Failover Packages; Eliminating Single Points Of Failure - HP StorageWorks 8000 - NAS User Manual

Failover Packages

Eliminating Single Points of Failure

Failover packages are the smallest unit of failover within the cluster. A
package contains necessary definitions and configuration information
relating to resources and their processes that must be failed over to the
secondary server in the event the primary server fails. Each cluster can have a
maximum of 30 packages running concurrently. For NAS, the package
defines the volumes (file systems) and their associated CIFS shares and NFS
exports that should be failed over. A given volume group can be defined in
only one package at a time, but a package can contain multiple volume
group definitions. The packages can fail over automatically when a server
fails, or they can be manually failed over one at a time. A given package can
be running on only one cluster node at a time.
Think of a package as a group of one or more volume groups (with their file
systems and shares/exports) that will fail over as a single unit. To fail over a
package manually, you need to:
Stop the existing package (in the case of a service, network, or resource
I
failure).
Start the new instance of the package on a different node.
I
You can manage failover packages on the Storage tab of the Command View
NAS web interface.
Most problems that result in service outages are single-level failures. High-
availability lets you quickly detect and handle these failures and minimize
downtime. Examples of single-level failures include:
NIC failures
I
NFS failure
I
SMB failure
I
Operating system failure
I
Power failure
I
NAS 8000 Concepts 27