Question on Scope: 8-byte vs 2-byte LUNs

George M. Ericson ericson at
Sun Apr 12 21:53:11 PDT 1998

* From the T10 (formerly SCSI) Reflector (t10 at, posted by:
* "George M. Ericson" <ericson at>

I've been watching SCC fairly closely for only the last 18 months or so.
Don't recall discussion of pass-through [bypass] commands in email or
meeting minutes.  If I'm rehashing an old subject, I apologize.

You wrote:
	I would have to support continuance of the status quo,
	wherein array devices can reject or pass through commands
	as they see fit and report whatever parameter data on the
	REPORT LUNS command that they consider appropriate.

Under this set of rules, an application client can not rely on an array
device to guarentee integrity of data within a Volume Set.  Therefore, the
application client must implement all data integrity constraints within its
scope.  I haven't figured out how an application client can reliably
implement these constraints in a multi-host, multi-initiator environment.
For example, I didn't find any words in SCC that prescribe how Reservation
would be applied across two LUs, one contained within the other.

I'll support the status quo if SCC can provide a profile which describes how
application clients can reliably insure that data integrity of Volume Sets
is not violated in a multi-host, multi-initiator environment.

In my last mail, I made the assumption that the standard as written wouldn't
allow such a profile to be written, so I suggested that we "could" add words
to specifically make aggregated LUs unaccessible. Perhaps I spoke too soon.
If not, to explicitly complete the thought, we alternatively could add words
that specifically say that access will continue to be allowed after
aggregation.  Or alternatively, we could add an option to SCC commands that
would select which of these two policies to use.  I support the status quo
only if we can create the profile above without implementing one of the
above options that specify, or allows to be specified, the accesibility
policy to use.

Looking forward to your thoughts.

George Ericson

* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at

More information about the T10 mailing list