On Thu, 11 Jan 2001, Jeremy Higdon wrote:
> On Jan 11, 12:44pm, Martin Peschke wrote:
> >
> > On Wed, 10 Jan 2001, Prasenjit Sarkar wrote:
> >
> > >
> > >
> > > I am aware of the 64-bit LUN specification, but can you clarify what you
> > > meant by 'id'?
> >
> > SCSI ID
> >
> > > If you meant the WWN-type identifier in FCP,
> >
> > Seems, that FCP-2 [ftp://ftp.t10.org/t10/drafts/fcp2/fcp2r04.pdf]
> > maps FCP term "address identifier of target port" (should be D_ID) to
> > SCSI term "target identifier" (SCSI ID, I think) - see Annex A.1.
> > Since D_IDs are volatile, I think it would be nice to use the WW(P)N of
> > targets for device discovery (see FCP-2, annex F.1).
> >
> > >then I think it should not be
> > > part of the scan
> > > procedure (as it is currently: 0 to max_id) and not something the SCSI
> > > upper layer
> > > should be concerned with until you implement multipathing.
> >
> > to find alternate paths by means of WWN? Yes.
>
> There will have to be distinction between the node name and the
> port name (from what I can see, WWN can refer to either, though
> maybe a more precise definition has been made).
Well, WWN may be either WWPN (port name) or WWNN (node name).
According to the mapping of targetID and D_ID, WWPN seems to be the right
choice.
> Of course, some devices allow access to a given storage through
> different nodes as well as different ports. There are a vendor
an example:
http://www.storage.ibm.com/hardsoft/products/ess/support/essfcwp.pdf
> unique methods of determining when that is the case, unfortunately.
> I presume such devices are beyond the scope of this proposal.
Do you know more about such methods?
Martin
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
