[CASSANDRA-19376] CompactionStrategyManagerTest.testAutomaticUpgradeConcurrency is flaky

• Type: Bug
• Status: Triage Needed
• Priority: Normal
• Resolution: Unresolved
• Fix Version/s: 5.x
• Component/s: Test/unit
• Bug Category: Correctness - Test Failure
• Platform: All
• Assignee: Unassigned
• Reporter: Andres de la Peña
• Created: 07/Feb/24 11:56
• Updated: 05/Dec/24 19:58

Description

The unit test CompactionStrategyManagerTest.testAutomaticUpgradeConcurrency fails occasionally at least in trunk:

unit.framework.AssertionFailedError: expected:<1> but was:<0>
	at org.apache.cassandra.db.compaction.CompactionStrategyManagerTest.testAutomaticUpgradeConcurrency(CompactionStrategyManagerTest.java:227)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
	at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

https://app.circleci.com/pipelines/github/adelapena/cassandra/3427/workflows/126307ce-d465-4736-a71e-82e0b8749598

The failure can be reproduced with:

.circleci/generate.sh -p \
  -e REPEATED_UTESTS=org.apache.cassandra.db.compaction.CompactionStrategyManagerTest \
  -e REPEATED_UTESTS_COUNT=1000

I have only tested trunk. I haven't seen it on Butler.

Observed flakiness is ~0.2% for all jobs except for tries, which ramps up to ~2.5%.

Issue Links

• Discovered while testing: CASSANDRA-19336 Repair causes out of memory