FW: FW: persistent reservations follow-up question
Ulrich, David
dulrich at lsil.com
Fri Nov 14 11:40:15 PST 2003
* From the T10 Reflector (t10 at t10.org), posted by:
* "Ulrich, David" <dulrich at lsil.com>
*
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.
------_=_NextPart_001_01C3AAE6.E43C01E6
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Hi Charles,=20
I had a very similar question at one time and got this reply from
George. I'm sure the wording has changed since then, but this was the
original intent.
David Ulrich=20
LSI Logic Storage Systems Division (today)=20
-----Original Message-----=20
From: George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com=20
[ mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com
<mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com> ]=20
Sent: Friday, December 01, 2000 8:04 PM=20
To: Ulrich, David=20
Cc: 't10 at t10.org'=20
Subject: Re: FW: persistent reservations follow-up question=20
David,=20
Not really. You can look at in one of two ways, both end in the same=20
result.=20
-The target does the preempt by clearing the reservation and
registration=20
then establishing a new reservation and new registration before
completing=20
the operation. The new reservation would be defined by the type and
scope=20
fields and the new registration key would be the reservation key fields =
value. OR=20
- The target does the preempt by clearing the reservation and does
nothing=20
with the registration. It knows not to do anything because the
reservation=20
key and the service action reservation key are equal. Because they are=20
equal the target knows only the reservation type and scope are going to
be=20
changed. This is all allowed because to the following statement in
section=20
5.5.3.6.3.3:=20
A PERSISTENT RESERVE OUT specifying a PREEMPT service action with the=20
SERVICE ACTION RESERVATION KEY value equal to the reservation key is =
not
an=20
error. In that case the device server shall establish the new
reservation.=20
Bye for now,=20
George Penokie=20
Dept 2C6 114-2 N212=20
E-Mail: gpenokie at tivoli.com=20
Internal: 553-5208=20
External: 507-253-5208 FAX: 507-253-2880=20
"Ulrich, David" <dulrich at lsil.com>@t10.org on 11/29/2000 11:11:55 AM=20
Sent by: owner-t10 at t10.org=20
To: "'t10 at t10.org'" <t10 at t10.org>=20
cc:=20
Subject: FW: persistent reservations follow-up question=20
George,=20
Thanks for the reply.=20
I have considered what you said below and while looking at figure 2 and =
following the flowchart another question has come to mind. Assuming
that=20
no error conditions are encountered wouldn't we eventually come to the=20
point of removing the registration for the initiator holding the=20
reservation, which in this case also happens to be the preemptor,
removing=20
the reservation, and then attempting create a new reservation for the=20
initiator we just removed the registration information for?=20
Regards,=20
David Ulrich=20
-----Original Message-----=20
From: George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com=20
[ mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com
<mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com> ]=20
Sent: Monday, November 27, 2000 10:00 AM=20
To: Ulrich, David=20
Cc: 't10 at t10.org'=20
Subject: Re: persistent reservations=20
* From the T10 Reflector (t10 at t10.org), posted by:=20
* George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com=20
*=20
David,=20
The logical unit would respond as no differently than if the PREEMPT =
AND
ABORT came from an initiator that did not hold the reservation. (i.e.,
it=20
would follow the flow chart in figure 2 to preempt and the description
in=20
section 5.5.3.6.4 to abort.=20
Bye for now,=20
George Penokie=20
Dept 2C6 114-2 N212=20
E-Mail: gpenokie at tivoli.com=20
Internal: 553-5208=20
External: 507-253-5208 FAX: 507-253-2880=20
"Ulrich, David" <dulrich at lsil.com>@t10.org on 11/20/2000 09:52:51 AM=20
Sent by: owner-t10 at t10.org=20
To: "'t10 at t10.org'" <t10 at t10.org>=20
cc:=20
Subject: persistent reservations=20
Is there any behavior defined if an initiator issues a PREEMPT AND =
ABORT
command to a logical unit that it has reserved?=20
*=20
* For T10 Reflector information, send a message with=20
* 'info t10' (no quotes) in the message body to majordomo at t10.org=20
------_=_NextPart_001_01C3AAE6.E43C01E6
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
FW: FW: persistent reservations follow-up question Hi Charles, I had a very similar question at one time and got = this reply from George. I'm sure the wording has changed since = then, but this was the original intent. David Ulrich
LSI Logic Storage Systems Division (today) -----Original Message-----
From: = George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com
[= mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com]
Sent: Friday, December 01, 2000 8:04 PM
To: Ulrich, David
Cc: 't10 at t10.org'
Subject: Re: FW: persistent reservations follow-up = question
David, Not really. You can look at in one of two ways, both = end in the same
result.
-The target does the preempt by clearing the = reservation and registration
then establishing a new reservation and new = registration before completing
the operation. The new reservation would be defined = by the type and scope
fields and the new registration key would be the = reservation key fields
value. OR
- The target does the preempt by clearing the = reservation and does nothing
with the registration. It knows not to do anything = because the reservation
key and the service action reservation key are = equal. Because they are
equal the target knows only the reservation type and = scope are going to be
changed. This is all allowed because to the = following statement in section
5.5.3.6.3.3: A PERSISTENT RESERVE OUT specifying a PREEMPT service = action with the
SERVICE ACTION RESERVATION KEY value equal to the = reservation key is not an
error. In that case the device server shall = establish the new reservation. Bye for now,
George Penokie Dept 2C6 114-2 N212
E-Mail: gpenokie at tivoli.com
Internal: 553-5208
External: 507-253-5208 FAX: = 507-253-2880
;Ulrich, David; = <dulrich at lsil.com>@t10.org on 11/29/2000 11:11:55 AM Sent by: owner-t10 at t10.org
To: ;'t10 at t10.org'; = <t10 at t10.org>
cc:
Subject: FW: persistent reservations follow-up = question
George, Thanks for the reply. I have considered what you said below and while = looking at figure 2 and
following the flowchart another question has come to = mind.=A0 Assuming that
no error conditions are encountered wouldn't we = eventually come to the
point of removing the registration for the initiator = holding the
reservation, which in this case also happens to be = the preemptor, removing
the reservation, and then attempting create a new = reservation for the
initiator we just removed the registration = information for? Regards,
David Ulrich -----Original Message-----
From: = George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com
[= mailto:George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com]
Sent: Monday, November 27, 2000 10:00 AM
To: Ulrich, David
Cc: 't10 at t10.org'
Subject: Re: persistent reservations * From the T10 Reflector (t10 at t10.org), posted = by:
* = George_Penokie/Tivoli_Systems%TIVOLI_SYSTEMS at us.ibm.com
*
David,
The logical unit would respond as no differently = than if the PREEMPT AND
ABORT came from an initiator that did not hold the = reservation. (i.e., it
would follow the flow chart in figure 2 to preempt = and the description in
section 5.5.3.6.4 to abort. Bye for now,
George Penokie Dept 2C6=A0 114-2 N212
E-Mail:=A0=A0=A0 gpenokie at tivoli.com
Internal:=A0 553-5208
External: 507-253-5208=A0=A0 FAX: = 507-253-2880 ;Ulrich, David; = <dulrich at lsil.com>@t10.org on 11/20/2000 09:52:51 AM Sent by:=A0 owner-t10 at t10.org To:=A0=A0 ;'t10 at t10.org'; = <t10 at t10.org>
cc:
Subject:=A0 persistent reservations
Is there any behavior defined if an initiator issues = a PREEMPT AND ABORT
command to a logical unit that it has = reserved?
*
* For T10 Reflector information, send a message = with
* 'info t10' (no quotes) in the message body to = majordomo at t10.org
------_=_NextPart_001_01C3AAE6.E43C01E6--
More information about the T10
mailing list