Skip to content

[ZEPPELIN-4906]. FLINK_CONF_DIR doesn't work in yarn interpreter mode#3820

Closed
zjffdu wants to merge 1 commit intoapache:masterfrom
zjffdu:ZEPPELIN-4906
Closed

[ZEPPELIN-4906]. FLINK_CONF_DIR doesn't work in yarn interpreter mode#3820
zjffdu wants to merge 1 commit intoapache:masterfrom
zjffdu:ZEPPELIN-4906

Conversation

@zjffdu
Copy link
Contributor

@zjffdu zjffdu commented Jun 24, 2020

What is this PR for?

This use env instead of properties for these flink related config. Because the env will be reset in yarn interpreter mode (https://github.com/apache/zeppelin/blob/master/zeppelin-plugins/launcher/yarn/src/main/java/org/apache/zeppelin/interpreter/launcher/YarnRemoteInterpreterProcess.java#L295)

What type of PR is it?

[Bug Fix]

Todos

  • - Task

What is the Jira issue?

How should this be tested?

  • Manually tested,
    Here's what I see after this PR.
 INFO [2020-06-24 10:46:06,461] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:153) - FLINK_HOME: /Users/jzhang/Temp/hadoop/yarn/usercache/jzhang/appcache/application_1591195707498_0116/container_1591195707498_0116_01_000001/flink
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:154) - FLINK_CONF_DIR: /Users/jzhang/Temp/hadoop/yarn/usercache/jzhang/appcache/application_1591195707498_0116/container_1591195707498_0116_01_000001/flink/conf
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:155) - HADOOP_CONF_DIR: /Users/jzhang/Java/lib/hadoop-2.7.7/etc/hadoop
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:156) - YARN_CONF_DIR: /Users/jzhang/Java/lib/hadoop-2.7.7/etc/hadoop
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:157) - HIVE_CONF_DIR: 
 INFO [2020-06-24 10:46:06,464] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:160) - Using flink: 1.10.0
 INFO [2020-06-24 10:46:06,465] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkShims.java[loadShims]:51) - Initializing shims for Flink 1.10

Screenshots (if appropriate)

Questions:

  • Does the licenses files need update? No
  • Is there breaking changes for older versions? No
  • Does this needs documentation? No

@asfgit asfgit closed this in 5dedab9 Jun 25, 2020
asfgit pushed a commit that referenced this pull request Jun 25, 2020
### What is this PR for?

This use env instead of properties for these flink related config. Because the env will be reset in yarn interpreter mode (https://github.com/apache/zeppelin/blob/master/zeppelin-plugins/launcher/yarn/src/main/java/org/apache/zeppelin/interpreter/launcher/YarnRemoteInterpreterProcess.java#L295)

### What type of PR is it?
[Bug Fix]

### Todos
* [ ] - Task

### What is the Jira issue?
* https://issues.apache.org/jira/browse/ZEPPELIN-4906

### How should this be tested?
* Manually tested,
Here's what I see after this PR.

```
 INFO [2020-06-24 10:46:06,461] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:153) - FLINK_HOME: /Users/jzhang/Temp/hadoop/yarn/usercache/jzhang/appcache/application_1591195707498_0116/container_1591195707498_0116_01_000001/flink
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:154) - FLINK_CONF_DIR: /Users/jzhang/Temp/hadoop/yarn/usercache/jzhang/appcache/application_1591195707498_0116/container_1591195707498_0116_01_000001/flink/conf
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:155) - HADOOP_CONF_DIR: /Users/jzhang/Java/lib/hadoop-2.7.7/etc/hadoop
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:156) - YARN_CONF_DIR: /Users/jzhang/Java/lib/hadoop-2.7.7/etc/hadoop
 INFO [2020-06-24 10:46:06,462] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:157) - HIVE_CONF_DIR:
 INFO [2020-06-24 10:46:06,464] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkScalaInterpreter.scala[initFlinkConfig]:160) - Using flink: 1.10.0
 INFO [2020-06-24 10:46:06,465] ({FIFOScheduler-interpreter_583675831-Worker-1} FlinkShims.java[loadShims]:51) - Initializing shims for Flink 1.10
```
### Screenshots (if appropriate)

### Questions:
* Does the licenses files need update? No
* Is there breaking changes for older versions? No
* Does this needs documentation? No

Author: Jeff Zhang <zjffdu@apache.org>

Closes #3820 from zjffdu/ZEPPELIN-4906 and squashes the following commits:

e38cc30 [Jeff Zhang] [ZEPPELIN-4906]. FLINK_CONF_DIR doesn't work in yarn interpreter mode

(cherry picked from commit 5dedab9)
Signed-off-by: Jeff Zhang <zjffdu@apache.org>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

Comments