5
Troubleshooting and Component
Replacement Guidelines
This chapter discusses the following topics:
•
The failure behavior of Fibre Channel components (Section 5.1)
•
The requirements and limitations for adding storage (Section 5.2)
•
Adding or replacing Fibre Channel devices (Section 5.3)
•
Adding or replacing a cluster member (Section 5.4)
5.1 Failure Behavior of Fibre Channel Components
This section describes the failure behavior of the following components; later
sections discuss considerations for the addition or replacement of each of
these components in a cluster:
•
Fibre Channel adapter (KGPSA) (Section 5.1.1)
•
Fibre Channel switch (Section 5.1.2)
•
Fibre Channel hub (Section 5.1.3)
•
HSG80 RAID array controller (Section 5.1.4)
•
Disks (Section 5.1.5)
5.1.1 Fibre Channel (KGPSA) Adapter Failures
The failure of a KGPSA adapter has similar effects as that of a KZPBA
adapter in a parallel SCSI configuration. Although either adapter type is
primarily a SCSI interface, they differ in medium. Certain failure modes
and their associated problems do not exist in the Fibre Channel domain. For
example, each Fibre Channel fabric connection is a simple point-to-point
connection, rather than the multidrop configuration of the parallel SCSI bus.
A failing Fibre Channel adapter generally drops off line, making storage
unavailable to its hosting member only.
In a TruCluster configuration, the ASE software detects the failure of a
KGPSA adapter and, depending on the service configuration, automatically
relocates services from the member with the failed adapter to a healthy
Troubleshooting and Component Replacement Guidelines 5–1
