You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Zhu Zhu (Jira)" <ji...@apache.org> on 2020/04/08 10:08:00 UTC
[jira] [Created] (FLINK-17047) Simplify
SchedulingStrategy#onPartitionConsumable(...) parameters
Zhu Zhu created FLINK-17047:
-------------------------------
Summary: Simplify SchedulingStrategy#onPartitionConsumable(...) parameters
Key: FLINK-17047
URL: https://issues.apache.org/jira/browse/FLINK-17047
Project: Flink
Issue Type: Sub-task
Components: Runtime / Coordination
Affects Versions: 1.11.0
Reporter: Zhu Zhu
Fix For: 1.11.0
I'd propose to simplify SchedulingStrategy#onPartitionConsumable(...) parameters as below:
1. take IntermediateResultPartitionID instead of ResultPartitionID
ResultPartitionID is a composition of IntermediateResultPartitionID and ExecutionAttemptID. SchedulingStrategy is not aware of ExecutionAttemptID so there is no need to expose it.
2. drop the executionVertexId param. executionVertexId does not provide extra information. The check in LazyFromSourcesSchedulingStrategy does not make much sense since the executionVertexId is just retrieved by the partitionId in an earlier stage. It makes things more complex since a blocking result partition can become consumable when a vertex who is not its producer finishes.
This simplification also eases the work of FLINK-14234.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)