Clearing ACA on CHECK CONDITION

Gerry_Houlder at notes.seagate.com Gerry_Houlder at notes.seagate.com
Thu Jun 1 10:48:17 PDT 2000


* From the T10 Reflector (t10 at t10.org), posted by:
* Gerry_Houlder at notes.seagate.com
*

You are correct that the previous ACA condition is cleared.  Also remember
that a new ACA (or CA if the NACA bit is 0) condition will be created
because of the new CHECK CONDITION. This new contingent allegiance should
still leave the drive in a state where the desired recovery can be done by
the intended initiator. Of course it is strongly recommended that, once an
initiator sets the NACA bit in commands, all commands should have the NACA
bit set so the error recovery behavior is consistent.





Costa Sapuntzakis <csapuntz at cisco.com>@t10.org on 06/01/2000 12:05:51 PM

Sent by:  owner-t10 at t10.org


To:   t10 at t10.org
cc:

Subject:  Clearing ACA on CHECK CONDITION


* From the T10 Reflector (t10 at t10.org), posted by:
* Costa Sapuntzakis <csapuntz at cisco.com>
*

In Section 5.6.1.1 of the SAM-2 03/27/00, it states that:

If the logical unit accepts a value of one for the NACA bit and
this bit was set to one in the CONTROL Byte of the faulting comand,
then the SCSI-2 rules for clearing a contignent aleegiance condition
shall not apply, In this case, the ACA condition shall only be cleared
by:
....
d) A command with the ACA attribute terminates with a CHECK CONDITION
status

---------------------

Part d) seems dangerous. For example, if during recovery, the initiator
sends an unsupported CDB and gets CHECK CONDITION,
ACA is cleared!

This seriously limits the suitability of ACA as a general error-recovery
mechanism.

-Costa

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



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




More information about the T10 mailing list