[
https://issues.apache.org/jira/browse/FLINK-18522?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17173126#comment-17173126
]
Matthias commented on FLINK-18522:
----------------------------------
ZooKeepers TestServer is implemented in a way that it selects a random
available port. Unfortunately, the port selection is separated from the
instantiation of the ZooKeeper instances. It looks like there's a race
condition between multiple ZooKeeper instances selecting the same port due to
the tests running in parallel.
We investigated the option of changing the default port from {{-1}} to 0 to
trigger the port selection at startup time. Unfortunately, the [ZooKeeper
code|https://github.com/apache/zookeeper/blob/branch-3.4.14/zookeeper-server/src/main/java/org/apache/zookeeper/server/quorum/QuorumPeerConfig.java#L316]
does a validation on the configuration leading to an IllegalArgumentException.
We decided to close this issue for now. We might have to consider running the
tests in sequential order if we see this error coming up more often.
> ZKCheckpointIDCounterMultiServersTest.testRecoveredAfterConnectionLoss failed
> with "Address already in use"
> -----------------------------------------------------------------------------------------------------------
>
> Key: FLINK-18522
> URL: https://issues.apache.org/jira/browse/FLINK-18522
> Project: Flink
> Issue Type: Test
> Components: Runtime / Checkpointing, Runtime / Coordination, Tests
> Affects Versions: 1.10.1
> Reporter: Dian Fu
> Assignee: Matthias
> Priority: Major
> Labels: test-stability
>
> [https://travis-ci.org/github/apache/flink/jobs/705770513]
> {code}
> 15:09:34.674 [ERROR]
> testRecoveredAfterConnectionLoss(org.apache.flink.runtime.checkpoint.ZKCheckpointIDCounterMultiServersTest)
> Time elapsed: 5.74 s <<< ERROR!
> java.net.BindException: Address already in use
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
