During normal operation, the cluster functions continuously
without intervention. When it becomes necessary to stop the entire
cluster for such operations as replacing hardware or physically
moving the nodes, you can manually halt the cluster and restart
it at a later time. Stopping the cluster in this way has the effect
of running the OPS halt scripts on each node to halt the operation
of OPS and OPS applications as the cluster shuts down. Restarting
the cluster has the effect of running the OPS start script on each
node (if it is configured) to bring up Oracle Parallel Server and
its applications.
Using SAM to Stop the Cluster |
 |
Run SAM, and choose the High Availability options.
Choose Cluster Administration, and select "Shut
Down Cluster."
Respond Yes to the verification prompt.
Using HP-UX Commands to Stop the Cluster |
|
To stop the entire cluster:
If you are running packages on your cluster, halt them.
Shutdown the OPS database.
If there are volume groups used by packages that
have mounted file systems, unmount them.
Deactivate each cluster-bound volume group.
# vgchange -a n /dev/vg_ops |
Use the cmhaltcl command from any one node to stop the entire cluster. cmhaltcl causes all nodes in a configured cluster to stop
their MC/LockManager daemons.
This command will halt all Oracle instances and stop all the MC/LockManager daemons
on all currently running systems. If you only want to shut down
a subset of nodes, the cmhaltnode command should be used instead.
Using SAM to Start the Cluster |
|
Run SAM, and choose the High Availability options.
Choose Cluster Administration, and select "Start
Cluster."
Select "Start Cluster on All Nodes."
Respond Yes to the verification prompt.
Printable version
