Configuring high availability and workload sharing of service integration

You can configure high availability and workload sharing of service integration without using messaging engine policy assistance.

Before you begin

Ensure that you want to use the following procedure. As an alternative, you can configure high availability and workload sharing of service integration by using messaging engine policy assistance when you add a server cluster to a bus. You create messaging engines and their associated policies as part of the procedure, by using the appropriate predefined messaging engine policy type. Alternatively, you can use a custom policy type and configure the messaging engine policy as you require, and the relevant core group policies and match criteria are created automatically.

About this task

When you set up a service integration environment, you create bus members, either servers or clusters, that run messaging engines. For high availability, where the messaging engine can fail over, or workload sharing, where multiple messaging engines share the load on a destination, you need to create a cluster bus member and configure high availability and workload sharing characteristics of the messaging engines.

If you do not require high availability or workload sharing, you can use a simple configuration and create a server bus member. You do not need the steps described in this topic.

The high availability and workload sharing characteristics of the messaging engines in the cluster are set by core group policies.

To see the policies that are configured in your system, you can use the administrative console to open the Policies page. In the navigation pane, click Servers -> Core groups -> Core group settings -> core_group_name -> [Additional Properties] Policies.

One of the available policies is the default service integration policy, Default SIBus Policy, which is the policy that a messaging engine uses unless you configure the system so that the engine uses another policy. The default policy is sufficient for many purposes and you might not need to alter the policy configuration. It is not advisable to alter the default service integration policy, because those changes will affect all messaging engines that the policy manages. Therefore, it is better to create and configure one or more new specific policies.

1. Optional: Create a cluster, if it is not created already.
2. Add the cluster to the service integration bus.

   A single messaging engine that uses the default service integration policy is created automatically.

   For high availability without workload sharing, you can use this configuration and do not need to change it further. If you want to configure the messaging engine behavior further, for example to specify preferred servers for the messaging engine, or enable the messaging engine to fail back, complete step 3.

3. Optional: For high availability when you want to configure the messaging engine behavior, create and configure a policy for the messaging engine.
   1. Create a policy with the One of N type.
   2. Configure a policy for the messaging engine.
4. Optional: For workload sharing without high availability, use the following steps:
   1. Add as many messaging engines as you require to the cluster. Typically, a workload sharing configuration has one messaging engine for each server in the cluster
   2. Create a policy for each messaging engine in the cluster. Create policies with the Static type.
   3. Configure a policy for each messaging engine.
5. Optional: For workload sharing with high availability, use the following steps:
   1. Add as many messaging engines as you require to the cluster. Typically, a workload sharing configuration has one messaging engine for each server in the cluster.
   2. Create a policy for each messaging engine in the cluster. Create policies with the One of N type.
   3. Configure a policy for each messaging engine.
6. Optional: To use an external high availability (HA) framework to manage high availability or workload sharing behavior, use the following steps:
   1. If you require workload sharing, add as many messaging engines as you require to the cluster. Typically, a workload sharing configuration has one messaging engine for each server in the cluster.
   2. Create one policy for the messaging engines in the cluster. Create a policy with the No operation type.
   3. Configure a core group policy for the messaging engine.

What to do next

If you created a high availability configuration for service integration, you might also want to configure high availability for the transaction service.

If you created a workload sharing configuration, you might want to deploy a queue destination to the cluster, so that the queue is partitioned across the set of messaging engines.