SUMMARY: disconnecting live SCSI bus

From: Bill Morrow (morrow@cns.ucalgary.ca)
Date: Tue Mar 30 1993 - 07:06:13 CST


Sorry for the delayed summary. Three weeks ago, I asked:

>
> Is it safe to disconnect a SCSI device with the CPU and other devices powered?
>
This was in reference to having a mobile CD drive.

The final tally was: 3 said it was safe to disconnect a SCSI device at any time,
9 people, among them kfp@qualcomm.com said words to the effect:

>It is possible that you can connect and disconnect without blowing the (SCSI
>bus) fuse.
>My record is about 30-1 with no blown fuse. If you do this, however, be sure
>to L1-A to halt UNIX to the prom monitor first. Then, after the cabling is
>done, use "ok go" or "> c" to continue the CPU. It is still possible to panic
>the kernel doing this, so I would "sync" before halting, just in case.
>Obviously, devices that are "in use" should never be removed.

But most of these recommended powering down.
Another 9 respondents were more emphatic, and said you *must* power
down before fiddling with the SCSI bus.

kfp@qualcomm.com also mentioned:
"Apunix has a SCSI repeater box that is <$200 and will allow for live
connection/disconnection with no ill effects to the host and other devices."

In the future, I plan to halt the machine (at least) in emergencies, and
power down for planned hardware moves. It doesn't seem worth the risk to
play with a live SCSI bus.

Thanks to:
 Andrew Watson <ajw@ansa.co.uk>
 david@capmkt.com
 daniel@CANR.Hydro.Qc.CA (Daniel Hurtubise)
 ekurgpol@develop-law.usc.edu (Elmar Kurgpold)
 hoel@cec.mtu.edu (Jim Hoel)
 kfp@qualcomm.com
 aad@lovecraft (Anthony A. Datri)
 Peter Samuel <peters@mega.com.au>
 andrews@Erin.GOV.AU (Andrew Scarman)
 todd@flex1.Eng.McMaster.CA (Todd Pfaff)
 eckhard@ts.go.dlr.de (Eckhard Rueggeberg)
 glk@pls.amdahl.com (Greg Kastanek)
 birger@vest.sdata.no (Birger A. Wathne)
 trdlnk!mike@uunet.UU.NET (Michael Sullivan)
 caxwgk@pki-nbg.philips.de (Wolfgang Kuehnel )
 cyerkes@jpmorgan.com (Chuck Yerkes)
 rauls@usb.ve (LDC - Raul Silvera)
 David Sansom <david@camscan.co.uk>
 map@europa.ecn.uoknor.edu (Michael A. Parker)
 msd@sharpwa.com (Mark Deason)
 ups!upstage!glenn@fourx.Aus.Sun.COM (Glenn Satchell - Uniq Professional Services)
 Mike Raffety <miker@il.us.swissbank.com>
__________________________________________________________
Bill Morrow voice: (403) 220-6275
Clinical Neurosciences fax: (403) 283-8770
University of Calgary e-mail: morrow@cns.ucalgary.ca
3330 Hospital Drive NW
Calgary, Alberta, CANADA
T2N 4N1



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:39 CDT