[CASSANDRA-15672] Testsuite: org.apache.cassandra.repair.consistent.CoordinatorMessagingTest Tests run: 7, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 7.878 sec

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 4.0-alpha4, 4.0
• Component/s: Test/unit
• Labels: pull-request-available
• Bug Category: Correctness
• Severity: Normal
• Complexity: Normal
• Discovered By: User Report
• Platform: All
• Since Version: 4.0-alpha

Source Control Link

https://gitbox.apache.org/repos/asf?p=cassandra.git;a=commit;h=b4e640a96e76f8d4a45937b1312b64ddc1aeb8ac

Test and Documentation Plan

I ran CoordinatorMessagingTest several hundred times and it turned out that there is more than 1 flaky test. testMockedMessagingHappyPath and all tests that execute testMockedMessagingPrepareFailure fail from time to time.

In testMockedMessagingHappyPath we check whether sessionResult.isDone after each step of repair communication and assert that it will only be true after the final step. However, doing these assertions we cannot be sure that the final step hasn't been finished yet. Without 208a23 this test fails ~10 times during 500 executions.

There is a similar issue with testMockedMessagingPrepareFailure. We check that there are no failures yet right after starting coordinator execution. If we are unlucky and at least one of the participants that is meant to fail responds before we load proposeFailed, the assertion will fail. b3c761 removes the flaky assertion.

Additionally I included small javadoc fixes in a6e9d0.

• Assignee: Alex Sorokoumov Assign to me
• Reporter: Ekaterina Dimitrova

Authors

Alex Sorokoumov

Reviewers

Stefania Alborghetti

• Created: 27/Mar/20 17:52
• Updated: 21/Dec/20 08:08
• Resolved: 06/Apr/20 12:36

Description

The following failure was observed:

Testsuite: org.apache.cassandra.repair.consistent.CoordinatorMessagingTest Tests run: 7, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 7.878 sec
[junit-timeout]
[junit-timeout] Testcase: testMockedMessagingPrepareFailureP1(org.apache.cassandra.repair.consistent.CoordinatorMessagingTest): FAILED
[junit-timeout] null
[junit-timeout] junit.framework.AssertionFailedError
[junit-timeout] at org.apache.cassandra.repair.consistent.CoordinatorMessagingTest.testMockedMessagingPrepareFailure(CoordinatorMessagingTest.java:206)
[junit-timeout] at org.apache.cassandra.repair.consistent.CoordinatorMessagingTest.testMockedMessagingPrepareFailureP1(CoordinatorMessagingTest.java:154)
[junit-timeout]
[junit-timeout]

Seen on Java8

Issue Links

• links to: GitHub Pull Request #504