Package org.apache.flink.kubernetes.operator.observer.deployment

Class AbstractFlinkDeploymentObserver

java.lang.Object

org.apache.flink.kubernetes.operator.observer.AbstractFlinkResourceObserver<FlinkDeployment>

org.apache.flink.kubernetes.operator.observer.deployment.AbstractFlinkDeploymentObserver

All Implemented Interfaces:

Observer<FlinkDeployment>

Direct Known Subclasses:

ApplicationObserver, SessionObserver

public abstract class AbstractFlinkDeploymentObserver
extends AbstractFlinkResourceObserver<FlinkDeployment>

Base observer for session and application clusters.

Field Summary

Fields

Modifier and Type	Field	Description
protected org.slf4j.Logger	logger

Fields inherited from class org.apache.flink.kubernetes.operator.observer.AbstractFlinkResourceObserver

configManager, eventRecorder

Constructor Summary

Constructors

Constructor	Description
AbstractFlinkDeploymentObserver(FlinkConfigManager configManager, EventRecorder eventRecorder)

Method Summary

Modifier and Type	Method	Description
protected void	clearErrorsIfDeploymentIsHealthy(FlinkDeployment dep)
protected boolean	isJmDeploymentReady(FlinkDeployment dep)
protected boolean	isSuspendedJob(FlinkDeployment deployment)
protected abstract void	observeFlinkCluster(FlinkResourceContext<FlinkDeployment> ctx)	Observe the flinkApp status when the cluster is ready.
void	observeInternal(FlinkResourceContext<FlinkDeployment> ctx)	Internal observer logic specific to each resource type.
protected void	observeJmDeployment(FlinkResourceContext<FlinkDeployment> ctx)
protected void	updateStatusToDeployedIfAlreadyUpgraded(FlinkResourceContext<FlinkDeployment> ctx)	Checks a resource that is currently in the UPGRADING state whether it was already deployed but we simply miss the status information.

Methods inherited from class org.apache.flink.kubernetes.operator.observer.AbstractFlinkResourceObserver

isResourceReadyToBeObserved, observe

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

logger

protected final org.slf4j.Logger logger

Constructor Detail

AbstractFlinkDeploymentObserver

public AbstractFlinkDeploymentObserver(FlinkConfigManager configManager,
                                       EventRecorder eventRecorder)

Method Detail

observeInternal

public void observeInternal(FlinkResourceContext<FlinkDeployment> ctx)

Description copied from class: AbstractFlinkResourceObserver

Internal observer logic specific to each resource type.

Specified by:

observeInternal in class AbstractFlinkResourceObserver<FlinkDeployment>

Parameters:

ctx - Resource context

observeJmDeployment

protected void observeJmDeployment(FlinkResourceContext<FlinkDeployment> ctx)

isJmDeploymentReady

protected boolean isJmDeploymentReady(FlinkDeployment dep)

clearErrorsIfDeploymentIsHealthy

protected void clearErrorsIfDeploymentIsHealthy(FlinkDeployment dep)

isSuspendedJob

protected boolean isSuspendedJob(FlinkDeployment deployment)

updateStatusToDeployedIfAlreadyUpgraded

protected void updateStatusToDeployedIfAlreadyUpgraded(FlinkResourceContext<FlinkDeployment> ctx)

Description copied from class: AbstractFlinkResourceObserver

Checks a resource that is currently in the UPGRADING state whether it was already deployed but we simply miss the status information. After comparing the target resource generation with the one from the possible deployment if they match we update the status to the already DEPLOYED state.

Specified by:

updateStatusToDeployedIfAlreadyUpgraded in class AbstractFlinkResourceObserver<FlinkDeployment>

Parameters:

ctx - Context for resource.

observeFlinkCluster

protected abstract void observeFlinkCluster(FlinkResourceContext<FlinkDeployment> ctx)

Observe the flinkApp status when the cluster is ready. It will be implemented by child class to reflect the changed status on the flinkApp resource.

Parameters:

ctx - the context with which the operation is executed