Clearing effects of logouts on different protocols

JoeBre at exabyte.com JoeBre at exabyte.com
Wed Feb 27 14:56:47 PST 2002


* From the T10 Reflector (t10 at t10.org), posted by:
* JoeBre at exabyte.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_01C1BFE2.09080840
Content-Type: text/plain;
	charset="iso-8859-1"

> * From the T10 Reflector (t10 at t10.org), posted by: 
> * "Rob Haydt" <robhay at windows.microsoft.com> 
> * 
> I didn't intend to imply either one. 
> 
> The addition of another mode page is only practical if it gets used. 
> Since that would require new software, most software wouldn't use it. 

While the proposal is not to add a new mode page, but rather to add a
new bit to an existing mode page, your concern is understood.

... 
> For iscsi and srp, I'm assuming that "losing the connection" 
> between the 
> initiator and the target means losing all "non-persistent state" (the 
> stuff in quotes is deliberately vague). From a software perspective, 
> it's easier to assume that all previous state information is invalid 
> than try to figure out what might still be valid. 

Unfortunately, the problem is not so simple. Picture a degenerate Fibre
Channel SAN made up of several workstations and one tape drive. It is
the nature of workstations that they are rebooted from time to time.
Upon booting, typical hosts log in to any device they can find on the
SAN. This login results in first processing an implicit logout to the
device. This logout destroys current mode parameters in the tape drive,
replacing them with default mode parameters, as mandated by FCP-2. At
best, this results in blowing away any backup operation in process. It
is not difficult to envision cases where this will also lead to data
loss. 

I am not sure that the proposed mechanism is the proper way to solve the
problem. I never quite understood why FCP could not allow mode
parameters to exist across login sessions. However, just ignoring the
problem will not eliminate it.

Joe Breher 
Exabyte Corp 

> -rob 
> 
> -----Original Message----- 
> From: Mallikarjun C. [ mailto:cbm at rose.hp.com 
] 
> Sent: Saturday, February 23, 2002 5:05 PM 
> To: Rob Haydt 
> Cc: t10 at t10.org 
> Subject: Re: Clearing effects of logouts on different protocols 
> 
> * From the T10 Reflector (t10 at t10.org), posted by: 
> * "Mallikarjun C." <cbm at rose.hp.com> 
> * 
> Rob, 
> 
> Can you please comment if you are implying that (Reserve-Release 
> managed)  reservations are not used by software? Or, software is 
> fine with FCP-like behavior, and so isn't likely to ever use the mode 
> page? 
> 
> Thanks. 
> 
> Mallikarjun 
> 
> ----- Original Message ----- 
> From: Rob Haydt 
> To: t10 at t10.org 
> Sent: Friday, February 22, 2002 8:38 AM 
> Subject: RE: Clearing effects of logouts on different protocols 
> 
> 
> * From the T10 Reflector (t10 at t10.org), posted by: 
> * "Rob Haydt" <robhay at windows.microsoft.com> 
> * 
> iScsi and SRP should clear. 
> 
> Don't add a mode page. Why would anyone want to implement this if 
> software isn't going to use it? 
> 
> -rob 
> 
> -----Original Message----- 
> From: Elliott, Robert [ mailto:Robert.Elliott at COMPAQ.com
 ] 
> Sent: Thursday, February 21, 2002 9:50 AM 
> To: t10 at t10.org 
> Subject: Clearing effects of logouts on different protocols 
> 
> <misc. text removed> 
> 
> The questions are: 
> * should iSCSI clear or not? 
> * should SRP clear or not? 
> * should we add a control mode page bit to allow not clearing? 
> * should mode pages/log pages/etc. all be handled the same? 
> 
> --- 
> 
> * 
> * 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 
> 


------_=_NextPart_001_01C1BFE2.09080840
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">

RE: Clearing effects of logouts on different protocols > * From the T10 Reflector (t10 at t10.org), posted = by: 
> * ;Rob Haydt; = <robhay at windows.microsoft.com> 
> * 
> I didn't intend to imply either one. 
> 
> The addition of another mode page is only = practical if it gets used. 
> Since that would require new software, most = software wouldn't use it. While the proposal is not to add a new mode page, but = rather to add a new bit to an existing mode page, your concern is = understood. ... 
> For iscsi and srp, I'm assuming that = ;losing the connection; 
> between the 
> initiator and the target means losing all = ;non-persistent state; (the 
> stuff in quotes is deliberately vague). From a = software perspective, 
> it's easier to assume that all previous state = information is invalid 
> than try to figure out what might still be = valid. Unfortunately, the problem is not so simple. Picture = a degenerate Fibre Channel SAN made up of several workstations and one = tape drive. It is the nature of workstations that they are rebooted = from time to time. Upon booting, typical hosts log in to any device = they can find on the SAN. This login results in first processing an = implicit logout to the device. This logout destroys current mode = parameters in the tape drive, replacing them with default mode = parameters, as mandated by FCP-2. At best, this results in blowing away = any backup operation in process. It is not difficult to envision cases = where this will also lead to data loss. I am not sure that the proposed mechanism is the = proper way to solve the problem. I never quite understood why FCP could = not allow mode parameters to exist across login sessions. However, just = ignoring the problem will not eliminate it. Joe Breher 
Exabyte Corp > -rob 
> 
> -----Original Message----- 
> From: Mallikarjun C. [mailto:cbm at rose.hp.com] 
> Sent: Saturday, February 23, 2002 5:05 = PM 
> To: Rob Haydt 
> Cc: t10 at t10.org 
> Subject: Re: Clearing effects of logouts on = different protocols 
> 
> * From the T10 Reflector (t10 at t10.org), posted = by: 
> * ;Mallikarjun C.; = <cbm at rose.hp.com> 
> * 
> Rob, 
> 
> Can you please comment if you are implying that = (Reserve-Release 
> managed)  reservations are not used by = software? Or, software is 
> fine with FCP-like behavior, and so isn't = likely to ever use the mode 
> page? 
> 
> Thanks. 
> 
> Mallikarjun 
> 
> ----- Original Message ----- 
> From: Rob Haydt 
> To: t10 at t10.org 
> Sent: Friday, February 22, 2002 8:38 AM 
> Subject: RE: Clearing effects of logouts on = different protocols 
> 
> 
> * From the T10 Reflector (t10 at t10.org), posted = by: 
> * ;Rob Haydt; = <robhay at windows.microsoft.com> 
> * 
> iScsi and SRP should clear. 
> 
> Don't add a mode page. Why would anyone want to = implement this if 
> software isn't going to use it? 
> 
> -rob 
> 
> -----Original Message----- 
> From: Elliott, Robert [mailto:Robert.Elliott at COMPAQ.c= om] 
> Sent: Thursday, February 21, 2002 9:50 = AM 
> To: t10 at t10.org 
> Subject: Clearing effects of logouts on = different protocols 
> 
> <misc. text removed> 
> 
> The questions are: 
> * should iSCSI clear or not? 
> * should SRP clear or not? 
> * should we add a control mode page bit to = allow not clearing? 
> * should mode pages/log pages/etc. all be = handled the same? 
> 
> --- 
> 
> * 
> * 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 
> 
------_=_NextPart_001_01C1BFE2.09080840--




More information about the T10 mailing list