Re: iscsi : default iscsi mode page settings.

["Sanjeev Bhagat \(TRIPACE/Zoetermeer\)" <iscsi_t10@sanjeevbhagat.com>]

Santosh,

Yes agreed!!, this works fine with me. But now the target will have to
support different values of FirstBurstSize for different initiators.

But should this field be kept in mode page if it will be dynamically
configurable and it value may vary for different inititators. This goes
against the semantics of mode pages. However keeping them in mode page may
mean similarity and compatibility with FCP-2

Sanjeev


----- Original Message -----
From: "Santosh Rao" <santoshr@cup.hp.com>
To: "Sanjeev Bhagat (TRIPACE/Zoetermeer)" <iscsi_t10@sanjeevbhagat.com>
Cc: <ips@ece.cmu.edu>
Sent: Monday, September 24, 2001 8:09 PM
Subject: Re: iscsi : default iscsi mode page settings.


> "Sanjeev Bhagat (TRIPACE/Zoetermeer)" wrote:
>
> > I think the FirstBurstSize be kept as login key but here we will have a
> > problem as to which value be over-ridden whether the one sent by
initiator
> > or target.
>
> Sanjeev,
>
> I don't think there is an ambiguity on which value is to be accepted. The
final
> negotiated value is agreed upon by both parties. See below definition of
> numerical key negotiation, taken from Section 2.2.4 of Rev 7.97 :
>
> " In numerical negotiations, the offering and responding party state a
> numerical value. The result of the negotiation is key dependent;
frequently the
> lower or the higher of the two values is used. "
>
> By defining the FirstBurstSize as a login key and specifying that the
lower of
> the 2 numbers is accepted, there is no ambiguity over the final chosen
value.
> This is true of all numerical negotiations. See the definition of
> MaxConnections as an example :
>
> "MaxConnections=<number-from-1-to-65535>
>
> Default is 8.
>
> Initiator and target negotiate the maximum number of connections
> requested/acceptable.  The lower of the 2 numbers is selected."
>
> Regards,
> Santosh
>
>
> > This depends upon whether
> >
> > a. the iSCSI target should be made so efficient as to support different
> > values of Firstburstsize for different initiators. If it is so then the
> > target accepts the Initiator value. This increases complexity on target
side
> > or
> > b. the iSCSI target should operate on one particular mode of
FirstBusSize.
> > If it is so then initiator accepts target value. But in this case also I
> > foresee a problem in this case when multiple intiators are accessing a
iSCSI
> > target and one of the iSCSI initiator changes the value of FirstBusSize
>