[
https://issues.apache.org/jira/browse/HDDS-16133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siyao Meng updated HDDS-16133:
------------------------------
Attachment: HDDS-16133.001.patch
> Intermittent timeout in
> TestReconTasks#testContainerHealthTaskDetectsEmptyMissingWhenAllReplicasLost
> ----------------------------------------------------------------------------------------------------
>
> Key: HDDS-16133
> URL: https://issues.apache.org/jira/browse/HDDS-16133
> Project: Apache Ozone
> Issue Type: Sub-task
> Components: Ozone Recon, test
> Reporter: Siyao Meng
> Priority: Major
> Attachments: HDDS-16133.001.patch
>
>
> {code}
> [ERROR] Tests run: 6, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 212.2
> s <<< FAILURE! -- in org.apache.hadoop.ozone.recon.TestReconTasks
> [ERROR]
> org.apache.hadoop.ozone.recon.TestReconTasks.testContainerHealthTaskDetectsEmptyMissingWhenAllReplicasLost
> -- Time elapsed: 59.98 s <<< ERROR!
> java.util.concurrent.TimeoutException: timeout: after 20000 millis
> at
> org.apache.ozone.test.LambdaTestUtils$GenerateTimeout.evaluate(LambdaTestUtils.java:258)
> at org.apache.ozone.test.LambdaTestUtils.await(LambdaTestUtils.java:144)
> at
> org.apache.hadoop.ozone.recon.TestReconTasks.testContainerHealthTaskDetectsEmptyMissingWhenAllReplicasLost(TestReconTasks.java:387)
> {code}
> The test writes a real block into the container, then waits for Recon to
> report it as EMPTY_MISSING (and not MISSING) after all replicas are lost.
> Recon classifies EMPTY_MISSING via {{getNumberOfKeys() == 0}}; once the
> datanode container report propagates a non-zero key count, the container is
> classified MISSING, so the awaited EMPTY_MISSING state can be missed. In the
> logs every scan reports {{Stored 1 MISSING, 0 EMPTY_MISSING}} until the 20s
> wait expires, so more time does not help. The sibling
> {{testContainerHealthTaskDetectsMissingForContainerWithKeys}} is
> deterministic because it sets the key count explicitly; this test could do
> the same (force key count to 0 after the datanode is stopped).
> Seen in:
> -
> https://github.com/adoroszlai/ozone-build-results/blob/master/2026/08/02/51843/integration-recon/hadoop-ozone/integration-test-recon/org.apache.hadoop.ozone.recon.TestReconTasks.txt
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
