StefanRRichter commented on a change in pull request #7571: [FLINK-10724]
Refactor failure handling in check point coordinator
URL: https://github.com/apache/flink/pull/7571#discussion_r276155223
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/PendingCheckpoint.java
##########
@@ -101,7 +100,7 @@
private final CheckpointStorageLocation targetLocation;
/** The promise to fulfill once the checkpoint has been completed. */
- private final CompletableFuture<CompletedCheckpoint>
onCompletionPromise;
+ private final CompletableFuture<CheckpointExecutionResult>
onCompletionPromise;
Review comment:
@yanghua Yes, but take a look at how `CheckpointTriggerResult` is used and
then tell me if you don't agree that it would be better to just get an
exception as usual. It just checks the boolean and then translates it into
`CompletableFuture` anyways. IMO, there is no value in having this class for
the code as it currently is. So instead of mimicking and double down on a
design that does not look useful, why not correct it?
I also disagree that two methods for the results will make it more clean,
just because there are less methods. Inside the methods you have to branch
between success and failure case anyways. I think the manager will not care
about a succesful trigger, only about the two failure scenarios (trigger,
execution) and whether or not there was a checkpoint completed (success).
----------------------------------------------------------------
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
