I am facing an issue when creating an EMR cluster on AWS. Currently, I have a CloudFormation (CF) script that is invoked by Lambda to create a cluster — this includes steps to setup authentication for Hue, Hive, and other applications. For the purpose of the project I have created a custom authentication for Hive and configured it through the CF script like so:
Configurations:
- Classification: hive-site
ConfigurationProperties:
hive.metastore.client.factory.class: 'com.amazonaws.glue.catalog.metastore.ClientFactory'
hive.server2.authentication: 'CUSTOM'
hive.server2.custom.authentication.class: 'com.common.code.auth.CustomAuthClass'
When I log into Hue on the cluster, none of my tables from Hive are loaded and I get the following error:
TSocket read 0 bytes (code THRIFTTRANSPORT):
TTransportException(‘TSocket read 0 bytes’,)
I ensured the Hive server was running. So this is not the issue. I even tried spinning up a cluster without Hive authentication setup and all my Hive tables were populated, so it must be something to do with the Hive custom authentication. I have played around with some of the hue.ini settings, but didn’t have any success. Does anyone have any advice or know if this is not possible?