Defining task schedules

You can use different combinations of parameters to define schedules for task executions.

Procedure

To define a new scheduled task:

Connect to the Db2 subsystem with sufficient authorization to call the ADMIN_TASK_ADD stored procedure.

The following task definitions show some common scheduling options.

To define	Do this
A task that executes only one time:	Set `max-invocations` to 1. Optionally, provide a value for the `begin-timestamp` parameter to control when execution happens. Leave other parameters null. For example, if `max-invocations` is set to 1 and `begin-timestamp` is set to 2008-05-27-06.30.0, the task executes at 6:30 AM on May 27, 2008. With this definition, the task executes one time. If `begin-timestamp` has been provided, execution happens as soon as permitted.
A regular repetitive execution:	Set `interval` to the number of minutes that you want to pass between the start of one execution and the start of the next execution. Optionally, provide values for the `max-invocations`, `begin-timestamp`, and `end-timestamp` parameters to limit execution. Leave other parameters null. For example, if `interval` is set to 5 and `begin-timestamp` is set to 2008-05-27-06.30.0, the task executes at 6:30 AM on May 27, 2008, then again at 6:35, 6:40, and so forth. With this definition, the task executes every `interval` minutes, so long as the previous execution has finished. If the previous execution is still in progress, the new execution is postponed `interval` minutes. Execution continues to be postponed until the running task completes.
An irregular repetitive execution:	Set `point-in-time` to a valid UNIX cron format string. The string specifies a set of times. Optionally, provide values for the `max-invocations`, `begin-timestamp` and `end-timestamp` parameters to limit execution. Leave other parameters null. For example, if `point-in-time` is set to 0 22 * * 1,5, the task executes at 10:00 PM each Monday and Friday. With this definition, the task executes at each time specified, so long as the previous execution has finished. If the previous execution is still in progress, the new execution is skipped. Subsequent executions continue to be skipped until the running task completes.
An execution that is triggered when another task completes:	Set `trigger-task-name` to the name of the triggering task. Optionally set `trigger-task-cond` and `trigger-task-code` to limit execution based on the result of the triggering task. The `trigger-task-cond` and `trigger-task-code` parameters must either both be null or both be non-null. Optionally, provide values for the `max-invocations`, `begin-timestamp` and `end-timestamp` parameters to limit execution. Leave other parameters null. For example, assume that a scheduled INSERT job has a task name of test_task. If `trigger-task-name` is test_task, `trigger-task-cond` is EQ, and `trigger-task-code` is 0, then this task executes when the INSERT job completes with a return code of 0. With this definition, the task executes at each time specified, so long as the previous execution has finished. If the previous execution is still in progress, the new execution is skipped. Subsequent executions continue to be skipped until the running task completes.
An execution that is triggered when Db2 starts:	Set `trigger-task-name` to DB2START. Optionally, provide values for the `max-invocations`, `begin-timestamp` and `end-timestamp` parameters to limit execution. Leave other parameters null. For example, if `trigger-task-name` is DB2START, `begin-timestamp` is 2008-01-01-00.00.0, and `end-timestamp` is 2009-01-01-00.00.0, the task executes each time that Db2 starts during 2008. With this definition, the task executes at each Db2 start, so long as the previous execution has finished. If the previous execution is still in progress, the new execution is skipped. Subsequent executions continue to be skipped until the running task completes.
An execution that is triggered when Db2 stops:	Set `trigger-task-name` to DB2STOP. Optionally, provide values for the `max-invocations`, `begin-timestamp` and `end-timestamp` parameters to limit execution. Leave other parameters null. With this definition, the task executes at each Db2 stop, so long as the previous execution has finished. If the previous execution is still in progress, the new execution is skipped. Subsequent executions continue to be skipped until the running task completes.