Shared Disk Problem Isolation

- Disk Manager

- Resource kit tools

- Fast!Util

- FAStT Check

- Simplicity Storage Manager

- Disk resources online, node ownership

- State of shared drives, SCSI ID, merge IDs, host/partner, stripe size

• RAID Configuration and Setup Utility

- SCSI host id settings

• Other programs -IPSHAHTO -IPSSEND

Chkdsk cannot be run on a shared disk while resources contained on that disk are online. If you want to run chkdsk on the quorum drive, you must move the quorum to another drive first.

In SCSI installations, cables and termination should be checked first. In Fibre installations, use the error lights on the hub to determine fiber cable problems (you can also use the <Alt-Q> Fast!Util to check for connections).

ICSM allows you to check which node owns the shared disk and whether the resource is online.

NT/2000 Disk Manager will allow you to determine the state of the drives. Remember to manage only the drives owned by the local node.

The RAID Configuration and Setup Utility will also allow you to check drive status. In ServeRAID environments:

• IPSSEND will allow you to check for BIOS level and cluster configuration.

• IPSHAHTO can be used to isolate problems that may pertain to NT or MSCS. Be careful when using IPSHAHTO after MSCS is installed. The cluster service will have to be stopped on both nodes and ownership will have to be documented so that the drives can be returned to their original owners.

In Fibre Channel environments, Fast!Util and Netfinity Storage Manager are the primary tools for shared drive problem determination.

CAUTION: DO NOT RUN DIAGNOSTICS ON THE SHARED DISKS WITHOUT CONFIRMING CLUSTER AND NODE STATUS. RUNNING SHARED DISK DIAGNOSTICS REQUIRES THE CLUSTER TO BE DOWN.

Was this article helpful?

0 0

Post a comment