Re: logout command

[julian_satran@il.ibm.com]

Mark,

Will - fix text - Julo

Mark Burton <markb@ordern.com> on 27/02/2001 20:59:37

Please respond to Mark Burton <markb@ordern.com>

To:   Julian Satran/Haifa/IBM@IBMIL
cc:
Subject:  Re: logout command





Julo,

Sorry to be dim, but can you please confirm that when a target
receives a logout PDU with a reason code of 0 (closing the session) it
should kill all of the connections associated with the session that
receives the logout PDU (and kill the session as well).

If that is true, the CID does not need to be specified when reason is 0.

Also, the spec says "after logout the TCP connection must be closed at
both ends". Please confirm that this means the TCP connection for the
connection that was logged out is closed (which is not necessarily the
same as the connection used to deliver the Logout request).

Thanks,

Mark

---------------

From: julian_satran@il.ibm.com
Subject: Re: logout command
Date: Tue, 27 Feb 2001 16:14:56 +0200

>
>
> Mark,
>
> 2.14.3 reason code 0
>
> I will rephrase it to be more specific (like:
>
> 0 - closing the session
>
> ExpStaSN - when the connection is closed for recovery it tells the target
> what was the last
> StatSN it has seen as the target might not have seen previous acks and
the
> initiator may think the target has seen them (this issue was raised a
while
> ago by Pierre Labat.
>
> Julo
>
>
> Mark Burton <markb@ordern.com> on 27/02/2001 14:42:07
>
> Please respond to Mark Burton <markb@ordern.com>
>
> To:   ips@ece.cmu.edu
> cc:
> Subject:  logout command
>
>
>
>
>
> Hello again,
>
> A couple of issues re the logout command (section 2.14 of 04)
>
> >   An initiator MAY use a logout command to remove a connection from a
> >   session or to close an entire session.
>
> How is the entire session closed? there doesn't seem to be any option
> in the PDU that says "zap all the connections in this session"
>
> >   This is the ExpStatSN for the connection to be closed.
>
> I don't understand the need for this, please explain.
>
> many thanks,
>
> Mark
>
>
>
>