[
https://issues.apache.org/jira/browse/FLINK-25423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17464577#comment-17464577
]
Seth Wiesman commented on FLINK-25423:
--------------------------------------
This sounds reasonable. I’m actively working on porting the api from dataset to
datastream. I’ll pick this up as follow up to that work
> Enable loading state backend via configuration in state processor api
> ---------------------------------------------------------------------
>
> Key: FLINK-25423
> URL: https://issues.apache.org/jira/browse/FLINK-25423
> Project: Flink
> Issue Type: Improvement
> Components: API / State Processor, Runtime / State Backends
> Reporter: Yun Tang
> Assignee: Seth Wiesman
> Priority: Major
> Fix For: 1.15.0
>
>
> Currently, state processor API would load savepoint via explictly
> initalizated state backend on client side, which is like
> {{StreamExecutionEnvironment#setStateBackend(stateBackend)}}:
> {code:java}
> Savepoint.load(bEnv, "hdfs://path/", new HashMapStateBackend());
> {code}
> As we all konw, stream env also support to load state backend via
> configuration to provide flexibility to load state backends especially some
> customized state backend. This could also benefit state processor API with
> similiar ability.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
