FCP-2 recovery problem

Baldwin, Dave Dave.Baldwin at emulex.com
Mon Jun 26 12:17:35 PDT 2000


* From the T10 Reflector (t10 at t10.org), posted by:
* "Baldwin, Dave" <Dave.Baldwin at emulex.com>
*
Carl,

There is no need to reinvent the wheel here. My proposal plugs this hole
completely and simply. It works with multi-LUN devices. It will even
work with Class 2! Changing existing structures to add 8 bytes of LUN is
not something I would look forward to. This would break many existing
designs that are out in the field. It would also put extra resource
requirements on initiators that will cause them to reduce existing
resources to make room for the new information.

Do you feel my proposal does not work for some reason?

Best regards,
Dave Baldwin

"Binford, Charles" wrote:

>
>
> I don't think need to address anything beyond the specific whole Dave
> Baldwin described (which his proposal handles).  The Multi-Lun issue
> came up to explain why other solutions (e.g. mine!) don't work.  I am
> unaware of any other cases that are broken, even with multi-Lun
> considerations.
>
> Charles Binford
> LSI Logic Storage Systems
> (316) 636-8566
>
> -----Original Message-----
> From: Zeitler, Carl [mailto:Carl.Zeitler at compaq.com]
> Sent: Sunday, June 25, 2000 12:53 PM
> To: 'David A. Peterson'; Binford, Charles
> Cc: T10 at t10.org; 'FC Reflector'
> Subject: RE: FCP-2 recovery problem
>
> All my thinking the past has been implicitly based on an I-T nexus and
> NEVER
> on an I-T-L nexus.  This adds a whole new dimension to the equation.
> If we
> are going to go back and add LUNs to qualify an Exchange, I believe we
> have
> some major, major re-thinking to do.  This would have to be a global
> change
> to Fibre Channel, at the FC-FS level.  Class 2 doesn't work either for
>
> I-T-L.  RX-IDs are just as ambiguous as OX_IDs; they are I-T based,
> not
> I-T-L based.  To be completely pure and pristine, which I believe we
> have to
> be at this level, we would have to add 8 bytes to the FC Header.
> Maybe we
> need to follow the FC-VI track of putting the LUN in a Device Header
> for
> FCP-2, if this is what needs to be done to solve this problem!!
>
> Regards,
>
> Carl

*
* 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