[
https://issues.apache.org/jira/browse/BEAM-10940?focusedWorklogId=507765&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-507765
]
ASF GitHub Bot logged work on BEAM-10940:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 04/Nov/20 20:20
Start Date: 04/Nov/20 20:20
Worklog Time Spent: 10m
Work Description: boyuanzz commented on a change in pull request #13105:
URL: https://github.com/apache/beam/pull/13105#discussion_r517607818
##########
File path:
runners/flink/src/main/java/org/apache/beam/runners/flink/translation/wrappers/streaming/ExecutableStageDoFnOperator.java
##########
@@ -502,6 +690,8 @@ public void close() throws Exception {
processWatermark1(Watermark.MAX_WATERMARK);
while (getCurrentOutputWatermark() <
Watermark.MAX_WATERMARK.getTimestamp()) {
invokeFinishBundle();
+ // Sleep for 5s to wait for any timer to be fired.
+ Thread.sleep(5000);
Review comment:
> Is there a way to only sleep if there are such pending timers and
then wait for precisely as long as the maximum of such timers?
One way I'm coming up with is we only sleep when we are executing
SDF/Process fn.
> Also, I'm surprised these timers get fired at all, since newer versions of
Flink stop processing timer execution once closing the operator (that's why we
are draining the timers manually in super.close()).
I haven't looked into Flink timer service yet but I guess it might be
related to watermark hold.
----------------------------------------------------------------
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]
Issue Time Tracking
-------------------
Worklog Id: (was: 507765)
Time Spent: 10h 20m (was: 10h 10m)
> Portable Flink runner should handle DelayedBundleApplication from
> ProcessBundleResponse.
> ----------------------------------------------------------------------------------------
>
> Key: BEAM-10940
> URL: https://issues.apache.org/jira/browse/BEAM-10940
> Project: Beam
> Issue Type: New Feature
> Components: runner-flink
> Reporter: Boyuan Zhang
> Assignee: Boyuan Zhang
> Priority: P2
> Time Spent: 10h 20m
> Remaining Estimate: 0h
>
> SDF can produce residuals by self-checkpoint, which will be returned to
> runner by ProcessBundleResponse.DelayedBundleApplication. The portable runner
> should be able to handle the DelayedBundleApplication and reschedule it based
> on the timestamp.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
