HP NetRAID-4M Configuration and Upgrade Guide (Release 5)

Chapter 9 Issues and Problem Resolution
44
Problem: One of the new features that disk drives offer is SMART (Self-Monitoring
And Reporting Technology). The new SMART compliant disk drives support
informational exception conditions, including the SCSI sense codes FAILURE
PREDITION THRESHOLD EXCEEDED, or WARNING. There are 7 MRIE
(Method of Reporting Informational Exceptions):
0 - The device will not report exception conditions.
1 - (Not Supported.)
2 - Report exception before a SCSI command.
3 - Report after a SCSI command if reporting of RECOVERED ERROR is
allowed.
4 - Report after a SCSI command as RECOVERED ERROR always.
5 - Report after a SCSI command as NO SENSE.
6 - Report only when polled.
The original design of NetRAID-4M had defaults to MRIE 6 with an option for the
end user to change the setting to any of these options. The MRIE default has been
changed from 6 to 2 to support multi-initiator.
Impact to Customer: Minimal. All drives support polling. The new feature will
decrease any possible overhead that could be associated with polling.
Frequency of Occurrence: At all times.
Resolution: The new functionality has been implemented and is available with
Release 3.
18 Drive firmware download process can cause drives to go offline
Problem: During the firmware update, on some hard disk drives with the NetRAID-
4M, the process fails and takes the drive offline-causing containers to become either
degraded or critical. The firmware flashing process causes the drives to not respond
to the SCSI Selection within a certain period of time. It occurs every time that the
Background Consistency is enabled.
Background: The hard disk drives go offline because they are not able to respond
to the NetRAID-4M’s SCSI selections at the final step of the firmware upgrade. If a
drive does not respond to a certain number of selections, the NetRAID-4M then
assumes that the hard disk drive is no longer there, takes the drive offline, and
changes the container to a degraded or critical state. The firmware upgrade is
successful, but sometimes does not appear so until you reboot the server.