[9353] in linux-scsi channel archive
Re: shared SCSI buses
daemon@ATHENA.MIT.EDU (Martin Peschke)
Fri Aug 11 08:28:39 2000
Date: Fri, 11 Aug 2000 14:02:41 +0200 (MET DST)
From: Martin Peschke <peschke@fh-brandenburg.de>
To: David Teigland <teigland@sistina.com>
cc: linux-scsi@vger.rutgers.edu, gfs-devel@sistina.com
In-Reply-To: <20000810193350.D15057@sistina.com>
Message-ID: <Pine.GSO.4.10.10008111215290.8523-100000@zeus.fh-brandenburg.de>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Thu, 10 Aug 2000, David Teigland wrote:
>
> I'm interested in combining two or more active hosts with multiple devices on a
> single parallel SCSI bus. I've successfully done this, but don't know the
> extent of problems which could arise when hosts or disks are added or removed
> (crashed) on the in-use bus.
>
> A) How likely is it that the scsi driver(s) will see errors when nodes and
> drives come and go and are there specific cases which are bad?
>
> B) What are the possibilities of a node surviving if it sees scsi errors?
>
> C) How much work would it take to make all these odd cases reliable?
>
> I'm interested in the status on both 2.2 and 2.4.
Reactions on error conditions depend on whether the old or new error
handling code is used. Thus it depends on the adapter and driver of your
choice. 2.4 provides both old and new eh code. I do not know whether 2.2
provides the new eh stuff.
The new default eh strategy works like this:
Probably it resets everything if a device failes (crashed). This is
observed by other initiators on the bus. But they do not expect this reset
and start resetting, as well. (Known as reset wars) There is a workaround
that allows to avoid endless reset (scsi_report_bus_reset). I do not know
whether your driver makes use of it and whether it works at all. There is
no driver in 2.4-test2 that uses scsi_report_bus_reset.
HBA drivers are allowed to implement their own eh strategy. Maybe your
driver has its own eh startegy which is aware of multiple initiator
issues. (You can partially derive the answer for C) from this statement.)
It should be safe to remove i.e. SCSI disk by means of "scsi
remove-single-device" within /proc/scsi/scsi for every host after umount
for a particular host. Use add-single-device to add a SCSI device on the
fly.
Regards
Martin Peschke
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
