[
https://issues.apache.org/jira/browse/KAFKA-8837?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anna Povzner updated KAFKA-8837:
--------------------------------
Description:
@After method in KafkaMetricReporterClusterIdTest calls
`TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)` before it calls
tearDown on ZooKeeperTestHarness (which shut downs ZK and zk client). If
verifyNonDaemonThreadsStatus asserts, the rest of the resources will not get
cleaned up.
Second, this and several other tests do not actually create KafkaServer with
the thread prefix, so this verification does not actually do anything (because
threads are not prefixed with the test name).
We should move `TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)`
to the end of `tearDown()`; and every test that calls
verifyNonDaemonThreadsStatus should actually create KafkaServer with their test
classname prefix.
The reason I started looking into this is because one of the recent builds
failed with many tests failing with "found unexpected threads...". So, fixing
the above may at least help us to find the root cause, because this will
increase the chance that a test failing with this error is the same test that
leaks the thread.
[https://builds.apache.org/job/kafka-trunk-jdk11/774/#showFailuresLink]
was:
@After method in KafkaMetricReporterClusterIdTest calls
`TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)` before it calls
tearDown on ZooKeeperTestHarness (which shut downs ZK and zk client). If
verifyNonDaemonThreadsStatus asserts, the rest of the resources will not get
cleaned up.
We should move `TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)`
to the end of `tearDown()`. However, would also be good to prevent people using
this method in tear down similarly in the future. Maybe just adding a comment
would help here.
Summary: Several integration tests incorrectly use
verifyNonDaemonThreadsStatus (was: KafkaMetricReporterClusterIdTest may not
shutdown ZooKeeperTestHarness)
> Several integration tests incorrectly use verifyNonDaemonThreadsStatus
> ----------------------------------------------------------------------
>
> Key: KAFKA-8837
> URL: https://issues.apache.org/jira/browse/KAFKA-8837
> Project: Kafka
> Issue Type: Bug
> Components: core
> Reporter: Anna Povzner
> Assignee: Anastasia Vela
> Priority: Major
>
> @After method in KafkaMetricReporterClusterIdTest calls
> `TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)` before it
> calls tearDown on ZooKeeperTestHarness (which shut downs ZK and zk client).
> If verifyNonDaemonThreadsStatus asserts, the rest of the resources will not
> get cleaned up.
> Second, this and several other tests do not actually create KafkaServer with
> the thread prefix, so this verification does not actually do anything
> (because threads are not prefixed with the test name).
> We should move
> `TestUtils.verifyNonDaemonThreadsStatus(this.getClass.getName)` to the end of
> `tearDown()`; and every test that calls verifyNonDaemonThreadsStatus should
> actually create KafkaServer with their test classname prefix.
> The reason I started looking into this is because one of the recent builds
> failed with many tests failing with "found unexpected threads...". So, fixing
> the above may at least help us to find the root cause, because this will
> increase the chance that a test failing with this error is the same test that
> leaks the thread.
> [https://builds.apache.org/job/kafka-trunk-jdk11/774/#showFailuresLink]
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
