[4993] in linux-scsi channel archive
Re: Howto: rescan scsi bus after a hot swap ??
daemon@ATHENA.MIT.EDU (Drew Eckhardt)
Sat Oct 24 18:11:42 1998
To: Michael Schwingen <michaels@stochastik.rwth-aachen.de>
cc: Marty Leisner <leisner@sdsp.mc.xerox.com>,
Chris Atenasio <chrisa@ultranet.com>,
Hardware Stuff <mrfixit@clouddancer.com>, linux-scsi@vger.rutgers.edu,
ncr53c810@Colorado.EDU
In-reply-to: Your message of "Sat, 24 Oct 1998 22:10:24 +0200."
<19981024221024.44660@nemesis.stochastik.rwth-aachen.de>
Date: Sat, 24 Oct 1998 15:37:43 -0600
From: Drew Eckhardt <drew@Poohsticks.Org>
In message <19981024221024.44660@nemesis.stochastik.rwth-aachen.de>, michaels@s
tochastik.rwth-aachen.de writes:
>On Fri, Oct 23, 1998 at 02:33:01PM -0600, Drew Eckhardt wrote:
>> The standard doesn't restrict bus interaction for powered off devices,
>> and some result in munged data when powered down (I think my old
>> DEC RZ55 did this).
>
>Hm. IIRC, the standard *does* specify that a powered-off device shall not
>disturb bus operation - so I always thought that these few (and old) devices
>which do (Segate ST296N is another one) simply are not standard-compliant.
From the last SCSI-II draft standard:
4.4.1.2. Input Characteristics
....
It is recommended that SCSI devices with power off also meet the above IIL
and IIH electrical characteristics on each signal.
recommended != required. That this needs to be said implies that other
characteristics may vary as well.
>I and some friends have been using hot-plugging quite extensively using
>those disk bays with a 50-pin centronics plug. It is definitely not
>standard-compliant, since the connector does not guarantee that you make the
>ground connectio first and break it last,
Through SCSI-II, the standard doesn't address connection order for
hot-plug. Ground, power, signal is just a good idea.
>but it works just fine (your mileage may vary).
It usually works fine. When I worked as a sysadmin in my student/slave
days, SOP was that "you can plug and unplug anything attached to an
external connector without power cycling." and as far as I can remember,
we never had any problems.
My current job involves disk based serial digital video recorders targeted
at post production houses and broadcast television for play to air. Major
networks wouldn't treat loss-of-signal from a bus glitch or the RST sledge
hammer, so each of our 20 hot swap bays has its own SCSI bus.
It depends on what sort of reliability you need or want.
--
<a href="http://www.poohsticks.org/drew/">Home Page</a>
For those who do, no explanation is necessary.
For those who don't, no explanation is possible.
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
