pvary commented on code in PR #18270:
URL: https://github.com/apache/iceberg/pull/18270#discussion_r4120026659
##########
flink/v1.20/flink/src/main/java/org/apache/iceberg/flink/sink/IcebergCommitter.java:
##########
@@ -138,8 +141,13 @@ public void
commit(Collection<CommitRequest<IcebergCommittable>> commitRequests)
}
IcebergCommittable last =
commitRequestMap.lastEntry().getValue().getCommittable();
+ // Only consult the table's commit history when the job was restored from
a checkpoint or
+ // savepoint. After a stateless restart the checkpoint counter starts
over, so reusing the
+ // previous run's max committed checkpoint id would silently discard every
new commit.
long maxCommittedCheckpointId =
- SinkUtil.getMaxCommittedCheckpointId(table, last.jobId(),
last.operatorId(), branch);
+ isRestored
Review Comment:
Instead of isRestored, do we want to compare the restored checkpoint id with
the one in the table? Like:
```
long maxCommitted = SinkUtil.getMaxCommittedCheckpointId(...);
if (maxCommitted > restoredCheckpointId) { // absent =>
INITIAL_CHECKPOINT_ID
maxCommitted = SinkUtil.INITIAL_CHECKPOINT_ID;
}
```
This would also fix if we try to restore the first checkpointid
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
