[
https://issues.apache.org/jira/browse/CASSANDRA-16758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17369694#comment-17369694
]
Andres de la Peña commented on CASSANDRA-16758:
-----------------------------------------------
Using the {{requireNetworkWithoutDriver}} method suggested in CASSANDRA-16663
in 4.0.0 definitely solves all the failures but the one in
{{testQueryUpdatesConcurrentMetricsUpdate}}, as it can be seen in [this
multiplexer
run|https://app.circleci.com/pipelines/github/adelapena/cassandra/614/workflows/974352fc-a2d7-453e-ae07-002280b9910e/jobs/6020].
Regarding the failure in {{testQueryUpdatesConcurrentMetricsUpdate}}, I think
that the creation of the startup message of {{SimpleClient}} is messing with
the metrics. We can simply use {{spinAssertEquals}}/{{Awaitility}} to wait
until the moment the startup message finishes and the client metric goes back
to zero, as it's done [this
way|https://github.com/adelapena/cassandra/commit/c0caca66eae1dd388a15cf415ffc25fc37190d2d].
With that change the test survives [5000
runs|https://app.circleci.com/pipelines/github/adelapena/cassandra/623/workflows/b69efc53-0e20-4466-a90a-78f162dde54b/jobs/6097].
> Flaky ClientResourceLimitsTest
> ------------------------------
>
> Key: CASSANDRA-16758
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16758
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Berenguer Blasi
> Assignee: Berenguer Blasi
> Priority: Normal
> Fix For: 4.0-rc2, 4.0, 4.0.x
>
>
> Flaky
> [ClientResourceLimitsTest|https://ci-cassandra.apache.org/job/Cassandra-4.0.0/35/testReport/junit/org.apache.cassandra.transport/ClientResourceLimitsTest/testBackPressureWhenEndpointLimitExceeded_cdc/]
> {noformat}
> Error Message
> Timed out after waiting 5 seconds for paused connections metric to increment
> due to backpressure expected:<11> but was:<24>
> Stacktrace
> junit.framework.AssertionFailedError: Timed out after waiting 5 seconds for
> paused connections metric to increment due to backpressure expected:<11> but
> was:<24>
> at org.apache.cassandra.Util.spinAssertEquals(Util.java:610)
> at
> org.apache.cassandra.transport.ClientResourceLimitsTest.backPressureTest(ClientResourceLimitsTest.java:218)
> at
> org.apache.cassandra.transport.ClientResourceLimitsTest.testBackPressureWhenEndpointLimitExceeded(ClientResourceLimitsTest.java:179)
> Standard Output
> INFO [main] 2021-06-22 01:56:41,489 YamlConfigurationLoader.java:93 -
> Configuration location:
> file:////home/cassandra/cassandra/build/test/cassandra.cdc.yaml
> DEBUG [main] 2021-06-22 01:56:41,491 YamlConfigurationLoader.java:112 -
> Loading settings from
> file:////home/cassandra/cassandra/build/test/cassandra.cdc.yaml
> DEBUG [main] 2021-06-22 01:56:41,551 InternalLoggerFactory.java:63 - Using
> SLF4J as the default logging framework
> DEBUG [main] 2021-06-22 01:56:41,565 PlatformDependent0.java:417 - -D
> ...[truncated 765272 chars]...
> org.apache.cassandra.transport.SimpleClient.execute(SimpleClient.java:275)
> at
> org.apache.cassandra.transport.ClientResourceLimitsTest.lambda$testQueryUpdatesConcurrentMetricsUpdate$11(ClientResourceLimitsTest.java:352)
> at java.lang.Thread.run(Thread.java:748)
> DEBUG [Test Cluster-nio-worker-0] 2021-06-22 01:57:08,353 Slf4JLogger.java:81
> - Freed 24 thread-local buffer(s) from thread: Test Cluster-nio-worker-0
> INFO [main] 2021-06-22 01:57:08,364 Server.java:171 - Stop listening for CQL
> clients
> {noformat}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
