Start of change

High availability clustering

This section gives information on considerations for migrating high availability (HA) clusters.

Start of changeTo update an HA cluster, plan the order of migration so that a system running WebSphere MQ Version 5.3 code can never access the files for a queue manager that is already running with Version 6.0 code. Implementing this rolling migration strategy to minimize total downtime will probably require you to disable failover between some systems in the cluster during the upgrade process.End of change

Because channels are objects in WebSphere® MQ Version 6.0, the channels and clntconn directories must be replicated between nodes for channel definitions. (The client channel table file is unchanged.)

On Windows® HA systems, the certificate stores must be migrated to GSKit on all nodes which host a WebSphere MQ Version 6.0 queue manager.



End of change