rkhachatryan commented on a change in pull request #13709:
URL: https://github.com/apache/flink/pull/13709#discussion_r509265695
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/ZooKeeperCompletedCheckpointStore.java
##########
@@ -143,6 +145,10 @@ public void recover() throws Exception {
int numberOfInitialCheckpoints = initialCheckpoints.size();
LOG.info("Found {} checkpoints in ZooKeeper.",
numberOfInitialCheckpoints);
+ if (haveAllDownloaded(initialCheckpoints)) {
Review comment:
Yes, if at least one checkpoint is missing (or some extra checkpoints
exist) it will re-download `initialCheckpoints`.
It's quite a sensitive part affecting correctness so I think it's better to
keep it as simple as possible and not optimize further.
With that in mind, I think the only use of returning missing checkpoints
would be to log them.
Is this what you propose?
----------------------------------------------------------------
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]
