The configuration restrictions in the subsections that follow apply only to Sun Cluster HA for WebSphere Message Broker.
Your data service configuration might not be supported if you do not observe these restrictions.
The Sun Cluster HA for WebSphere Message Broker data service can only be configured as a failover service.
Single or multiple instances of WebSphere Message Broker can be deployed in the cluster.
WebSphere Message Broker can be deployed in the global zone or a whole root zone. The See Restriction for multiple WebSphere Message Broker instances for more information about deploying in a zone.
The Sun Cluster HA for WebSphere Message Broker data service supports different versions of WebSphere Message Broker. Before proceeding with the installation of WebSphere Message Broker you must check that the Sun Cluster HA for WebSphere Message Broker data service has been verified against that version.
The WebSphere Message Broker files are the data files used by the broker in /var/mqsi. Within this document references will be made to the WebSphere Message Broker files which implies all of the contents of /var/mqsi, unless specified otherwise.
These WebSphere Message Broker files needs to be placed on shared storage as either a cluster file system or a highly available local file system. However, this placement will depend on how WebSphere Message Broker is being deployed, if a single or multiple instances are being deployed, and if that deployment will be in the global zone or zones.
Refer to Step 5 and Step 6 in How to Install and Configure WebSphere Message Broker for a more information.
WebSphere Message Broker requires WebSphere MQ and a database.
If you are installing WebSphere Business Integration Message Broker v5, the Sun Cluster HA for WebSphere Message Broker requires that the broker, queue manager and database are all registered within the same resource group. This implies that a remote database cannot be used for WebSphere Business Integration Message Broker v5.
This restriction is required because WebSphere Business Integration Message Broker v5 has very specific restart dependencies if the queue manager or database fails. More specifically, it is not possible for the cluster to manage the restart of a remote database that is outside of the cluster.
Table 2 describes the restart dependencies that the WebSphere Business Integration Message Broker v5 software has on additional software.
Table 2 WebSphere Business Integration Message Broker v5 restart dependencies
Failure |
Intended Action |
Actual Action |
---|---|---|
Broker |
Broker Start |
Sun Cluster Broker resource restarted |
Broker Queue Manager |
Broker Stop Broker Queue Manager Start Broker Start |
Sun Cluster Queue Manager resource restarted Sun Cluster Broker resource restarted |
Broker Database |
Broker Stop Broker Queue Manager Stop Broker Database Start Broker Queue Manager Start Broker Start |
Sun Cluster Database resource restarted Sun Cluster Queue Manager resource restarted Sun Cluster Broker resource restarted |
If you are installing WebSphere Message Broker v6, the restart dependency for the broker database listed in Table 2 is no longer required. This implies that a remote database can be used for WebSphere Message Broker v6. WebSphere Message Broker and WebSphere MQ are still required to be registered within the same resource group.
The broker database needs to be available for WebSphere Message Broker v6 to fully initialize. Therefore if you are deploying a remote broker database you must consider the availability of the broker database and the impact that can have on the broker if the broker database is not available.
The Sun Cluster HA for WebSphere Message Broker data service can support multiple WebSphere Message Broker instances, potentially with different versions.
If you intend to deploy multiple WebSphere Message Broker instances you will need to consider how you deploy WebSphere Message Broker in the global zone or whole root zones.
The purpose of the following discussion is to help you decide how to use the global zone or whole root zones to deploy multiple WebSphere Message Broker instances and then to determine what Nodelist entries are required.
The Nodelist entry is used when the resource group is defined using the clresourcegroup command. The Sun Cluster HA for WebSphere Message Broker must use the same resource group that is used for the WebSphere MQ and database resources.
You must therefore determine how the WebSphere Message Broker will be deployed in the cluster before the WebSphere MQ resource group is created so that you can specify the appropriate Nodelist entry.
Within these examples:
There are two nodes within the cluster, node1 and node2.
Both nodes have two zones named z1 and z2.
Each example listed simply shows the required Nodelist property value, via the -n parameter, which is used when creating a failover resource group.
Benefits and drawbacks are listed within each example as + and -.
Create a single failover resource group that will contain all the WebSphere Message Broker instances that will run in the global zones across node1 and node2.
# clresourcegroup create -n node1,node2 RG1 |
+ Only the global zone per node is required.
- Multiple WebSphere Message Broker instances do not have independent failover as they are all within the same failover resource group.
- Under normal operation, only one node of the cluster at any time is actively processing the WebSphere Message Broker workload.
Create multiple failover resource groups that will each contain one WebSphere Message Broker instance that will run in the global zones across node1 and node2.
# clresourcegroup create -n node1,node2 RG1 # clresourcegroup create -n node2,node1 RG2 |
+ Only the global zone per node is required.
+ Multiple WebSphere Message Broker instances have independent failover in separate failover resource groups.
+ Under normal operation, each node of the cluster is actively processing a WebSphere Message Broker workload, thereby utilizing each node of the cluster.
Create a single failover resource group that will contain all the WebSphere Message Broker instances that will run in the same zones across node1 and node2.
# clresourcegroup create -n node1:z1,node2:z1 RG1 |
+ Only one zone per node is required.
- Although all zones are booted, only one zone at any time is actively processing the WebSphere Message Broker workload.
- Multiple WebSphere Message Broker instances do not have independent failover as they are all within the same failover resource group.
- Multiple WebSphere Message Broker instances are not isolated within their own separate zones.
Create multiple zones, where each zone pair will contain just one WebSphere Message Broker instance that will run in the same zones across node1 and node2.
# clresourcegroup create -n node1:z1,node2:z1 RG1 # clresourcegroup create -n node2:z2,node1:z2 RG2 |
+ Multiple WebSphere Message Broker instances have independent failover in separate failover resource groups and separate zones.
+ All WebSphere Message Broker instances are isolated within their own separate zones.
- Each resource group requires a unique zone per node.