RE: iSCSI: 2.2.6. Naming & mapping

[Peter Johansson <PJohansson@ACM.org>]

At 03:18 PM 9/14/00, DeKoning, Rod wrote:

>Since he hit one of my hot buttons regarding mapping (and therefore, 
>discovery), . . . I heartily endorse Pierre's assertion that the Device 
>Identification Page be used to identify Logical Units (Specifically the 
>Volume WWN flavor).  This is extremely important in an environment that 
>may access volumes (logical units) through many paths and potentially 
>through multiple underlying ports or physical devices.

I very much agree with both Pierre's and Rod's observation: LUs should have 
distinctive unique identifiers, else it is impractical to sort out multiple 
path access to the same device instance.

>The only question is: does this fit in a protocol spec such as iSCSI?

YES. It is entirely appropriate for a part of iSCSI to profile what it 
expects of the targets that utilize it. From the SAM, SPC, etc. 
smorgasbord, iSCSI may insist on some features (autosense, unique 
identifiers for each LU and so on), prohibit others and note the 
optionality of the remainder.




Regards,

Peter Johansson

Congruent Software, Inc.
98 Colorado Avenue
Berkeley, CA  94707

(510) 527-3926
(510) 527-3856 FAX

PJohansson@ACM.org