[9362] in linux-scsi channel archive
Re: shared SCSI buses
daemon@ATHENA.MIT.EDU (Steven Buller/Tucson/IBM)
Fri Aug 11 10:32:40 2000
To: linux-scsi@vger.rutgers.edu
From: "Steven Buller/Tucson/IBM" <sjbuller@us.ibm.com>
Date: Fri, 11 Aug 2000 08:30:33 -0600
Message-ID: <OFE689F16F.5705648F-ON07256938.004D8732@LocalDomain>
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Multi-initiator discussions come up every other month, but the answer as to
whether or not the multi-initiator configs are a good idea flip/flops.
The discussions around 14 June indicated that because the mid-layer doesn't
handle reserves and reset wars, multi-initiator configs are a bad idea on
Linux and are exposed to data corruption. Since the same mid-layer is used
for FC, there are multi-initiator issues in SANs as well.
I understood multi-initiator configs are at risk until Linux drivers (and
devices) implement the SCSI-3 persistent reserves and ACA.
This time around the discussion is indicating a different story.
Are multi-initiator SCSI/FC configs a robust, recommended topology or not?
I'm not looking for the anecdotal "works for me", I'm asking from a driver
design perspective.
Regards,
Steve
std discliamer: Views expressed are those of my prevailing personality of
the day and not those of my employer.
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
