After you create a new "No operation" core group policy
for a messaging engine, you can continue the configuration for a messaging
engine to be managed by an external high availability (HA) framework.
You can also configure the frequency of messaging engine monitoring.
About this task
A "No operation" policy allows an external HA cluster
to control when and where a messaging engine runs.
Procedure
- Associate the messaging engine with an externally managed
resource group by creating an HA cluster resource for it. Refer to
the documentation for your external HA product.
- Write scripts for the external HA framework to enable it
to use the HAManager to
start or stop the messaging engine. The scripts invoke operations
on the HAManager MBean
on the server that is taking ownership of the resource that represents
the messaging engine.
- Ensure that the messaging engine can always reach its data
store. If the messaging engine is configured to fail over,
its data store must be accessible from any server in the cluster on
which it might run. The set of possible servers depends on how you
have configured the external HA resource group. A typical configuration
is to include the data store as an additional resource in the resource
group managed by the external HA cluster. The HA cluster will then
ensure that the data store and the messaging engine failover together
and remain collocated after the failover. Whether this is the case,
or whether a network server is used to make the data store available,
the data store must be accessible from any server that might run the
messaging engine.
- If the messaging engine must always be accessible through
the same IP address, for example because it is the receiving end of
a WebSphere MQ link, you must arrange for
that IP address to remain collocated with the messaging engine. You
can do this by creating an IP address resource in the same external
HA resource group as the resource that represents the messaging engine.
- If required, enter a value in the Is
alive timer field. This value specifies the
interval of time, in seconds, at which the high availability manager
(HAManager) checks that
a messaging engine is running properly. When this value is 0 (zero),
the default value of 120 seconds is used.
The "No operation" policy
is intended primarily for when you use an external high availability
framework such as IBM HACMP. In this situation, you might create a
monitoring script that the external framework calls periodically.
If you set the value of Is alive timer greater
than or equal to 0, the HAManager performs
health monitoring and the external monitoring script can retrieve
the state from the HAManager MBean.
Alternatively, if you set the value of Is alive timer to
-1, monitoring by the HAManager is
disabled, and the external monitoring script can retrieve the state
from the messaging engine MBean.
- Click OK.
- Save your changes to the master configuration.