Github user StefanRRichter commented on a diff in the pull request:
https://github.com/apache/flink/pull/3770#discussion_r113920542
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinator.java
---
@@ -1025,11 +1026,11 @@ public boolean restoreLatestCheckpointedState(
LOG.info("Restoring from latest valid checkpoint: {}.",
latest);
// re-assign the task states
-
- final Map<JobVertexID, TaskState> taskStates =
latest.getTaskStates();
+ final Map<OperatorID, OperatorState> operatorStates =
latest.getOperatorStates();
StateAssignmentOperation stateAssignmentOperation =
- new StateAssignmentOperation(LOG,
tasks, taskStates, allowNonRestoredState);
+ new StateAssignmentOperation(LOG,
tasks, operatorStates, allowNonRestoredState);
--- End diff --
Not sure why this is implemented in a way that a logger is passed to the
`StateAssignmentOperation`. I guess the class should simply have its own
logger. I think this could be changed. But seems like this was introduced
earlier and is unrelated to this PR. But I wouldn't to have this refactored to
the normal logger scheme before we merge.
---
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.
---
