FC Mode Page DSA/RHA Bit

Bob Snively Bob.Snively at EBay.Sun.COM
Tue Oct 19 15:21:49 PDT 1999


* From the T10 Reflector (t10 at t10.org), posted by:
* Bob Snively <Bob.Snively at EBay.Sun.COM>
*

Jim raised an interesting question.

First, the hard address choice is only relevant in loop environments where
assignment is made by the loop initialization process.  We have not put
any wording in to handle any other cases.  So it certainly only applies
to loops, whether public or private.

Second, only the low order byte of the D_ID is extracted based on the
AL_PA.  All other bytes are fabric assigned.  As a result, I would expect
that you might be able to justify the choice of the low-order byte as
a hardware selected value, but the remainder of the address space is
not known and may be reassigned by the fabric in any manner it desires,
independently of the state of the RHA bit.

Now the question to all of you.

	Do you want to fix the low-order byte of the D_ID of a loop
	target to the hardware location using the RHA bit?
	Note that there is no necessary capability implied of parsing the
	actual physical location of the device.
	
No response or a yes will be taken as a "yes" and the RHA will be applied to
the private loop address and the low-order byte of the public loop address
in revision 4 of FCP-2.

A negative response will put this item on the T10 agenda in Monterey
for more complete discussion.

Bob

>* From the T10 Reflector (t10 at t10.org), posted by:
>* Jim_Coomes at notes.seagate.com
>*
>The Joint T10/T11 Activity Working Group Minutes from the meeting
>last week indicates there was discussion of the FC Control mode
>page RHA (alias DSA) bit. The minutes indicate the discussion
>included disallowing the RHA bit for fabric attachment.
>
>The requirement for a hard address is still valid for fabric
>environments. Switches are increasingly being incorporated into
>storage systems. Those private loop JBOD products are migrating to
>public behavior. The association of an address to a physical
>location is still a desired feature.
>
>Fabrics already have to handle address changes due to
>configuration changes and power cycling. A change of address due
>to a hard address change should be an infrequent event due to a
>configuration error.
>
>Am I missing a problem with RHA support in public configurations?
>
>Jim

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






More information about the T10 mailing list