Handling Exceptions

Errors in Batch Job Stream

A batch job submitted as part of a Batch Job Stream can run into error with one of its threads being in error. It could be because of business scenario or environment issue etc. In such a case, the corresponding Batch Job Stream stops execution and its step ends in failed state. The cause for the failure is noted on the Additional Details column on the stream step. You can fix the issue and restart the stream via the multi-select Restart Stream option on the batch job stream operations completed query option. Restarting a stream will result in the Batch Job Stream continuing execution from the step that failed and finish through to end.

Note:

Once a Batch Job Stream fails, its future scheduled runs no longer execute until the failed run is addressed.

Batch Job Stream stuck in running status

A Batch Job Stream can be stuck in Running status as one of its steps is in running status for too long. In this situation there are various options.

Cancelling a batch job stream: The corresponding batch job submission is stuck in pending status as the thread pool name set on the GLOBAL program options via a previous stream step is invalid. In such case, the running Batch Job Stream can be cancelled via the multi-select Cancel option. The thread pool name can then be set to a valid value and the Batch Job Stream can be submitted manually via the Run Manually option on the Batch Job Stream Definition or wait until it’s scheduled to run next.

Interrupting a batch job stream: The corresponding batch job submission is stuck in started status while the respective batch run or batch threads have ended. In this case, the running batch job stream can be interrupted via the multi-select Interrupt option. This sets the batch job stream and respective steps in “Failed” state. The Restart Stream multi-select option can then be used to re-start the stream to continue execution from the step that failed and complete processing of stream.

Renew Publish a batch job stream: If either of the above options don’t work, then Renew Publish multi-select option can be used. This option re-publishes the stream in DBMS scheduler with a different name(same name suffixed with version number and incremented for every renew click) so further scheduling can continue. For example, if the stream name is HOURLY, clicking on renew publish would re-publish stream as HOURLY2 in DBMS scheduler.

Renew Publish a batch job stream with sub-streams: When renew publishing a batch job stream with sub-streams in it, the respective sub-streams are not re-published. Typically, if the main batch job stream is stuck in Running status, it does not necessarily mean the sub-streams would be stuck in Running status too. It is recommended to check if the sub-streams are also stuck in Running status and if so, “renew publish” them first before main stream. If the sub-streams are not stuck in Running status, then the sub-streams can be left as is and does not require “renew publish” action.

Note: “Renew Publish” option should only be used when a stream is stuck/stranded in “Running” state. Additionally, this option should be used only after using the Interrupt option and it does not work and job stream continues to be stuck in “Running” state.

A job stream that is re-published via Renew Publish option has it’s name suffixed with incremental numbers in DBMS Scheduler. For example, if the stream name is HOURLY, clicking on renew publish would re-publish stream as HOURLY2 in DBMS scheduler. It is important to take note of this to make sure there is no other stream in the system with stream name suffixed with numbers such as HOURLY2 in given example as then the “Renew Publish” option would fail.