alirezazamani commented on a change in pull request #1142:
URL: https://github.com/apache/helix/pull/1142#discussion_r455406410



##########
File path: 
helix-core/src/test/java/org/apache/helix/integration/task/TestDeleteJobFromJobQueue.java
##########
@@ -44,10 +44,10 @@ public void testForceDeleteJobFromJobQueue() throws 
InterruptedException {
     // Create two jobs: job1 will complete fast, and job2 will be stuck in 
progress (taking a long
     // time to finish). The idea is to force-delete a stuck job (job2).
     JobConfig.Builder jobBuilder = 
JobConfig.Builder.fromMap(WorkflowGenerator.DEFAULT_JOB_CONFIG)
-        .setMaxAttemptsPerTask(1).setWorkflow(jobQueueName)
+        .setWorkflow(jobQueueName)

Review comment:
       I reverted this change for this test. However, we still need to increase 
number of attempts for some test that we stop the participant and start them 
again. 
   For example, in the 
TestForceDeleteWorkflow.testDeleteCompletedWorkflowForcefully:
   In super.beforeClass() we start the participant, however, once the test 
started, we stop those participants and start them again. I have observed that 
if we start the workflow very close to the time that we start the participant 
one by one (kind of like rolling upgrade), Master might jump several times 
between the nodes. Hence if we set maxNumberOfAttempts to 1, then the test will 
fail. Since we just started to respect maxNumberOfAttempts in this PR.
   
   About your second comment: Yeah the newly added test checks 
maxNumberOfAttempts for targetJobChange. For "rebalanceRunningTask", I checked 
other tests and I saw we this logic is already covered in 
"TestRebalanceRunningTask" extensively for different scenarios. So if rebalance 
running task is set and we need to drop the task in one instance and run in 
other one (Send INIT->RUNNING), then this is new scheduling and we are 
respecting max number of attempts. So rebalanceRunningTask is happening until 
we reach maxNumberOfAttempts. In other word, once a task reaches 
maxNumberOfAttempts, it won't be scheduled again. So no new INIT->RUNNING once 
we reach maxNumberOfAttempts.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to