HP StorageWorks HSG80 Enterprise Modular Storage RAID Array Fibre Channel Solution Software V8.8 for Tru64 UNIX Release Notes (AA-RV1WA-TE, March 2005)

ManualsBrandsHP ManualsSoftwareHP HSG80 and HSG60 ACS Patches

ACS Enhancements and Fixes

25HSG80 Enterprise Modular Storage RAID Array Fibre Channel Solution Software V8.8 for Tru64

UNIX Release Notes

12. Controller inoperability and

LFC 64030104

With this ACS version, there is no longer a conflict with the

use of the previously unsupported SET HOST/SCSI

OpenVMS command.

Additionally, Issues surrounding controller inoperability

problems which resulted from two different entities

executing send and receive diagnostic commands to the

controller and resulting in LFC 64030104 has been

resolved.

13. Controller inoperability due to

LFC 018F2087

The issue involving controller inoperability problems while

using FRUTIL which resulted in LFC 018F2087 has been

resolved.

14. Controller inoperability due to

metadata errors with

single-member mirror sets and

LFC 12000103 on both

controllers

An issue involving a controller inoperability event, as a

result of an unrecoverable read on container metadata

(medium error) and the mirror unit, is a single member

mirror. The controller failure is recursive, with an

LFC 12000103 reason code. A new and unique reason

code with a new repair action has been created. The

recursive failure has been eliminated, and now, after the

controller restarts, access to the rest of the storage occurs.

The Repair Code directs activities necessary to recover the

unit impacted by the device metadata read issue.

Note: If a mirrorset member is added to a current single

member mirrorset, the controller completely reads the

metadata to validate that the mirrorset is without error. If

the controller is under a heavy I/O load (near 0% idle)

while the controller validates mirrorset metadata, it can

take up to 4 or 5 minutes (on 146 GB drives) to read the

metadata before adding the new member to the single

member mirrorset. If the mirrorset member is smaller, the

amount of time it takes to validate mirrorset metadata is

reduced proportionally. If no load exists on the controller,

the metadata check completes within 2 to five seconds.

15. Controller ejecting devices

after bus device resets

ACS now ejects any device (if a member of the redundant

storageset) that is responsible for excessive

controller-initiated SCSI bus resets.

Table 4: ACS fixes (Continued)

Item Subject Description