Re: iSCSI: NOPs on discovery session

[Mark Bakke <mbakke@cisco.com>]

If a client wants to receive async messages when new iSCSI targets
become available, or wants to poll once in a while with SentTargets,
it is better to keep the discovery session there.  This is useful
in cases where iSCSI gateways are connected through firewalls; only
the iSCSI protocol needs to be allowed through the firewall to do
this type of discovery.

--
Mark

John Hufferd wrote:
> 
> I do not see any reason to keep the decovery session up.  Why would it be
> needed to stay up?
> 
> .
> .
> .
> John L. Hufferd
> Senior Technical Staff Member (STSM)
> IBM/SSG San Jose Ca
> Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
> Home Office (408) 997-6136, Cell: (408) 499-9702
> Internet address: hufferd@us.ibm.com
> 
> "Ayman Ghanem" <aghanem@cisco.com>@ece.cmu.edu on 12/05/2001 09:37:34 PM
> 
> Sent by:  owner-ips@ece.cmu.edu
> 
> To:   <ips@ece.cmu.edu>
> cc:
> Subject:  RE: iSCSI: NOPs on discovery session
> 
> Julian,
> 
> I would vote for  allowing it. Is the spec going to require the initiator
> to logout immediately  after the SendTargets response is received?. Should
> the target drop the session  after it had sent the response?. I am not sure
> how the  discovery session could be misused. It has been authenticated like
> any other  session and has only limited set of commands. One could also
> argue the need for  logout on a discovery session as opposed to just
> closing the  connection.
> 
> -Ayman
> 
> -----Original Message-----
> From: owner-ips@ece.cmu.edu  [mailto:owner-ips@ece.cmu.edu]On Behalf Of
> Julian  Satran
> Sent: Tuesday, December 04, 2001 1:05 AM
> To:  ips@ece.cmu.edu
> Subject: Re: iSCSI: NOPs on discovery  session
> 
> Ayman,
> 
> It is true that I agreed that we may want  to have it. However some members
> of the list have expressed immediately a  strong opposition to it claiming
> that it was no strictly needed (true) and  that discovery section are meant
> to be brief.
> I felt thus that we don't have a consensus on that and we may want to  try
> again.
> I assume that the main  concern was that this will enable discovery
> sessions to be become long lived  and be used as a back door monitoring
> channel.
> 
> Regards,
> Julo
> 
> 
>                              "Ayman Ghanem"
>                              <aghanem@cisco.com>              To:
>                              Sent by:                 <ips@ece.cmu.edu>
>                              owner-ips@ece.cmu.edu            cc:
>                                                                Subject:
>                              12-04-01 05:30 AM        iSCSI: NOPs on
>                                                       discovery  session
> 
> 
> 
> 
> Julian,
> 
> I thought we agreed to adding NOPs to the operations  allowed on the
> discovery session. I couldn't find that in draft-09.  Thanks.
> 
> -Ayman