[FLINK-34560] JoinITCase seems to fail on a broader scale (MiniCluster issue?)

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 1.18.1
• Component/s: Tests
• Labels: github-actions test-stability
• Assignee: Unassigned
• Reporter: Matthias Pohl
• Created: 01/Mar/24 10:19
• Updated: 01/Mar/24 11:36
• Resolved: 01/Mar/24 11:36

Description

https://github.com/apache/flink/actions/runs/8105495458/job/22154140154#step:10:11906

The o.a.f.table.planner.runtime.stream.table.JoinITCase test methods share the same MiniCluster. Job 0d045f91b959c15f49ab18f0613aebd6 is submitted in test JoinITCase#testJoinWithMultipleKeys[StateBackend=ROCKSDB]. This is the last test that successfully ran. But the job gets reassigned to the cluster before the next test testInnerJoinWithExpressionPreds[stateBackend=ROCKSDB] submits its job e54a8aec0b89012d66a1cf7e4b3c283f. 0d045f91b959c15f49ab18f0613aebd6 occupies all the resources (4 slots) which won't become available for the subsequent test runs and make them fail.

Missing resources:
         Job e54a8aec0b89012d66a1cf7e4b3c283f
                ResourceRequirement{resourceProfile=ResourceProfile{UNKNOWN}, numberOfRequiredSlots=4}
Current resources:
        TaskManager 3be92028-41f6-4c4c-891b-57ec9d150bfd
                Available: ResourceProfile{cpuCores=0, taskHeapMemory=0 bytes, taskOffHeapMemory=0 bytes, managedMemory=0 bytes, networkMemory=0 bytes}

Issue Links

• duplicates: FLINK-34227 Job doesn't disconnect from ResourceManager