[jira] [Created] (ZEPPELIN-4077) Interpreter supports dynamic configuration

["Xun Liu (JIRA)" <ji...@apache.org>]

Xun Liu created ZEPPELIN-4077:
---------------------------------

             Summary: Interpreter supports dynamic configuration
                 Key: ZEPPELIN-4077
                 URL: https://issues.apache.org/jira/browse/ZEPPELIN-4077
             Project: Zeppelin
          Issue Type: Improvement
         Environment: There are too many interpreters, it is difficult to maintain, and the user experience is not good.

!interpreter dynamic parameter.gif!
            Reporter: Xun Liu
            Assignee: Xun Liu
         Attachments: interpreter dynamic parameter.gif

In a multi-user scenario with multiple Hadoop clusters, Because different users and user groups have different permissions, Need to specify different Hadoop clusters and YARN queues.
The current interpreter configuration is fixedly configured, such as an interpreter configured with a YARN queue.
So, this creates a lot of interpreters that need to be configured in a zeppelin service.

There are too many interpreters, it is difficult to maintain, and the user experience is not good.

 

I think about this for extended support:
1. When the user logs in, it provides an extended interface for obtaining the user's queue information and kerberos information. It is not configured by default.
2. When the user can execute the interpreter related to the hadoop cluster configuration (for example: spark, hdfs file interpreter), provide a drop-down list that allows the user to select the HADOOP_CONF configuration item.
3. When the user can execute the interpreter related to kerberos configuration (for example: spark, hdfs file, jdbc interpreter), the user's kerberos information is automatically used to replace the keytab and principal configuration in the interpreter.



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