Question on Scope: 8-byte vs 2-byte LUNs

George Penokie gop at us.ibm.com
Wed Apr 1 14:55:45 PST 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* George Penokie <gop at us.ibm.com>
*
From: George Penokie/Rochester/IBM @ IBMUS
Subject: RE: Question on Scope: 8-byte vs 2-byte LUNs

George,
It looks like we are close. I have attach a file that contains a mini-tutorial
on addressing and configuration that may help in getting some of the details
and terminology worked out. (Note: for those only interested in the addressing
ignore the configuration steps in the second half of the document and replace
SCC device with bridge controller).

Some more on your comments:
a- I haven't though of the application client as being at level 0 but it works
so I have placed it in the mini-tutorial. Look at figure 1 to see how it fits
in.
b-a- Correct SCC commands only operate on level they are addressed to.
b-b- yes
c-a- yes
c-b- Zeros are filled in if there are no devices at those levels. Otherwise
there should be a value if the SCC device allows commands to be directed to
them. See mini-tutorial for detailed example.

d-a- I was a little lose with my terminology. You are correct the target does
not do the processing the copy manager does. The point I was trying to make was
that the copy managers address is a target/LUN and the source and destination
address information in the copy command includes target and LUN addressing
information.
d-b- I agree
d-c- Right the copy manager does not know want level it is at but should know
everything about what is attached to levels below it (see example) so it could
restructure the copy command and send it on. (I do not think this is a good
idea but it is possible)
d-d-a- Agree except that if the copy manager is the SCC device it could move
data between buses if those buses are attached to it and controlled by it.
d-d-b- I don't quite follow this but if it agrees with the example then OK.
d-d-c- Yes- you got it (or at least most of it)

This is a good exercise especially with the multi-level address moving into
SAM-2 (Ralph you out there). If you have any more comments/questions let me
know.

Note: Tutorial has been removed because it is too big for the reflector. If you
want a copy let me know.


Bye for now,
George Penokie


ericson at worldnet.att.net on 03/25/98 09:54:51 PM
Please respond to ericson at worldnet.att.net
To: t10 at Symbios.COM, George Penokie/Rochester/IBM at ibmus, owner-t10 at Symbios.COM
cc:
Subject: RE: Question on Scope: 8-byte vs 2-byte LUNs


This is a multi-part message in MIME format.
--------------------------------------------------------------------------------
George

Thanks for the info.  I think I understand the picture now.  However, I
didn't entirely follow your examples.
Are the following assertions correct?

    a.. Application client is at level 0, initiator port is at level 1, and
target LUs are at levels 1-4.
    b.. For an SCC command to a level n SCCS Logical Unit
        a.. SCC commands are unable to operate on LUs at lower (or higher)
levels than the SCCS LU.
        b.. The application client can directly send commands to LUs named
in 2-byte LUNs returned by an SCC command if the application client:
            a.. places the 2-byte LUN at level n in the 8-byte LUN
hierarchy, and
            b.. uses the same upper level (n-1) 2-byte LUNs as with the SCCS
LU.
    c.. For a REPORT LUNS command to a level n LU:
        a.. The LU processing the REPORT LUNS command will act as if it was
at level 1.
        b.. The returned 8-byte LUNs will always have the lower (n-1) level
LUNs filled with zeros.
        c.. The application client must shift the returned 8-byte LUNs by
(n-1) levels down and fill the upper level (n-1) 2-byte LUNs with the same
upper level (n-1) 2-byte LUNs used to address the LU that processed the
REPORT LUNS command.
    d.. For the COPY command you give several cases with the assumption that
the "Target" is doing the processing.
        a.. This doesn't match my understanding of SAM/SPC.
        b.. I'm expecting that the LU addressed by the COPY command will do
the processing (is the "copy manager") and that the source and destination
LUs may or may not be different from the copy manager LU.
        c.. The copy manager does not know what level it is at, so is never
in a position to manipulate the source or destination 8-byte LUNs based on
the copy manager's level.
        d.. Given this:
            a.. The application client must assume the responsibility of
casting the source and destination 8-byte LUNs into a form that can be
processed as if the copy manager, source, and destination LUs are all
connected via a common bus and that the copy manager can act as an initiator
on that bus and the source and destination LUNs will correctly address the
respective LUs when issued by the copy manager.  (In fact, the copy manager
will have to be rather clever to be able to know without some probing as to
whether or not either the source or destination LU is the same as the copy
manager LU.)
            b.. The above requirement implies that the application client
must shift up the source and destination 8-byte LUNs by (n-1) levels,
filling the lower (n-1) levels with zeros, where n is the level of the copy
manager LU.
            c.. This requirement says that a copy manager may only copy
between LUs at its level or lower (higher number).
Hope I've now got it.
Working though this has been helpful for me.
Thanks,
George Ericson


*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at symbios.com





More information about the T10 mailing list