flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Till Rohrmann (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-10771) Replace hard code of job graph file path with config option for FileJobGraphRetriever
Date Mon, 05 Nov 2018 12:54:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-10771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16675094#comment-16675094
] 

Till Rohrmann commented on FLINK-10771:
---------------------------------------

For the {{AbstractYarnClusterDescriptor}} it does not make much sense to honour the setting
of {{JOB_GRAPH_FILE_PATH}}. Instead this config option should be written by this class.

> Replace hard code of job graph file path with config option for FileJobGraphRetriever
> -------------------------------------------------------------------------------------
>
>                 Key: FLINK-10771
>                 URL: https://issues.apache.org/jira/browse/FLINK-10771
>             Project: Flink
>          Issue Type: Improvement
>          Components: YARN
>            Reporter: vinoyang
>            Assignee: vinoyang
>            Priority: Major
>
> There is a config option to configure the value :
> {code:java}
> public static final ConfigOption<String> JOB_GRAPH_FILE_PATH = ConfigOptions
>    .key("internal.jobgraph-path")
>    .defaultValue("job.graph");
> {code}
> However, we used the default hard code in AbstractYarnClusterDescriptor. 
> This is just a preliminary refactoring, and I finally recommend that we use Zookeeper-based
storage. It has been implemented in FLINK-10292.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message