Github user zentol commented on a diff in the pull request:
https://github.com/apache/flink/pull/3770#discussion_r113903349
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/jobgraph/JobVertex.java ---
@@ -125,6 +132,8 @@ public JobVertex(String name) {
public JobVertex(String name, JobVertexID id) {
this.name = name == null ? DEFAULT_NAME : name;
this.id = id == null ? new JobVertexID() : id;
+ this.operatorIDs.add(new OperatorID(this.id.getLowerPart(),
this.id.getUpperPart()));
+ this.operatorIdsAlternatives.add(null);
--- End diff --
We only need a single alternative ID for each operatorID, the one set using
`setUIDHash()`.
For savepoints created with 1.3 other alternatives aren't required since
1.3 doesn't use the old hasher.
1.0-1.2 savepoints are converted to the 1.3 format using the alternative
job vertex IDs (see `SavepointV2#convertToOperatorStateSavepointV2`), which is
why we can't remove them. After the conversion however the are identical to a
1.3 savepoint.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
