Mappings that run on Hortonworks HDP 3.1 cluster might fail when the following conditions are true:

The mapping reads from custom data object with LLAP enabled.
The mapping runs on the Blaze engine.

You might see an error like:

"Data partitioning is not supported for the mapping due to [Index:
0, Size: 0]

Read the Data Integration Service logs for more information about the failure.

Cloudera ticket reference number: 00228458

When you have a mapping with a Hive ACID source and target tables that runs on the Spark engine, the Summary Statistics view does not reflect any throughput statistics for the mapping job.

Cloudera ticket reference number: 00225986

When a mapping that runs on a Hortonworks HDP 3.1 cluster uses an Update Strategy transformation, and you enabled the Hive Warehouse Connector, the mapping writes incorrect data to rows marked with DD_INSERT.

Cloudera ticket reference number: 00228458

Due to a MapR performance issue, mappings that run on a non-Kerberos MapR cluster might hang with a message like "HDFS encryption zones are different."

On a Hortonworks HDP 3.1 cluster, Spark engine incorrectly converts string data type to binary data type.

Hortonworks ticket reference number: 270