From owner-freebsd-current Sun Mar 5 06:51:27 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id GAA07489 for current-outgoing; Sun, 5 Mar 1995 06:51:27 -0800 Received: from cardhu.cs.hut.fi (hsu@cardhu.cs.hut.fi [130.233.192.95]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id GAA07481 for ; Sun, 5 Mar 1995 06:51:25 -0800 Received: by cardhu.cs.hut.fi id AA06069 (5.65c8/HUTCS-C 1.3 for freebsd-current@freefall.cdrom.com); Sun, 5 Mar 1995 16:50:48 +0200 Date: Sun, 5 Mar 1995 16:50:48 +0200 From: Heikki Suonsivu Message-Id: <199503051450.AA06069@cardhu.cs.hut.fi> To: Peter Dufault Cc: hsu@cs.hut.fi, freebsd-current@freefall.cdrom.com Subject: Re: Wiring In-Reply-To: <199503051129.GAA01718@hda.com> References: <199503050749.JAA08984@smile.clinet.fi> <199503051129.GAA01718@hda.com> Organization: Helsinki University of Technology, Otaniemi, Finland Sender: current-owner@FreeBSD.org Precedence: bulk Peter Dufault writes: > > Would it be better to by default try to wire the first 4 disks? > I'm not sure what you mean. Where would we put the disks on my system: Something I would assume it should be: disk sd0 at scbus0 target 0 unit 0 disk sd1 at scbus0 target 1 disk sd2 at scbus0 target 2 disk sd3 at scbus0 target 3 disk sd4 at scbus0 target 4 disk sd5 at scbus0 target 5 disk sd6 at scbus0 target 6 tape st0 at scbus0 target 4 tape st1 at scbus0 target 5 device cd0 at scbus0 target 6 This really doesn't help as luns are not taken into account. SunOS uses a kludge, target x lun y becomes sd(2*x+1). Not nice, but at least it has got a fixed mapping. How about ctl, beefed up with sp for logical devices? > I don't think it is practical to do anything that changes the > default behavior present in 2.0 and earlier. With what I've put I don't know when this changed, but Mach and older NetBSD drivers had fixed mapping. I didn't notice this until I saw comments in 2.0 LINT. > had. I shudder to think what the last week would have been like > if I had changed the default behavior for disk assignments. I have been shuddering since I learned about allocating the ids dynamically. Fortunately our database test engine is a Sun, it has got 12 disks in 4 external boxes (in addition to internal boot disk). Managing such a system with sd's wandering around when one of the boxes doesn't spin up wouldn't be fun. Not even if we were lucky and one of the filesystem disks didn't happen to be remapped into database disks or tmp. -- Heikki Suonsivu, T{ysikuu 10 C 83/02210 Espoo/FINLAND, hsu@cs.hut.fi home +358-0-8031121 work -4513377 fax -4555276 riippu SN