Clearing effects table for FC-Tape

Rob Basham robbyb at us.ibm.com
Wed Jul 8 07:33:51 PDT 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* Rob Basham <robbyb at us.ibm.com>
*
=0ACharles,
 I understand the need for multiple port clearing in some cases, but I =
have
issue with some of these cases.  Let's start with power cycle...

For a normal tape drive with two ports and one LUN all controlled by th=
e same
power supply, the clearing on a power cycle makes sense as it stands in=
 the
table.

Let's suppose that instead you have a tape subsystem with the following=
:
N ports, each with a separate power source.
M LUNs, each with a separate power source.

Suppose that you lose power on an idle port.  Why should that affect I/=
Os on
other ports?  Also, suppose that you lose power on one of your LUNs.  W=
hy
should that affect I/Os to other LUNs?  This is not a hypothetical subs=
ystem by
any means.  We have small SCSI subsystems of this type today and with t=
he
advent of Fibre Channel, the values of N and M will only grow.

Also, power loss and power restoration need to be treated separately in=
 my
opinion.  There is one set of things to do on power loss and another se=
t of
things that need to be done on power restoration.

I plan on being at the meeting in Maine, and this is an issue I would l=
ike to
see raised.

Rob Basham
IBM Tucson


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* "Binford, Charles" <charles.binford at symbios.com>
*
Many actions effect multiple ports in the table because of the SCSI
queueing model.  Clear Task Set, for example, is defined by T10 to clea=
r
IOs for ALL initiators.  When that functionality is mapped to a
multiported FC node running FCP that means IOs received from multiple
ports are cleared.  Other actions such as power cycle and reset will
obviously effect multiple ports.

Charles Binford
Symbios Inc.
316-636-8566

> -----Original Message-----
> From: Rob Basham [SMTP:robbyb at us.ibm.com]
> Sent: Wednesday, July 01, 1998 4:25 PM
> To: T10 at symbios.com
> Subject: Clearing effects table for FC-Tape
>
> * From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by=
:
> * Rob Basham <robbyb at us.ibm.com>
> *
>
> Dave,
> This is a comment on the Clearing effects table for FC-Tape.  I have =
a
> question.  Why do so many actions affect more than just the one port
> on which
> the Initiator Action takes place?  It seems that with the wrong
> hardware
> defect, this has the potential to take down all ports (effectively) t=
o
> the tape
> drive, reducing the "value add" of alternate ports for high
> availability.
>
> Also, with the wrong topology choice I could envision a single
> initiator action
> on one loop cascading and causing problems on all loops.  Even with
> smart
> topology choices, it can at least cascade to the next loop over in a
> common
> dual loop topology.
> Regards,
> Rob
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at symbios.com=

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


=
*
* 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