Re: [Ips] iSCSI-specific unit attention conditions

Paul Koning <> Wed, 01 April 2009 15:50 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id EA9523A67F7 for <>; Wed, 1 Apr 2009 08:50:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -104.024
X-Spam-Status: No, score=-104.024 tagged_above=-999 required=5 tests=[AWL=-0.121, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, RCVD_IN_NJABL_RELAY=2.696, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id aLkJcV7benOP for <>; Wed, 1 Apr 2009 08:50:16 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id AE03D3A63EB for <>; Wed, 1 Apr 2009 08:50:16 -0700 (PDT)
X-IronPort-AV: E=Sophos;i="4.39,307,1235973600"; d="scan'208";a="393035242"
Received: from unknown (HELO ([]) by with SMTP; 01 Apr 2009 10:51:16 -0500
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <>
Date: Wed, 1 Apr 2009 11:51:12 -0400
From: Paul Koning <>
References: <> <>
X-Mailer: VM 7.17 under 21.4 (patch 22) "Instant Classic" XEmacs Lucid
X-OriginalArrivalTime: 01 Apr 2009 15:51:15.0308 (UTC) FILETIME=[B05D9AC0:01C9B2E1]
Subject: Re: [Ips] iSCSI-specific unit attention conditions
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: IP Storage <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Wed, 01 Apr 2009 15:50:18 -0000

>>>>> "dcuddihy" == dcuddihy  <> writes:

 dcuddihy> It seems to me that the more important question is how
 dcuddihy> useful these unit attention codes are.  (For example,
 dcuddihy> ATTO's Xtend San initiator doesn't make use of them.)  If
 dcuddihy> initiators don't care about this information, precisely
 dcuddihy> defining these unit attention codes (instead of depricating
 dcuddihy> them) will be a change for the worse.

That's one of my concerns.

It seems we're just speculating what purpose these codes were intended
to serve.  Not only don't we know for sure what that purpose was, we
also don't know if that purpose is actually achieved.

The other concern is that these codes could be interpreted to impose a
new requirement on targets to generate them in certain situations.  Of
course we don't know what those situations are, or why targets should
do this, but clearly someone could argue that those numbers exist and
therefore are supposed to be generated.

Unless there is a solid proposal that assigns a clear meaning, and
that meaning is valuable to initiators, I believe that the only
correct answer is to consider what happened as a glitch in the
standards process and remove, to the extent possible, the debris left
behind by that glitch.

I don't see anything in the recent discussion that gets us to this
clear meaning and useful purpose.  In particular, I see absolutely NO
trace of "rough consensus and running code" to support the notion that
the iSCSI standard should support these new codes.

David, can we put in motion the deprecation of these codes?