Martijn Visser created FLINK-41046:
--------------------------------------
Summary: HBaseConnectorITCase.testTableSinkDisabledBufferFlush
expected [+I[1, 6]] but was [+I[1, 1]]
Key: FLINK-41046
URL: https://issues.apache.org/jira/browse/FLINK-41046
Project: Flink
Issue Type: Bug
Components: Connectors / HBase
Affects Versions: hbase-4.0.0
Reporter: Martijn Visser
HBaseConnectorITCase.testTableSinkDisabledBufferFlush fails intermittently on
the v4.0 branch in the weekly runs (Flink 1.19.3). Most recent, weekly
2026-10-18, JDK 8, and the re-run of that job on 2026-10-19:
https://github.com/apache/flink-connector-hbase/actions/runs/39081906348/job/117540414198
https://github.com/apache/flink-connector-hbase/actions/runs/39081906348/job/118003776981
{code}
[ERROR]
org.apache.flink.connector.hbase2.HBaseConnectorITCase.testTableSinkDisabledBufferFlush
Time elapsed: 0.645 s <<< FAILURE!
org.junit.ComparisonFailure:
Different elements in arrays: expected 1 elements and received 1
expected: [+I[1, 6]]
received: [+I[1, 1]] expected:<+I[1, [6]]> but was:<+I[1, [1]]>
at
org.apache.flink.connector.hbase2.HBaseConnectorITCase.testTableSinkDisabledBufferFlush(HBaseConnectorITCase.java:725)
{code}
The same failure on JDK 11 on 2026-09-06 and 2026-08-02:
https://github.com/apache/flink-connector-hbase/actions/runs/34000731246/job/101399010108
https://github.com/apache/flink-connector-hbase/actions/runs/30725174553/job/91435387881
That is 3 of the 8 weeklies since 2026-08-02 in which a v4.0 job ran to the
end. v4.0 has not changed since November 2024 and contains the FLINK-36188 fix
this test was added for. On main the test has been disabled since the async
sink migration in FLINK-35280. Because the weekly matrix is not fail-fast:
false, this failure also cancels the main cells.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
