[
https://issues.apache.org/jira/browse/BEAM-7520?focusedWorklogId=294928&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-294928
]
ASF GitHub Bot logged work on BEAM-7520:
----------------------------------------
Author: ASF GitHub Bot
Created on: 14/Aug/19 18:46
Start Date: 14/Aug/19 18:46
Worklog Time Spent: 10m
Work Description: je-ik commented on pull request #9190: [BEAM-7520] Fix
timer firing order in DirectRunner
URL: https://github.com/apache/beam/pull/9190#discussion_r314028231
##########
File path:
runners/direct-java/src/main/java/org/apache/beam/runners/direct/StatefulParDoEvaluatorFactory.java
##########
@@ -244,27 +249,40 @@ public void
processElement(WindowedValue<KeyedWorkItem<K, KV<K, InputT>>> gbkRes
delegateEvaluator.processElement(windowedValue);
}
+ Instant lastFired = null;
for (TimerData timer : gbkResult.getValue().timersIterable()) {
checkState(
timer.getNamespace() instanceof WindowNamespace,
"Expected Timer %s to be in a %s, but got %s",
timer,
WindowNamespace.class.getSimpleName(),
timer.getNamespace().getClass().getName());
- WindowNamespace<?> windowNamespace = (WindowNamespace)
timer.getNamespace();
- BoundedWindow timerWindow = windowNamespace.getWindow();
- delegateEvaluator.onTimer(timer, timerWindow);
+ checkState(
+ lastFired == null || !lastFired.isAfter(timer.getTimestamp()),
+ "lastFired was %s, current %s",
+ lastFired,
+ timer.getTimestamp());
+ if (lastFired != null && lastFired.isBefore(timer.getTimestamp())) {
+ pushedBackTimers.add(timer);
+ } else {
+ lastFired = timer.getTimestamp();
+ WindowNamespace<?> windowNamespace = (WindowNamespace)
timer.getNamespace();
+ BoundedWindow timerWindow = windowNamespace.getWindow();
+ delegateEvaluator.onTimer(timer, timerWindow);
Review comment:
I will add the test and check if there are any issues. Regarding the logic
you described - generally, yes, that would be the ideal approach. It would also
solve some other issues, that still remain, even after this PR - but - it would
require a larger refactoring, because there is currently no simple (or simple
and apparent to me) way of knowing if `@OnTimer` method actually setup any new
timer. The `delegateEvaluator` would have to signal that, and that boils down
to `PushbackSideInputDoFnRunner.onTimer` would have to return the new timers.
That is API change in core and I wanted to avoid that. But maybe I've
overlooked some other way.
----------------------------------------------------------------
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
Issue Time Tracking
-------------------
Worklog Id: (was: 294928)
Time Spent: 4h 40m (was: 4.5h)
> DirectRunner timers are not strictly time ordered
> -------------------------------------------------
>
> Key: BEAM-7520
> URL: https://issues.apache.org/jira/browse/BEAM-7520
> Project: Beam
> Issue Type: Bug
> Components: runner-direct
> Affects Versions: 2.13.0
> Reporter: Jan Lukavský
> Assignee: Jan Lukavský
> Priority: Major
> Time Spent: 4h 40m
> Remaining Estimate: 0h
>
> Let's suppose we have the following situation:
> - statful ParDo with two timers - timerA and timerB
> - timerA is set for window.maxTimestamp() + 1
> - timerB is set anywhere between <windowStart, windowEnd), let's denote that
> timerB.timestamp
> - input watermark moves to BoundedWindow.TIMESTAMP_MAX_VALUE
> Then the order of timers is as follows (correct):
> - timerB
> - timerA
> But, if timerB sets another timer (say for timerB.timestamp + 1), then the
> order of timers will be:
> - timerB (timerB.timestamp)
> - timerA (BoundedWindow.TIMESTAMP_MAX_VALUE)
> - timerB (timerB.timestamp + 1)
> Which is not ordered by timestamp. The reason for this is that when the input
> watermark update is evaluated, the WatermarkManager,extractFiredTimers() will
> produce both timerA and timerB. That would be correct, but when timerB sets
> another timer, that breaks this.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
