JSR352 Partition Mapper- Restart of Failed Parallel Partitions Wont Work Properly (Doc ID 2977425.1)

Last updated on OCTOBER 08, 2024

Applies to:

Symptoms

For the client, using JSR-352 batch application (along with WebLogic 12.2.1.4) for the batch processes. When implementing the parallel processing using the PartitionMapper option available in JSR-352, in a successful flow, this works fine. However, in case of failures in one or more partitions, then during restarts , noticed some unexpected behavior. The restart,of the failed partition thread, fails with an error that “Execution Id is not the most recent execution”. After the first failure, when the job is restarted the 1st time, it creates a new execution id for the child partitions. Sometimes the value of this execution id is lesser than that of the previous execution id and sometimes it is greater. So in case the first restart also fails, when the job is restarted the 2nd time, it fails when the first restart execution id is lesser than that of the first original submission id. So once this error occurs, not able to restart that job. This happens randomly. Sometimes, it works properly (i.e, the first restart execution id is greater than the first job submission execution Id) and sometimes it fails.

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.