[
https://issues.apache.org/jira/browse/FLINK-20919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17264576#comment-17264576
]
Yang Wang commented on FLINK-20919:
-----------------------------------
It is still a little strange. When you run the Flink per job in attach mode,
Flink client should wait for the execution result and then exit. But in your
situation, it seems that the client directly exits without getting the result.
I am not sure whether it is related with {{TableEnvironment}}. But I have run a
batch word count(DataSet API), it could finish normally.
> After the flink batch job is completed, the yarn application cannot be
> completed.
> ---------------------------------------------------------------------------------
>
> Key: FLINK-20919
> URL: https://issues.apache.org/jira/browse/FLINK-20919
> Project: Flink
> Issue Type: Bug
> Components: Deployment / YARN
> Affects Versions: 1.11.2
> Reporter: Wu
> Priority: Major
> Attachments: flink-Overview.png, flink-completed.png,
> flink-jobManager.png, jobmanager.log, kaLr8Coy.png
>
>
> I submit flink batch job in yarn-cluster mode. After the flink batch job is
> completed, the yarn application cannot be completed. The yarn application
> still occupies a vcore. How to automatically close the yarn application.
>
> {code:java}
> //代码占位符
> EnvironmentSettings settings =
> EnvironmentSettings.newInstance().useBlinkPlanner().inBatchMode().build();
> TableEnvironment tableEnv = TableEnvironment.create(settings);
> tableEnv.executeSql("create table file_table");
> tableEnv.executeSql("create table print_table");
> String sql = "select count(1) from file_table";
> Table table = tableEnv.sqlQuery(sql);
> tableEnv.createTemporaryView("t", table);
> tableEnv.from("t").executeInsert("print_table");
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
