pnowojski commented on a change in pull request #15146:
URL: https://github.com/apache/flink/pull/15146#discussion_r594466733
##########
File path:
flink-test-utils-parent/flink-test-utils/src/main/java/org/apache/flink/streaming/util/TestStreamEnvironment.java
##########
@@ -95,6 +96,14 @@ private static void randomize(Configuration conf) {
final PseudoRandomValueSelector valueSelector =
PseudoRandomValueSelector.create(testName != null ?
testName : "unknown");
valueSelector.select(conf,
ExecutionCheckpointingOptions.ENABLE_UNALIGNED, true, false);
+ valueSelector.select(
+ conf,
+ ExecutionCheckpointingOptions.ALIGNMENT_TIMEOUT,
+ Duration.ofSeconds(0),
+ Duration.ofSeconds(1),
+ Duration.ofSeconds(5),
+ Duration.ofSeconds(30),
Review comment:
Are three of those settings really necessary? Maybe we can stick to `0`,
short value (100ms) and long value (1s? 2s?)? At least until it's proven that
we really need to test more cases?
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/OneInputStreamTask.java
##########
@@ -143,7 +143,8 @@ private CheckpointedInputGate createCheckpointedInputGate()
{
inputGates,
getEnvironment().getMetricGroup().getIOMetricGroup(),
getTaskNameWithSubtaskAndId(),
- mainMailboxExecutor);
+ mainMailboxExecutor,
+ timerService);
Review comment:
I think as long as we are not checkpointing active timeout timers it
should be fine. I mean, if the timerservier is overloaded, that's probably a
degenerate case potentially not existing in the real world, that we can ignore?
----------------------------------------------------------------
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]
