow and buggy.   It'll work okay most of the time,
|> but under reasonable load I was getting almost a panic per day.

Could you please send details on the problems you had.
(Just in case you still have some logfiles or other 
suitable information.)

We can't fix a problem in the driver, if we never hear
about it. If we lear about it, it's generally fixed within 
a few days ... (at least if we have any chance to reproduce 
the fault).

The NCR driver seems to work very well for a number of
people who sent success reports (and benchmark results:
Thanks a lot !). (Available on request ...)
Those benchmark results indicate, that the driver in fact 
isn't "slow". (FreeBSD-2.0 had some file system optimisations 
that were missing from NetBSD-1.0, but both -current systems 
should be the same.)

It is used on a number of highly loaded systems without
the kind of problems you talk about.

|> Don't be afriad to make recomendataions in the release notes.  It would
|> have saved me a lot of time had I known that the 53c810 doesn't work
|> well with NetBSD-1.0 or that the 3c509 doesn't work well with FreeBSD-2.0.

Hmm, I know the NCR driver works well for a lot of people.
And if it didn't work for you, then we really would have 
appreciated a bug report!

Regards, 
-- 
 Stefan Esser				Internet:	<se@ZPR.Uni-Koeln.DE>
 Zentrum fuer Paralleles Rechnen	Tel:		+49 221 4706019
 Universitaet zu Koeln			FAX:		+49 221 4705160
 Weyertal 80
 50931 Koeln