This section explains the possible scenarios when a use of message processor is run in a worker-manager cluster nodemode. In such situations, only workers are required to process/serve client requests while managers are mainly required to carry out administration activities and synchronize the administrative information of the workers in the cluster. A message processor may be executed by one or more worker nodes.
Usually, a message processor is executed by one worker node with the current ESB implementation. However, there can be scenarios where a message processor is simultaneously executed by more than one worker node in the cluster to optimise performance. The number of worker nodes that is allowed to execute the message processor at a given time can be controlled by specifying the required number When WSO2 ESB is deployed in a clustered mode, you can increase the number of tasks for a particular message processor by specifying the task count in the message processor configuration as follows:
| Code Block | ||
|---|---|---|
| ||
<parameter name="member.count">2</parameter> |
The above sample configuration ensures that there are two tasks running for the given message processor. These tasks can be running in one or more worker nodes in the clustered setup.
| Info |
|---|
This configuration does not guarantee that the message processor task will run in all worker nodes in the clustered setup. If the task count is less than the number of worker nodes in the cluster, the task will not run in some workers nodes. On the other hand, if the task count is larger than the worker nodes in the cluster, some worker nodes may run more than one task. |
All configuration changes done to the message processor will take effect after the next ESB instance/cluster restart.
Shutting down a member with one or more message processors
When a worker node where on which one or more worker nodes message processors are running run is shut down by an administrator or due to a failure, those message processors are scheduled on the other members of can be transferred to other worker nodes in the cluster.