1u0 commented on a change in pull request #8602: [FLINK-12313] Add workaround
to avoid race condition in SynchronousCheckpointITCase test
URL: https://github.com/apache/flink/pull/8602#discussion_r293019590
##########
File path:
flink-streaming-java/src/test/java/org/apache/flink/streaming/runtime/tasks/SynchronousCheckpointITCase.java
##########
@@ -125,16 +126,24 @@ public void
taskCachedThreadPoolAllowsForSynchronousCheckpoints() throws Excepti
assertEquals(ExecutionState.RUNNING,
task.getExecutionState());
assertEquals(CheckpointingState.NONE,
synchronousCheckpointPhase.getState());
+ // Hack: we are triggering a checkpoint with
advanceToEndOfEventTime = true, to be sure that
+ // triggerCheckpointBarrier has reached the sync
checkpoint latch (by verifying in
+ //
SynchronousCheckpointTestingTask.advanceToEndOfEventTime) and only then
proceeding to
+ // notifyCheckpointComplete.
+ // Without such synchronization, the
notifyCheckpointComplete execution may be executed first and leave this
+ // test in a deadlock.
task.triggerCheckpointBarrier(
42,
156865867234L,
new
CheckpointOptions(CheckpointType.SYNC_SAVEPOINT,
CheckpointStorageLocationReference.getDefault()),
- false);
+ true);
Review comment:
Do you mean to get a future that would wait for result of
`invokable.triggerCheckpoint()` call?
In that case, I think what you are proposing would deadlock the test.
----------------------------------------------------------------
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:
us...@infra.apache.org
With regards,
Apache Git Services
