Github user SivaKaviyappa commented on the issue:
https://github.com/apache/zeppelin/pull/2231
This is works for me on aws emr. I am on emr 5.13.0, livy 0.40 , spark
2.3.0. The important change one has to make is by default aws emr brings up
livy in yarn mode. So if you change that yarn-cluster mode and stop livy
server, zeppelin (emr automatically restarts the services once stopped) then
this above issue does not happen. Thanks to @zjffdu for pointing that out.
# What spark master Livy sessions should use.
livy.spark.master yarn-cluster
---
