You can avoid SAP Web AS Java failure in the SAP central services by using the methods described below.
The message service controls the communications between the cluster managers of the Java instances. It distributes the cluster events and enables service-to-service communication. It also provides information for load balancing over the SAP Web dispatcher.
If the message service fails, this disturbs the communication between the cluster nodes. The message service is protected by a watchdog that restarts it in the event of failure. However, the only way to protect the message service from hardware failure is to use switchover software. After the switchover software restarts the message service, cluster communication continues.
The enqueue service provides a central lock table for logical locks on database objects in order to protect these objects from unsynchronized writes by the J2EE programs.
When the enqueue service fails, all open transactions have to be rolled back in order to prevent inconsistencies. If the enqueue service fails, the applications with requested locks get an exception that has to be handled by the application.
You can solve this problem in the following ways:
· Basic solution – switchover software
You need to at least protect the enqueue service with switchover software. This makes sure that the enqueue service can be restarted and is available again. However, even with switchover software, transactions open at the time of failure have to be rolled back.
· Advanced solution – standalone enqueue replication service
You can also use the standalone enqueue replication service, which maintains a replica of the lock table. This replica can be used in the event of failover to restore the lock table without the need to roll back transactions.
Note that there might be restrictions for the product availability of the standalone enqueue replication service.