Trying to fix test failures caused by Testcontainers running in Docker on Jenkins. The Cassandra Docker image/container is forked inside Docker running in Jenkins CI and is currently throwing:
Caused by: org.testcontainers.containers.ContainerLaunchException: Timed out waiting for container port to open (172.17.0.1 ports: [49401] should be listening)
at org.testcontainers.containers.wait.strategy.HostPortWaitStrategy.waitUntilReady(HostPortWaitStrategy.java:49)
at org.testcontainers.containers.wait.strategy.AbstractWaitStrategy.waitUntilReady(AbstractWaitStrategy.java:35)
at org.testcontainers.containers.GenericContainer.waitUntilContainerStarted(GenericContainer.java:892)
at org.testcontainers.containers.GenericContainer.tryStart(GenericContainer.java:440)
... 85 more
In the Testcontainers project, I found:
https://github.com/testcontainers/testcontainers-java/issues/4360
And specifically:
https://github.com/testcontainers/testcontainers-java/issues/4360#issuecomment-892481828
The Jenkins servers (worker nodes) were recently upgraded to:
docker-ce (5:20.10.8~3-0~ubuntu-bionic)
containerd.io (1.4.9-1)
Test and assert client Region DataPolicy configuration (PROXY).
Test and assert Environment configuration does not set the 'spring.data.gemfire.cache.client.region.shortcut' property.
Add preconditions for a Kubernetes runtime environment and logging set to INFO.
Test and assert client Region DataPolicy configuration (PROXY).
Test and assert Environment configuration does not set the 'spring.data.gemfire.cache.client.region.shortcut' property.
Add precoditions for a CloudFoundry runtime environment and logging not set to INFO.
Remove all logging from the overridden matches(..) method.
Override the configureTopology(..) method to do nothing.
Override the logConnectedRuntimeEnvironment(..) and logUnconnectedRuntimeEnvironment(..) to log appropriate messages for cloud runtime environments.
Refactor the StandaloneClusterAvailableCondition match to only act when not in a cloud runtime envrionment.
Remove logging from configureTopology(..) and add call to logRuntimeEnvironmet(..) from doMatch(..).
Add ClusterAwareCondition.set(:Boolean) class method to allow Condition subclasses to set the clusterAvailable state; refactor reset() to call set(null).
Edit Javadoc.
Using 'strictMatch' enables fail-fast behavior so that users can configure their Spring Boot, Apache Geode ClientCache applications to fail on startup if no cluster is available across any environment.
Additionally, enhanced the log output to give users better information about the runtime environment and whether an Apache Geode-based cluster was found and available.
Resolves gh-99.
Applied Spring's @NonNull and @Nullable annotations to the API.
Added a NotKubernetesEnvironmentCondition class to the ClusterNotAvailableConfiguration class.
Added a @Conditional on the NotKubernetesEnvironmentCondition class annotated IsNotKubernetesEnvironmentCondition class to the ClusterNotAvailableConfiguration.AllClusterNotAvailableConditions class.
Removed all @ConditionalOnCloudPlatform annotated Condition classes in the ClusterAvailableConfiguration.AnyClusterAvailableCondition class, for both CloudFoundry and Kubernetes.
Edit Javadoc.
Resolves gh-99.
Specifically, the listener state will track:
* The number of processEvents(:List<AsyncEvent>) method invocations.
* Whether the processEvents(..) method has ever been invoked.
* And, whether the processEvents(..) method has been invoked since the last check.
Introduces a protected doProcessEvents(:List<AsyncEvent>) method and changes processEvents(:List<AsyncEvent>) method to final.
This fixes a bug that manifests through GemFire/Geode's Management infrastructure, specifically with the MBean's ObjectName used to managed the AEQs across the cluster.
