After you create a new One of N core group policy
for a messaging engine, you configure the policy to specify the messaging
engine behavior, such as which server the messaging engine runs on,
and whether the messaging engine can fail over or fail back. You can
also configure the frequency of messaging engine monitoring.
About this task
You can use a One of N policy to run a messaging engine
in a cluster to enable failover. One server in the cluster
server runs the messaging engine and other servers in the cluster
act as standby servers, ready to run the messaging engine if it cannot
run in its current server. You can use a One of N policy
to run a messaging engine in a server bus member, but this is equivalent
to a cluster with only one server, that is, the value of N is 1, so
the messaging engine cannot fail over.
You can configure a One
of N policy for a messaging engine in a cluster to provide high
availability, or workload sharing with high availability, depending
on how you set the configuration options. See Policies for service integration.
Procedure
- Open the Policies page for the policy you are configuring.
Click .
- Optional: If required, define which servers
the messaging engine prefers to run on in a preferred servers list:
- Under Additional Properties, click Preferred
servers.
- Select the servers you require from the Core group servers
list, then click Add to add them to the Preferred
servers list.
Ensure that the servers you select are in
the server cluster where the messaging engine runs.
- Use Move up and Move
down to adjust the order of the list as required.
The earlier a server is in the preferred servers list, the stronger
the preference for that server.
- Click OK.
The messaging engine runs in the first available server in the
preferred servers list and fails over to the next available server
in the preferred servers list. If no preferred server is available,
the messaging engine can fail over to any other server in the cluster.
You
might use the preferred servers list if one server has more resources
available to it or typically performs less work than the others. You
might use the preferred servers list to help spread workload across
the cluster by configuring multiple policies and specifying a different
preferred server for each messaging engine. If you do not define a
preferred server list, the messaging engine runs on the first available
server in the cluster.
- Optional: If you defined a preferred servers
list, if required, restrict the messaging engine to run only on preferred
servers, by selecting the Preferred servers only check
box.
The messaging engine runs in the first available server
in the preferred servers list and fails over to the next available
server in the preferred servers list. The messaging engine cannot
run on a server that is not in the preferred servers list. If no preferred
server is available, the messaging engine cannot fail over.
You
can use this option together with a single server in the preferred
servers list to restrict a messaging engine to a specific server in
a workload sharing configuration. You can use this option together
with a preferred servers list to create a configuration that provides
workload sharing and aspects of high availability, for example, one
primary server and one failover server for each messaging engine.
If you require high availability, use this option with care, because
you can reduce or remove the high availability of the messaging engine.
- Optional: If you defined a preferred servers
list, if required, specify that the messaging engine automatically
fails back to a more preferred server by selecting the Fail
back check box.
If a messaging engine
is running on a server that is low in the preferred servers list,
or in the cluster but not in the preferred servers list (for example,
the messaging engine has failed over), the messaging engine automatically
fails back to a more preferred server when one becomes available.
- Ensure that the messaging engine can always reach its data
store or file store.
If the messaging engine can fail over,
that is, for any configuration with high availability characteristics,
the data store or file store must be accessible from any server in
the cluster on which the messaging engine might run.
The set of
possible servers depends on whether you defined a preferred servers
list and whether you selected the Preferred servers only option.
For example, a configuration might have a cluster of three servers,
server1, server2, and server3, with a single messaging engine that
uses a policy configured so that the messaging engine can fail over
to any of the servers in the cluster. The message store for
the messaging engine must be accessible from all three servers. However,
if the configured policy specified a preferred server list of server1
and server2, and the Preferred servers only option
is selected, only server1 and server2 would need access to the data
store or file store for that messaging engine.
- Optional: 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.
- Click OK.
- Save your changes to the master configuration.