Fix race condition in scheduler and quartz's ConcurrentExecutionTest

[jaikiran]

This PR[1] ran into an unrelated failure[2] which appears to be a race condition in the ConcurrentExecutionTest. The PR here tries to prevent that race. There were 2 ConcurrentExecutionTest tests, one in quartz extension and one in scheduler. I updated them both given that they both are doing the same logic.

Having said that, there might be more going on here? The reason I say that is, this exception stacktrace of the failure:

2020-07-03T02:58:36.5610209Z [INFO] Running io.quarkus.quartz.test.ConcurrentExecutionTest
2020-07-03T02:58:37.2370796Z 2020-07-03 02:58:37,222 INFO  [io.quarkus] (main) Quarkus 999-SNAPSHOT on JVM started in 0.286s. 
2020-07-03T02:58:37.2371554Z 2020-07-03 02:58:37,230 INFO  [io.quarkus] (main) Profile test activated. 
2020-07-03T02:58:37.2372133Z 2020-07-03 02:58:37,230 INFO  [io.quarkus] (main) Installed features: [cdi, scheduler]
2020-07-03T02:58:39.2334866Z 2020-07-03 02:58:39,229 INFO  [io.quarkus] (main) Quarkus stopped in 0.003s
2020-07-03T02:58:39.2357907Z [ERROR] Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 2.67 s <<< FAILURE! - in io.quarkus.quartz.test.ConcurrentExecutionTest
2020-07-03T02:58:39.2359910Z [ERROR] io.quarkus.quartz.test.ConcurrentExecutionTest.testNonconcurrentExecution  Time elapsed: 1.987 s  <<< FAILURE!
2020-07-03T02:58:39.2364006Z org.opentest4j.AssertionFailedError: expected: <1> but was: <2>
2020-07-03T02:58:39.2368414Z 	at org.junit.jupiter.api.AssertionUtils.fail(AssertionUtils.java:55)
2020-07-03T02:58:39.2373236Z 	at org.junit.jupiter.api.AssertionUtils.failNotEqual(AssertionUtils.java:62)
2020-07-03T02:58:39.2377864Z 	at org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:150)
2020-07-03T02:58:39.2382881Z 	at org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:145)
2020-07-03T02:58:39.2387302Z 	at org.junit.jupiter.api.Assertions.assertEquals(Assertions.java:510)
2020-07-03T02:58:39.2394124Z 	at io.quarkus.quartz.test.ConcurrentExecutionTest.testNonconcurrentExecution(ConcurrentExecutionTest.java:29)

Notice that it says io.quarkus.quartz.test.ConcurrentExecutionTest.testNonconcurrentExecution Time elapsed: 1.987 s. I can't see how that test can take anything less than 2.x seconds (given the LATCH and the schedule on the concurrent() method). Perhaps the JUnit5 reporting is at fault or maybe there indeed is some other issue going on?

[1] #10442
[2] https://github.com/quarkusio/quarkus/pull/10442/checks?check_run_id=832968580

[mkouba]

Hi @jaikiran and thanks for spotting this. The test is indeed unreliable. Unfortunately, I did not see a better way to implement a proper test. I'll investigate this today and we should probably ignore these tests for now.

[machi1990]

LGTM

[mkouba]

@geoand You're too fast. I didn't even manage to look at the changes ;-). Anyway, here's the PR to disable the tests: #10445

[geoand]

Oh shoot! I saw your comment and I thought you had approved. My bad!

[mkouba]

No problem at all! :-)