Configure schedulers

Before the application can make use of the scheduler service, configure a scheduler using the administrative console, configuration service or scripting. Conceptually, a scheduler is similar to a data source in that specify various configuration attributes, including a JNDI name where the instance is bound. Once defined, an application using the Scheduler API or WASScheduler MBean can look up the scheduler object and call various methods to manage tasks.

The scheduler service is always enabled. In previous versions of the product, the scheduler service could be disabled using the administrative console or configuration service. Scheduler service configuration objects are present in the configuration service, but the enabled attribute is ignored.

To achieve high availability, we can configure a duplicate scheduler on each server in a cluster, or create a scheduler at the cluster scope. For example, each server containing a scheduler with the JNDI name sched/MyScheduler, with the same database configuration parameters (data source and table prefix) behaves as a single clustered scheduler. Each server in the scheduler cluster has a running scheduler instance, which increases the number of poll daemons and allows automatic failover. For more information on creating clusters for high availability, see the article, "WebSphere Enterprise Scheduler planning and administration guide."

Typically, create schedulers at the server or cluster scope. Scheduler poll daemons run in each server within the configured scope, which means that if we create a scheduler at the node or cell scope, the scheduler poll daemon can attempt to run tasks on any of the servers in the node or cell. If applications are not mapped uniformly over each server in that scope, the scheduler might not run tasks correctly. Because applications are mapped to servers and clusters, there is less chance for error and less competition between daemons to run tasks.

Depending on the preferred method of configuration, select one of the following steps to configure schedulers.

1. Configure schedulers using the administrative console.

2. Configure schedulers using JMX API (JMX).

Results

A scheduler is configured and ready to use.

• Configure scheduler default transaction isolation
  The scheduler uses read-committed transaction isolation, by default, when reading tasks using the get or find APIs on the com.ibm.websphere.scheduler.Scheduler interface and WASScheduler MBean. The default behavior for a scheduler can be changed to read-uncommitted, which allows the get and find methods to return the current or next state of the task in the database.

• Configure schedulers using the administrative console
  Schedulers can be created or configured using the administrative console.

• Configure schedulers using JMX
  Schedulers can be created or configured using the JMX API using one of several scripting languages or Java.

• Create a scheduler resource reference
  When you define schedulers in the server configuration, the object instance is bound into the global name space under the configured Java Naming Directory Interface (JNDI) name. Use a resource reference to avoid manually coding this JNDI name into the application. Using a resource reference allows administrators to map applications to the appropriate schedulers.

Related tasks

Develop and scheduling tasks

Related information:

WebSphere Enterprise Scheduler planning and administration guide