Re: Connection recovery using same TSID-TSIH-CID

[Julian Satran <Julian_Satran@il.ibm.com>]

Eddy - connection allegiant is a bit
fuzy when you are talking about a connection that this not exist anymore.
For those cases we decided that the task will go into a "frozen"
state for a while. If it is not reassigned during this interval it will
"evaporate". Every task has to be specifically reassign. Special
attention should be paid to tasks that the initiator is not sure they exist
(the CmdSN ack was not seen). Those have to be resent and reassigned (and
the target will either ignore the reassignment or the resend).

Regards,
Julo

"Eddy Quicksall" <eddy_quicksall_iVivity_iSCSI@comcast.net>
Sent by: ips-bounces@ietf.org

18/09/04 22:25

To	"Mallikarjun C." <cbm@rose.hp.com>, <ips@ietf.org>
cc
Subject	Re: Connection recovery using same TSID-TSIH-CID

Doesn't the following paragraph imply that a task
stays connection allegiant
until it has been reassigned?



6.5.  Implicit Termination of Tasks



      b)  When a connection fails and the connection
state eventually

          times out (state transition M1 in Section
7.2.2 State

          Transition Descriptions for Initiators
and Targets) and there

          are active tasks allegiant to that connection.



Eddy

----- Original Message ----- 
From: "Mallikarjun C." <cbm@rose.hp.com>
To: <ips@ietf.org>
Sent: Friday, September 17, 2004 2:21 PM
Subject: Re:  Connection recovery using same TSID-TSIH-CID


>what is the rational for requiring the TMF if the task is on
 > the same connection?

Task is actually on no connection - it's a free agent since the
connection reinstatement is an implicit logout.  Again 5.3.4 says
it.

The principle we applied was to keep the reassignment process always the
same - implicit or explicit logout, connection reinstatement or no
reinstatement, reassignment all onto one connection or to multiple
connections.
-- 
Mallikarjun

Mallikarjun Chadalapaka
Networked Storage Architecture
Network Storage Solutions
Hewlett-Packard MS 5668
Roseville CA 95747
cbm [at] rose.hp.com



Eddy Quicksall wrote:

> Interesting, what is the rational for requiring the TMF if the task
is on
> the same connection?
>
> What I'm referring to is:
>    Connection reinstatement is the process of an initiator
logging in
>
>    with an ISID-TSIH-CID combination that is possibly active
from the
>
>    target's perspective, which causes the implicit logging
out of the
>
>    connection corresponding to the CID,  and reinstating
a new Full
>
>    Feature Phase iSCSI connection in its place (with the
same CID).
>
> Eddy
>
> ----- Original Message ----- 
> From: "Mallikarjun C." <cbm@rose.hp.com>
> To: <ips@ietf.org>
> Sent: Thursday, September 16, 2004 5:56 PM
> Subject: Re:  Connection recovery using same TSID-TSIH-CID
>
>
>
>>After closing the connection, can the initiator simply login with
an
>
>  > ISID-TSIH-CID combination that is the same as the failed
connection,
>  > send the appropriate SNACK and continue – without doing
the Task
>  > Reassign TMF?
>
> No, task reassignment is always explicit.
>
> Section 5.3.4 says: "If the operational ErrorRecoveryLevel is
2,
> connection reinstatement enables future task reassignment."




_______________________________________________
Ips mailing list
Ips@ietf.org
https://www1.ietf.org/mailman/listinfo/ips


_______________________________________________
Ips mailing list
Ips@ietf.org
https://www1.ietf.org/mailman/listinfo/ips

_______________________________________________
Ips mailing list
Ips@ietf.org
https://www1.ietf.org/mailman/listinfo/ips