|
|
A general discussion of configuring the Transaction Manager failover with the No Operation policy is provided in 9.6, Transaction Manager failover with No Operation policy. In this section we discuss how to configure VCS to correctly failover the Transaction Manager with the No Operation (NoOP) recovery policy.
In order to describe the configuration we refer to the example defined in Transaction Manager with No Operation policy scenario. In our setup, we have two WAS nodes. One node is the system named Bottechia and the other is the system named SystemB. Each node hosts a single appserver: TradeServer1 on Bottechia and TradeServer2 on SystemB. Both appservers are cluster members of TradeCluster.
For this scenario, we need to define the following service groups in VCS:
This group will be configured as parallel (Active/Active) and will start, stop, and monitor the Node Agent and the cluster members.
This group will be configured as failover (Active/Passive) and will activate, deactivate and monitor the TM on system Bottechia (primary) and failover to system SystemB (secondary).
This group will be configured as failover (Active/Passive) and will activate, deactivate and monitor the TM on system SystemB (primary) and failover to system Bottechia (secondary).
We also need to configure links (dependencies) between each group. The TM activate groups are linked to the WAS servers group. This means that the TM activate groups will not start until the WAS servers group completes starting. The TM activate groups are not linked to each other.
