HDDS-10079. Remove validation of test method in flaky-test-check

[adoroszlai]

What changes were proposed in this pull request?

If flaky-test-check is triggered for a specific method, it checks if the method exists in the test class. This does not work for test methods inherited from parent class.

Test method testValidateBlockLengthWithCommitKey does not exist in TestSecureOzoneRpcClient.java.Stopping!

https://github.com/adoroszlai/ozone/actions/runs/7431424440/job/20222159860#step:3:33

However, the same test method is accepted for the parent class, which cannot be run, since it's abstract.

I propose removing this safety check. If the specified method cannot be run, build output will show Tests run: 0.

https://issues.apache.org/jira/browse/HDDS-10079

How was this patch tested?

https://github.com/adoroszlai/ozone/actions/runs/7431422827

[sadanand48]

If the specified method cannot be run, build output will show Tests run: 0.

I guess the job would still pass though giving a false sense of success to user. I'm okay with current change given the parent class problem, but this would mean we need to check thoroughly when someone posts a successful job link. Should we fail if no tests are run?

[adoroszlai]

If the specified method cannot be run, build output will show Tests run: 0.

I guess the job would still pass though giving a false sense of success to user. I'm okay with current change given the parent class problem, but this would mean we need to check thoroughly when someone posts a successful job link. Should we fail if no tests are run?

You are right, the job would still pass. I've filed HDDS-10080 to fail fast (after first iteration) in junit.sh if no tests were run.

BTW, this can happen even prior to this patch, e.g. if I try to run TestOzoneRpcClientAbstract.

[adoroszlai]

Should we fail if no tests are run?

This is implemented in #6036.

[adoroszlai]

@sadanand48 with HDDS-10080 implemented, is this good to go?

[sadanand48]

@sadanand48 with HDDS-10080 implemented, is this good to go?

Yes , Thanks for the fix. LGTM

[adoroszlai]

Thanks @sadanand48 for the review.