Re: iSCSI: scope of keys

["John Hufferd" <hufferd@us.ibm.com>]

I do not think that this has anything to do with sending stale values.
Whenever it is sent, it will be the value at that time.  Regardless of how
set.  No need to have async updates.   That is the way normal reporting
values work.

.
.
.
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


"Mallikarjun C." <cbm@rose.hp.com>@ece.cmu.edu on 12/14/2001 04:44:45 PM

Sent by:  owner-ips@ece.cmu.edu


To:   "ips" <ips@ece.cmu.edu>
cc:
Subject:  Re: iSCSI: scope of keys



Julian,

I actually suggest something along the following lines at the beginning of
Appendix D, than labelling each key.  This would also allow us to not
specifically
call out each "IO" key as we currently do.

  All keys with LO or  FFPO, or qualified as Declarative use have
session-wide
  applicability.  All keys with IO or ALL usage have connection-wide
applicability.

With that said, some general comments on keys -

- InitiatorAlias, TargetAlias and TargetAddress keys should be qualified
with
   the "Declarative" label.

- I recommend changing InitialR2T, BidiInitialR2T and ImmediateData as IO,
   fom the current ALL (with the one-time change restrictions).   This
makes
it
   simple, and besides I can not quite picture a scenario the current
allowance
   could be useful in.

- InitiatorAlias and TargetAlias are keys that merely report an alias set
in a
  different management domain (hence declarative), with the currently
specified
  usage as ALL.  I assume it is so because the alias could be changed while
the
  session is in progress outside of iSCSI.  iSCSI does not however require
this
  key to be sent everytime the alias is changed during a session.  We
should
either
  require this to avoid stale values, or much better, drop aliases
altogether
from
  iSCSI (can be asynchronously set in a different domain, and not useful
for
iSCSI
  protocol interactions)....

- Editorial: MaxRecvPDULength - s/b "Use: All" with "Use: ALL".
--
Mallikarjun

Mallikarjun Chadalapaka
Networked Storage Architecture
Network Storage Solutions Organization
Hewlett-Packard MS 5668
Roseville CA 95747

----- Original Message -----
From: "Julian Satran" <Julian_Satran@il.ibm.com>
To: <ips@ece.cmu.edu>
Sent: Friday, December 14, 2001 12:50 PM
Subject: Re: iSCSI: scope of keys


Eddy,

I think the text says it - but if people love headers better I can add it
(some voices needed).

Julo




"Eddy Quicksall" <Eddy_Quicksall@ivivity.com>
Sent by: owner-ips@ece.cmu.edu
14-12-01 22:12


        To:     "ips@ece. cmu. edu \(E-mail\)" <ips@ece.cmu.edu>
        cc:
        Subject:        iSCSI: scope of keys




Would it make sense to add a ?scope? to each key definition? The IO and LO
?use:? labels almost do that but not in all cases. For example:

 SW = Session wide
 CO = Connection only

 Use: IO
 Who can send: Initiator
 Scope: SW

 Key=<value>

Eddy