zoning Load step clarification

Johnson, Steve Steve.Johnson at lsi.com
Thu Mar 5 14:58:25 PST 2009


* From the T10 Reflector (t10 at t10.org), posted by:
* "Johnson, Steve" <Steve.Johnson at lsi.com>
*
Greg,
I agree the OPEN_REJECT (ZONE VIOLATION) should be returned as OPEN_REJECT
(RETRY) for the reason you have stated, however, I see the case where
OPEN_REJECT (NO DESTINATION) is possible as well and should also be returned
as OPEN_REJECT (RETRY) for the following reason:
The management device server in a zoning expander device only accepts SMP
zone configuration function requests, SMP ZONE ACTIVATE requests, and SMP
ZONE UNLOCK requests while it is locked. (See section 4.9.6.1)
Zone configuration requests allowed while locked are:
CONFIGURE ZONE MAMANGER PASSWORD
CONFIGURE ZONE PHY INFORMATION
CONFIGURE ZONE PERMISSION TABLE
CONFIGURE ZONE PHY INFORMATION can change the zone group of a phy/s this may
require route table updates throughout the topology. During these route table
changes it is possible that OPEN_REJECT (NO DESTINATION) could temporarily be
encountered. Therefore we still need to handle the NO DESTINATION case.
So something like this put into section 4.9.6.1?
"While the ZONE CONFIGURING bit is set to one, the expander device shall
return OPEN_REJECT (RETRY) for any connection requests that would otherwise
have resulted in OPEN_REJECT (NO DESTINATION) or OPEN_REJECT (ZONE
VIOLATION)."
Steve
Steve Johnson
LSI Storage Components Group
M: 719 337 3135
-----Original Message-----
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Greg Tabor
Sent: Thursday, March 05, 2009 2:48 PM
To: t10 at t10.org
Subject: Re: zoning Load step clarification
* From the T10 Reflector (t10 at t10.org), posted by:
* Greg Tabor <greg.tabor at maxim-ic.com>
*
Gerry,
No, I don't think there is ever a case where an expander would return 
OPEN_REJECT (NO DESTINATION) due to a transitory condition during 
lock/load/activate/unlock zone configuration.  Zone configuration only 
updates the ZPT, not the route table, and the ZPT only affects the 
decision between allowing a connection request to a destination versus 
returning OPEN_REJECT (ZONE VIOLATION); it has no effect on non zone 
related routing decisions such as those that would result in OPEN_REJECT 
(NO DESTINATION) because a route to a destination can not be found.
So I think that:
(1) my point is valid, and the text in 4.9.6.3 should be changed to 
"OPEN_REJECT (ZONE VIOLATION)" from "OPEN_REJECT (NO DESTINATION)".
(2) Steve's point is valid, but probably insofar as the text is 
misleading because it is located specifically in the Load step section 
yet it applies to the whole lock/load/activate/unlock process.	Perhaps 
the fourth paragraph in 4.9.6.3 (starting "While the ZONE CONFIGURING 
bit is set to one...") could be fixed as I have requested and then moved 
up to the 4.9.6.1 Zone configuration overview section for clarity.
Any takers?
-- 
Greg Tabor
Maxim Integrated Products
greg.tabor at maxim-ic.com
719-278-2063
On Wednesday 4 March 2009 1:01:32 pm you wrote:
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Gerry.Houlder at seagate.com
> *
> So does this discussion imply that the original text (referring to
> OPEN REJECT (NO DESTINATION)) is correct? I suppose this response is
> intended to cover cases where no matching destination exists in the
> old (i.e., currently active) route table. After the new values are
> loaded and activated, that destination may become valid, so a retry
> would be reasonable.
>
>
>
>
>	       "Elliott, Robert
>	       (Server Storage)"
>	       <Elliott at hp.com> 				       
>   To Sent by: 		 "t10 at t10.org" <t10 at t10.org>
> owner-t10 at t10.org					     cc No
> Phone Info
>	       Available					    
> Subject RE: zoning Load step clarification
>
>	       03/04/2009 12:31
>	       PM
>
>
>
>
>
>
>
> * From the T10 Reflector (t10 at t10.org), posted by:
> * "Elliott, Robert (Server Storage)" <Elliott at hp.com>
> *
> True.  For a group of expanders being
> locked/loaded/activated/unlocked together, it covers the time from
> the first expander being activated until the last expander is
> activated.
>
> ---
> Rob Elliott (elliott at hp.com)
> HP Industry Standard Server Storage Advanced Technology
>
>
> -----Original Message-----
> From: Johnson, Steve [mailto:Steve.Johnson at lsi.com]
> Sent: Wednesday, March 04, 2009 10:12 AM
> To: Elliott, Robert (Server Storage); t10 at t10.org
> Subject: RE: zoning Load step clarification
>
> Since the load step is not suppose to affect the expanders current
> zoning values and only the shadow values why would an expander return
> any OPEN_REJECT (ZONE VIOLATION) caused by the load step? This step
> should not have any affect on current operations, that happens during
> the activate step correct?
>
>
> From section 4.9.6.3:
> ...
> SMP zone configuration functions change the zoning expander shadow
> values and do not affect the zoning expander current values. The
> zoning expander shadow values become zoning expander current values
> during the activate step (see 4.9.6.4).
> ...
>
> Steve Johnson
> LSI Storage Components Group
> M: 719 337 3135
>
>
> -----Original Message-----
> From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of
> Elliott, Robert (Server Storage)
> Sent: Tuesday, March 03, 2009 12:35 PM
> To: t10 at t10.org
> Subject: RE: zoning Load step clarification
>
> * From the T10 Reflector (t10 at t10.org), posted by:
> * "Elliott, Robert (Server Storage)" <Elliott at hp.com>
> *
> I agree.  That was the intent of 06-384r1 and is correct in 4.6.6.3,
> 4.9.3.3, 4.9.3.5, 7.2.6.10, 7.13.4.2.5, but broken in 4.9.6.3.  I
> suggest that correction be included in the version sent for second
> public review at the March meetings.
>
> While searching for those uses, I also noted that OPEN_REJECT
> (RESERVED CONTINUE 0) and OPEN_REJECT (RESERVED CONTINUE 1) are
> mistakenly referred to as OPEN_REJECT (CONTINUE 0) and OPEN_REJECT
> (CONTINUE 1) in 10.2.7.4, 10.4.3.4, and 10.4.3.18.  I suggest that
> these be corrected as well.
>
> ---
> Rob Elliott (elliott at hp.com)
> HP Industry Standard Server Storage Advanced Technology
>
>
> -----Original Message-----
> From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Greg
> Tabor Sent: Sunday, March 01, 2009 9:45 AM
> To: t10 at t10.org
> Subject: zoning Load step clarification
>
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Greg Tabor <greg.tabor at maxim-ic.com>
> *
> In the 4.9.6.3 Load step section of SAS-2, it states:
>
> "While the ZONE CONFIGURING bit is set to one, the expander device
> shall return OPEN_REJECT (RETRY) for any connection requests that
> would otherwise have resulted in OPEN_REJECT (NO DESTINATION) (see
> 4.6.6.3)."
>
> Should that read:
>
> "...that would otherwise have resulted in OPEN_REJECT (ZONE
> VIOLATION)" ?
>
> If the existing text is correct, why would an expander return
> OPEN_REJECT (NO DESTINATION) during the zoning Load step?  During
> this time the ZPT is being updated, not the route table (right?), so
> a transient condition in the ZPT would potentially result in the
> expander returning OPEN_REJECT (ZONE VIOLATION), not OPEN_REJECT (NO
> DESTINATION).
>
> Thanks
>
> Greg Tabor
> Maxim Integrated Products
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org
*
* 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