[FLINK-10401] Port ProcessFailureCancelingITCase to new code base #6749
Conversation
| @@ -130,6 +131,13 @@ public T getDispatcher() { | |||
| } | |||
| } | |||
|
|
|||
| @VisibleForTesting | |||
| public WebMonitorEndpoint<?> getWebMonitorEndpoint() { | |||
There was a problem hiding this comment.
Coming back to my longer comment on #6749, I think this is only required because the collaborator is lazily created in a lifecycle method. I wonder if we could not have a building object the will create all collaborators at the place where currently startComponent is called because clusterComponent = createClusterComponent(configuration) and clusterComponent.startComponent(...) are even called in a straight sequence. Tests could then simply pass WebMonitorEndpoint into the component.
There was a problem hiding this comment.
Yes you're right. I've incorporated your feedback from #6743 and with that it should no longer be necessary to have this getter.
Introduce a ClusterComponent which encapsulates the logic for starting the cluster components, Dispatcher, RestServerEndpoint and the ResourceManager. The individual components are created by using a factory instance. The ClusterEntrypoint is now only responsible for managing the required services needed by the ClusterComponent. This design should make the testing of these components easier, improve reusability and reduce code duplication.
Move the logic of when to exit the JVM process out of the ClusterEntrypoint so that the caller is now responsible to make this call. This improves the usage of the ClusterEntrypoint for testing purposes.
This commit introduces the DispatcherResourceManagerComponentFactory which is used to create a DispatcherResourceManagerComponent. That way, it is possible to eagerly initialize all fields of the DispatcherResourceManagerComponent making it possible to make all fields final and remove the lock.
|
Thanks for the review @StefanRRichter. Addressing your comment and then merging this PR. |
a4f768e
to
77ca1e2
Compare
What is the purpose of the change
This PR is based on #6743 and #6742 and is part of the removal of Flink's legacy mode.
Port
ProcessFailureCancelingITCaseto new code base using theSessionClusterComponent.Verifying this change
Does this pull request potentially affect one of the following parts:
@Public(Evolving): (no)Documentation