RE: LU access through an iSCSI session

[Roger Cummings <roger.cummings@veritas.com>]

Charles,

Jim Hafner said the same thing when the question came up with regards to his
Access Control proposal in T10, and a bunch of people (including myself)
went and looked at SAM as a result, and I believe that there was a good
consensus in T10 that this is correct.

Regards,





Roger

> -----Original Message-----
> From: Charles Monia [mailto:cmonia@NishanSystems.com]
> Sent: Saturday, September 23, 2000 7:59 PM
> To: meth@il.ibm.com; Robert Snively; ips@ece.cmu.edu
> Subject: RE: LU access through an iSCSI session
> 
> 
> Hi:
> 
> Strangely enough, while we're on the subject, I don't believe SAM-xx
> requires a target to present the same view of its logical 
> units to every
> initiator regardless of the number of NICs.  I suspect there 
> may even be
> products out there that exploit this "feature".
> 
> Charles
> 
> 
> > -----Original Message-----
> > From: meth@il.ibm.com [mailto:meth@il.ibm.com]
> > Sent: Saturday, September 23, 2000 1:19 PM
> > To: Robert Snively; ips@ece.cmu.edu
> > Subject: RE: LU access through an iSCSI session
> > 
> > 
> > 
> > 
> > 
> > There should be no problem. It is similar to the case where 
> > there are 2
> > different SCSI busses connecting a host to a controller. Each 
> > iSCSI session
> > is like a separate virtual SCSI bus. The LU names for the 
> > same device that
> > is reached via the different paths may be different. The way 
> > for the host
> > to tell that they are actually the same device is though the 
> > Vital Product
> > Data, which uniquely identifies the LU.
> > 
> > - Kalman
> > 
> > Robert Snively <rsnively@Brocade.COM> on 21/09/2000 18:42:44
> > 
> > Please respond to Robert Snively <rsnively@Brocade.COM>
> > 
> > To:   Kalman Meth/Haifa/IBM@IBMIL, Pierre Labat 
> <pierre_labat@hp.com>,
> >       ips@ece.cmu.edu
> > cc:
> > Subject:  RE: LU access through an iSCSI session
> > 
> > 
> > 
> > 
> > 
> > >  LU numbers are identical across all of the TCP connections of a
> > >  single iSCSI session.
> > 
> > This is an intriguing concept.  At present, SCSI does not require
> > logical units to have the same LUN if they are accessed through
> > different NICs.  However, present SCSI is essentially one NIC 
> > per session.
> > 
> > If you have an iSCSI session using NIC A and NIC B and a 
> > second session
> > using NIC B and NIC C as physical paths, by this definition,
> > NIC B could have two different LUNs for the same logical unit, one
> > from each session.
> > 
> > That should not be a problem, right?
> > 
> > 
> > 
> > 
>