You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Stephan Ewen (JIRA)" <ji...@apache.org> on 2017/04/20 12:20:04 UTC
[jira] [Closed] (FLINK-4321) Initialize TaskExecutions directly
with their starting state
[ https://issues.apache.org/jira/browse/FLINK-4321?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Stephan Ewen closed FLINK-4321.
-------------------------------
Resolution: Won't Fix
This change is actually not needed
> Initialize TaskExecutions directly with their starting state
> ------------------------------------------------------------
>
> Key: FLINK-4321
> URL: https://issues.apache.org/jira/browse/FLINK-4321
> Project: Flink
> Issue Type: Sub-task
> Components: Distributed Coordination
> Affects Versions: 1.1.0
> Reporter: Stephan Ewen
> Assignee: Stephan Ewen
>
> Currently, the CheckpointCoordinator injects state into the Executions of each vertex.
> For more file grained recovery, not all executions need state at the same time, and different executions may be scheduled at different points in time with different completed checkpoints.
> Executions should hence be eagerly initialized with their checkpoint state handles.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)