[KUDU-3578] De-flaking effort

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Labels: flaky-test
• Assignee: Bakai Ádám
• Reporter: Marton Greber
• Created: 17/May/24 12:30
• Updated: 17/May/24 13:53

Description

We have quite the number of flaky tests: http://dist-test.cloudera.org:8080/
This makes verifying patches a tedious process.
The idea is to use the above dashboard, to work though the flaky tests.
This should be an umbrella jira.
If you wish to contribute please create sub-tasks for each test de-flaking effort.
Testing remarks:
--stress_cpu_threads to simulate testing under load.

Thanks!

Sub-Tasks

• auto_leader_rebalancer-test: Open
• MasterAuthzITest.TestAuthzListTablesConcurrentRename fails from time to time: Open
• TestNewOpsDontGetScheduledDuringUnregister sometimes fail: Open
• AutoIncrementingItest.BootstrapNoWalsNoData fails sometimes: Open
• testReadDataFrameAtSnapshot(org.apache.kudu.spark.kudu.DefaultSourceTest) sometimes fails: Open
• AutoRebalancerTest.TestMaxMovesPerServer is flaky: Open
• TableKeyRangeTest.TestGetTableKeyRange fails from time to time: Resolved
• ClientTest.ClearCacheAndConcurrentWorkload fails from time to time in TSAN builds: Resolved
• SecurityITest.TestNonDefaultPrincipalMultipleMaster is flaky: Resolved
• update certs in test_certs.cc: Resolved
• scan_token-test is failing on ASAN: In Review
• tablet_copy_service-test is flaky in TSAN builds: Resolved
• org.apache.kudu.client.TestSecurity.testExternallyProvidedSubjectRefreshedExternally is flaky: Resolved
• EchoSubprocessTest.TestSubprocessMetricsOnError is flaky: Resolved
• ksck_->CheckMasterConsensus() is flaky in TSAN build: In Progress
• TSAN error in CountDownLatch destructor when using InternalMiniCluster: Open
• Threadpool check flakiness in ksck_remote-test during MiniMaster shutdown: In Progress