Error connecting via Spark Thrift -> LLAP on secure cluster

[#1]

Hi all.

I've got a fully secured 2.6.3 cluster and am attempting to test the spark-llap bits so we can take advantage of SparkSQL with Ranger. All the Ranger stuff is working fine and all works fine if connecting directly to LLAP/HS2 on port 10500.

However, following the instructions in this recent HCC article, I am not able to connect from beeline via Thrift on port 10016.

Here is my connection string:

beeline -u 'jdbc:hive2://hdp26-master1:10016/default;principal=hive/_HOST@YOMAMMA.COM'

The error appears to come when ThriftCLIService is negotiating with LLAP for a connection:

17/09/19 14:39:09 WARN ThriftCLIService: Error opening session:
java.sql.SQLException: Could not open client transport with JDBC Uri: jdbc:hive2://xxxxxxxxxxxxxxx:10500/;hive.server2.proxy.user=barney: Peer indicated failure: Unsupported mechanism type PLAIN
at org.apache.hive.jdbc.HiveConnection.openTransport(HiveConnection.java:233)
at org.apache.hive.jdbc.HiveConnection.<init>(HiveConnection.java:178)
at org.apache.hive.jdbc.HiveDriver.connect(HiveDriver.java:105)
at java.sql.DriverManager.getConnection(DriverManager.java:664)
at java.sql.DriverManager.getConnection(DriverManager.java:247)
at com.hortonworks.spark.sql.hive.llap.JDBCWrapper.getConnector(HS2JDBCWrapper.scala:174)
at org.apache.spark.sql.hive.llap.LlapSessionState.connection(LlapSessionState.scala:57)
at org.apache.spark.sql.hive.llap.LlapExternalCatalog$$anonfun$databaseExists$1.apply(LlapExternalCatalog.scala:124)
at org.apache.spark.sql.hive.llap.LlapExternalCatalog$$anonfun$databaseExists$1.apply(LlapExternalCatalog.scala:124)
at org.apache.spark.sql.hive.llap.LlapExternalCatalog.org$apache$spark$sql$hive$llap$LlapExternalCatalog$$tryWithResource(LlapExternalCatalog.scala:389)

... etc.

Additionally, impersonation is disabled.

Any ideas about this?

[#2]

Hi @Mark Stephenson,

From the below line, this looks like impersonation error.

jdbc:hive2://xxxxxxxxxxxxxxx:10500/;hive.server2.proxy.user=barney: Peer indicated failure: Unsupported mechanism type PLAIN

Can you please double check if hadoop.proxyuser.hive.user and hadoop.proxyuser.hive.hosts are set to ('*' or the user and host which you are running the query on)

Thanks,

Aditya

[#3]

Thanks for the tiop, Aditya.

I have set the following, but still receive the same error:

hadoop.proxyuser.hive.hosts=*
hadoop.proxyuser.hive.groups=*
hadoop.proxyuser.hive.users=*

Do you have other ideas?

Thanks.

[#4]

FIX UPDATE:

Problem was that not all settings (if any?) inside "Custom Spark2-Defaults" get picked up by runtime, so I had to add one to "Custom spark2-thrift-sparkconf":

spark.sql.hive.hiveserver2.jdbc.url.principal="hive/_HOST@YOMAMMA.COM"