Sergey Soldatov created HDDS-16080:
--------------------------------------
Summary: TestClientRetryContainerStateMachineFailures takes 12
minutes due to a 30s heartbeat override
Key: HDDS-16080
URL: https://issues.apache.org/jira/browse/HDDS-16080
Project: Apache Ozone
Issue Type: Test
Components: test
Affects Versions: 2.2.0
Reporter: Sergey Soldatov
Assignee: Sergey Soldatov
On my box, TestClientRetryContainerStateMachineFailures takes over 700s for 5
tests, of which ~73% is MiniOzoneCluster setup: ~100s to build a 3-datanode
cluster. Method init() sets hdds.heartbeat.interval to 30s, overriding the 1s
that hadoop-ozone/integration-test/src/test/resources/ozone-site.xml pins for
the whole module. Cluster readiness then advances in 30-second rounds, and
init() is @BeforeEach, so it is paid 5 times. Setting it to 1s brings the class
to less than 300s (more than 2x times faster).
[~sumitagrawl], the 30s looks like a deliberate override of the module default
rather than an omission. Was there an intent behind it that these runs would
not surface? I haven't found any regression using a 200 ms or 1s heartbeat.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
