FW: 00-140r0 - Bug in SAM-2 Task Identifier Definition

Ericson, George gericson at clariion.com
Fri Mar 24 09:35:15 PST 2000


* From the T10 Reflector (t10 at t10.org), posted by:
* "Ericson, George" <gericson at clariion.com>
*
Gene,

Comments Inline:

George

> -----Original Message-----
> From: Gene_Milligan at notes.seagate.com
> [mailto:Gene_Milligan at notes.seagate.com]
> Sent: Friday, March 24, 2000 9:53 AM
> To: owner-t10 at t10.org
> Cc: T10, Reflector; Ericson, George
> Subject: Re: FW: 00-140r0 - Bug in SAM-2 Task Identifier Definition
> 
> 
> 
> <<"each port is a unique SCSI Device."
> The SMU (SCSI Multi-port Unit) concept was created
> in an attempt to address that concern.>>
> 
>      Not to muddy the water but the above should be ""each 
> port appears to
> be a SCSI Device."
> The SMU (SCSI Multi-port Unit) concept was created in an 
> attempt to address
> that concern."
> 
>      My distinction is that, at least for parallel, until all 
> this gets
> sorted out there is no such thing as a port - just the SCSI 
> device. I also
> take exception to the "unique" since if in deed if the 
> toaster has two of
> those no such things there are two SCSI devices sharing a 
> common device
> server.
-------------------------------------------------------------
GME: I disagree. Don't forget about all those RAID Subsystems out there that
have multiple SCSI Parallel Interface Service Delivery Ports.  These devices
are conformant to SAM and earlier versions of SAM-2.  SAM-2 must not be
modified to make them non-conformant.  More detail on these devices follows.


These RAID Subsystems are partitioned into one or more targets, each having
a disjoint set of one or more Service Delivery Ports and a common set of LUs
(device servers.)  So by SAM definitions, a RAID subsystem logically
implements one or more SCSI Devices (that are each SCSI Targets).

As far as I know, all of the vendors that produce these RAID Subsystems
conform to SAM and SPC in that an LU Reset generates a Unit Attention
condition that will be noticed by all initiators on their next access to the
LU, regardless of Service Delivery Port.  Also, a Target Reset delivered
through any Service Delivery Port of a target will have the same affect as a
Target Reset delivered through some other Service Delivery Port of the same
target. 
-------------------------------------------------------------
> 
> <<As of this writing, the SMU concept is falling into
> some disfavor. >>
> 
>      I still favor it since the SMU and a vanilla SCSI Device 
> have very
> diferent ramifications. I am not necessarily opposed to 
> bringing some of
> the additional refinements into the SMU description but I do 
> not want to
> see all that baggage lumped under the monichor SCSI Device.
-------------------------------------------------------------
GME: I suggest a different approach.  Start with SAM and figure out what has
to be removed to create a definition of a SCSI Device with only one Service
Delivery Port.  You will find that "all that baggage" is not such a heavy
load.
-------------------------------------------------------------

Regards,
George Ericson 

CLARiiON Advanced Technology 
EMC Corporation, 4 Coslin Drive, MS C44, Southboro, MA 01772
Office: (508) 480-7349; Mobile: (508) 498-8461; Fax: (508) 480-7913
*
* 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