Github user zjffdu commented on the issue:
https://github.com/apache/zeppelin/pull/2422
@Leemoonsoo @jongyoul @prabhjyotsingh @felixcheung Could you help review
that ?
Here's some suggestion for you how to review it since it is a very large PR
* The overall change is the same as I described in the design doc. I would
suggest you to read the unit test first. These unit test is very readable and
easy to understand what the code is doing now. InterpreterFactoryTest,
InterpreterGroupTest, InterpreterSettingTest, InterpreterSettingManagerTest,
RemoteInterpreterTest.
* Remove the reference counting logic. Now I will kill the interpreter
process as long as all the sessions in the same interpreter group is closed. (I
plan to add another kind of policy for the interpreter process lifecycle
control, ZEPPELIN-2197)
* The RemoteFunction I introduced is for reducing code duplicates when we
use RPC.
* The changes in Job.java and RemoteScheduler is for fixing the race issue
bug. This bug cause the flaky test we see often in
ZeppelinSparkClusterTest.pySparkTest
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
