sunhaibotb commented on a change in pull request #10151: [FLINK-14231] Handle
the pending processing-time timers to make endInput semantics on the operator
chain strict
URL: https://github.com/apache/flink/pull/10151#discussion_r373999581
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/StreamTask.java
##########
@@ -1046,10 +1046,16 @@ TimerService getTimerService() {
return timerService;
}
- public ProcessingTimeService getProcessingTimeService(int
operatorIndex) {
- Preconditions.checkState(timerService != null, "The timer
service has not been initialized.");
- MailboxExecutor mailboxExecutor =
mailboxProcessor.getMailboxExecutor(operatorIndex);
- return new ProcessingTimeServiceImpl(timerService, callback ->
deferCallbackToMailbox(mailboxExecutor, callback));
Review comment:
If we keep this method, we need to add a map in `StreamTask` to store the
corresponding `ProcessingTimeService` of all operators. In order to properly
handle `quiesce` of `ProcessingTimeService`, the `ProcessingTimeService`
obtained by passing the same `operatorIndex` must always be the same. The idea
of introducing `ProcessingTimeServiceAware` is to store the
`ProcessingTimeService` object in the operator. What do you think? @pnowojski
----------------------------------------------------------------
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
