GJL commented on a change in pull request #9663: [WIP][FLINK-12433][runtime]
Implement DefaultScheduler stub
URL: https://github.com/apache/flink/pull/9663#discussion_r324535361
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/scheduler/DefaultScheduler.java
##########
@@ -75,10 +137,281 @@ public DefaultScheduler(
slotRequestTimeout,
shuffleMaster,
partitionTracker);
+
+ this.restartBackoffTimeStrategy =
checkNotNull(restartBackoffTimeStrategy);
+ this.slotRequestTimeout = checkNotNull(slotRequestTimeout);
+ this.slotProvider = checkNotNull(slotProvider);
+ this.delayExecutor = checkNotNull(delayExecutor);
+ this.userCodeLoader = checkNotNull(userCodeLoader);
+ this.schedulingStrategyFactory =
checkNotNull(schedulingStrategyFactory);
+ this.failoverStrategyFactory =
checkNotNull(failoverStrategyFactory);
+ this.executionVertexOperations =
checkNotNull(executionVertexOperations);
+ this.executionVertexVersioner =
checkNotNull(executionVertexVersioner);
+ this.conditionalFutureHandlerFactory = new
ConditionalFutureHandlerFactory(executionVertexVersioner);
}
+ //
------------------------------------------------------------------------
+ // SchedulerNG
+ //
------------------------------------------------------------------------
+
@Override
- public void startScheduling() {
- throw new UnsupportedOperationException();
+ public void startSchedulingInternal() {
+ initializeScheduling();
+ schedulingStrategy.startScheduling();
+ }
+
+ private void initializeScheduling() {
+ executionFailureHandler = new
ExecutionFailureHandler(failoverStrategyFactory.create(getFailoverTopology()),
restartBackoffTimeStrategy);
+ schedulingStrategy =
schedulingStrategyFactory.createInstance(this, getSchedulingTopology(),
getJobGraph());
+ executionSlotAllocator = new
DefaultExecutionSlotAllocator(slotProvider, getInputsLocationsRetriever(),
slotRequestTimeout);
+ setTaskFailureListener(new
UpdateTaskExecutionStateInDefaultSchedulerListener(this,
getJobGraph().getJobID()));
+ prepareExecutionGraphForScheduling();
+ }
+
+ @Override
+ public boolean updateTaskExecutionState(final TaskExecutionState
taskExecutionState) {
+ final Optional<ExecutionVertexID> executionVertexIdOptional =
getExecutionVertexId(taskExecutionState.getID());
+ if (executionVertexIdOptional.isPresent()) {
+ final ExecutionVertexID executionVertexId =
executionVertexIdOptional.get();
+ updateState(taskExecutionState);
+
schedulingStrategy.onExecutionStateChange(executionVertexId,
taskExecutionState.getExecutionState());
Review comment:
> Seems the schedulingStrategy is not aware of internal state changes in
this way, like DEPLOYING or SCHEDULED, and can only trigger scheduling upon the
external state changes, like RUNNING, FINISHED, etc.
That's a correct observation. At the moment states other than `FINISHED` are
not needed but the inconsistency should be fixed eventually. The
`DefaultScheduler` could invoke
`SchedulingStrategy#onExecutionStateChange(...)` at the right places.
> Some state changes like CANCELED or FAILED might be notified when it's
from task manager and not notified from scheduler internal errors.
`UpdateTaskExecutionStateInDefaultSchedulerListener` is supposed to
propagate internal errors to `DefaultScheduler`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
