Spring Batch: Terminating the current running job

Question

I am having an issue in terminating the current running spring batch. I wrote

Set<Long> executions = jobOperator.getRunningExecutions("Job-Builder");
jobOperator.stop(longExecutions.iterator().next());`

in my code after going through the spring documentation.

The problem I am facing is at times the termination of the job is happening as expected and the other times the termination of job is not happening. In fact every time I call stop on joboperator it is updating the BATCH_JOB_EXECUTION table. When the termination happens successfully the status of the job is updating to STOPPED by killing the jobExecution in my batch process. The other times when it fails it is completing the rest of the different flows of the batch and updating the status to FAILED on BATCH_JOB_EXECUTION table.

But every time I call stop in the job operator I see a message in my console

2020-09-30 18:14:29.780 [http-nio-8081-exec-5] INFO  o.s.b.c.l.s.SimpleJobOperator:428        - Aborting job execution: JobExecution: id=33058, version=2, startTime=2020-09-30 18:14:25.79, endTime=null, lastUpdated=2020-09-30 18:14:28.9, status=STOPPING, exitStatus=exitCode=UNKNOWN;exitDescription=, job=[JobInstance: id=32922, version=0, Job=[Job-Builder]], jobParameters=[{date=1601504064263, time=1601504064262, extractType=false, JobId=1601504064262}]

My project has a series of flows and steps with in it.

Over all my batch process looks like this:
JobBuilderFactory has 3 flows
Each flow has a stepbuilder and two tasklets.
each stepbuilder has a partitioner and a chunk(size is 100) based itemReader, itemProcessor and itemWriter.

I am calling the stop method when I am executing the very first flow in my jobBuilderFactory. The over all process to complete takes about 30 mins. So, it has close to around 20-25 mins from the time I call the stop method and the chunk size is 100 with in each and every flow and I am dealing with more than 500k records.

So, my question is why is jobExecution stopping at times when called stop methos(which is what I wanted) and why it isn't able to stop the jobExecution the remaining times.

Thanks in advance

Answer 1

So, my question is why is jobExecution stopping at times when called stop methos(which is what I wanted) and why it isn't able to stop the jobExecution the remaining times.

It's not easy to figure out the reason for that from what you shared, but I can give you a couple of notes about stopping jobs:

• jobOperator.stop does not guarantee that the job stops, it only sends a stop signal to the job execution. From what you shared, you are not checking the returned boolean that indicates if the signal has been correctly sent or not, so you should be doing that first.
• You did not share your code, but you need to use StoppableTasklet instead of Tasklet to make sure the stop signal is correctly sent to your steps.