Planning the HA for WebSphere MQ Installation and Configuration

Language:

This section contains the information you need to plan your HA for WebSphere MQ installation and configuration.

Configuration Restrictions

The configuration restrictions in the subsections that follow apply only to HA for WebSphere MQ.

Caution - Your data service configuration might not be supported if you do not observe these restrictions.

Restriction for the Supported Configurations of HA for WebSphere MQ

The HA for WebSphere MQ data service can only be configured as a failover service.

Single or multiple queue managers can be deployed in the cluster.

Multiple installations are not currently supported by HA for WebSphere MQ. Only a primary installation is supported on /opt/mqm. After installing IBM WebSphere v8.0 or later, you must also execute the WebSphere MQ setmqinst script on each of the cluster node as follows:

# /opt/mqm/bin/setmqinst -i -p /opt/mqm

IBM MQ can be deployed in either the global zone or the zone cluster. See Restriction for Multiple IBM MQ Instances for more information.

Restriction for the Location of IBM MQ Files

The IBM MQ files are where the queue manager data files /var/mqm/qmgr/queue-manager and /var/mqm/log/queue-manager are stored.

These IBM MQ files needs to be placed on shared storage as either a cluster file system or a highly available local file system.

Restriction for Multiple IBM MQ Instances

The HA for WebSphere MQ data service can support multiple IBM MQ instances, potentially with different versions.

If you intend to deploy multiple IBM MQ instances with different versions you will need to consider deploying IBM MQ in separate zone clusters.

Configuration Requirements

The configuration requirements in this section apply only to HA for WebSphere MQ.

Caution - If your data service configuration does not conform to these requirements, the data service configuration might not be supported.

Determine Which Oracle Solaris Zone IBM MQ Will Use

Oracle Solaris Zones provides a means of creating virtualized operating system environments within an instance of the Oracle Solaris 11 OS. Oracle Solaris Zones allow one or more applications to run in isolation from other activity on your system. For complete information about installing and configuring Oracle Solaris Zones, see Creating and Using Oracle Solaris Zones .

You must determine which zone IBM MQ will run in. IBM MQ can run within a global zone or a zone cluster.

Requirements If Multiple IBM MQ Instances Are Deployed on Cluster File Systems

If a cluster file system is being used for the IBM MQ files, it is possible to manually start the queue manager on one node of the cluster and at the same time to also manually start the same queue manager on another node of the cluster.

Note - Although it is possible, you should not attempt this as doing so will cause severe damage to the IBM MQ files.

Although it is expected that no-one will manually start the same queue manager on separate nodes of the cluster at the same time the HA for WebSphere MQ provides a mechanism to prevent someone from doing so, albeit by mistake.

To prevent against this happening you must implement one of the following two solutions.

Use a highly available local file system for the IBM MQ files.
This is the recommended approach as the IBM MQ files would be mounted only on one node of the cluster at a time. This then limits starting the queue manager on only one node of the cluster at a time.
Create a symbolic link for /opt/mqm/bin/strmqm and /opt/mqm/bin/endmqm to /opt/SUNWscmqs/mgr/bin/check-start.
/opt/SUNWscmqs/mgr/bin/check-start provides a mechanism to prevent manually starting or stopping the queue manager, by verifying that the start or stop is being attempted by the HA for WebSphere MQ data service.
/opt/SUNWscmqs/mgr/bin/check-start will report the following error if an attempt to manually start or stop the queue manager.
```
$ strmqm qmgr1
$ Request to run </usr/bin/strmqm qmgr1> within Oracle Solaris Cluster has been refused
```
If a cluster file system is used for the IBM MQ files, you must create a symbolic link for strmqm and endmqm to /opt/SUNWscmqs/mgr/bin/check-start and inform the HA for WebSphere MQ data service of this change.
To do this, you must perform the following on each node of the cluster.
```
# cd /opt/mqm/bin
#
# mv strmqm strmqm_sc3
# mv endmqm endmqm_sc3
#
# ln -s /opt/SUNWscmqs/mgr/bin/check-start strmqm
# ln -s /opt/SUNWscmqs/mgr/bin/check-start endmqm
```
After renaming strmqm and endmqm you must use these new program names (strmqm_sc3 and endmqm_sc3) for the START_CMD and STOP_CMD variables when you edit the /opt/SUNWscmqs/mgr/util/mgr_config file in Step 7 in How to Register and Configure HA for WebSphere MQ.
Note - If you implement this workaround, then you must back it out whenever you need to apply any maintenance to IBM MQ. Afterwards, you must again apply this workaround.
Instead the recommended approach is to use a highly available local file system for the IBM MQ files.