Page History

...

To configure Hive execution to Spark, set the following property to "spark":

• hive.execution.engine

Some configuration properties for Hive on Spark are documented in the Tez section:

...

hive.

...

spark.

...

job.monitor.timeout

• Default Value: 60 seconds
• Added In: Hive 1.1.0 with HIVE-9337

...

Whether to show explain result at user level for Hive-on-Spark queries. When enabled, will log EXPLAIN output for the query at user level.

hive.prewarm.spark.timeout

• Default Value: 5000ms
• Added In: Hive 3.0.0 with HIVE-17362

Time to wait to finish prewarming Spark executors when hive.prewarm.enabled is true.

Note:  These configuration properties for Hive on Spark are documented in the Tez section because they can also affect Tez:

Remote Spark Driver

The remote Spark driver is the application launched in the Spark cluster, that submits the actual Spark job. It was introduced in HIVE-8528. It is a long-lived application initialized upon the first query of the current user, running until the user's session is closed. The following properties control the remote communication between the remote Spark driver and the Hive client that spawns it.

...

Controls the number of containers to prewarm for Tez (0.13.0 to 1.2.x) or Tez/Spark (1.3.0+). This is for Hadoop 2 only.

hive.prewarm.spark.timeout

• Default Value: 5000ms
• Added In: Hive 3.0.0 with HIVE-17362

...

.

hive.merge.tezfiles

• Default Value: false
• Added In: Hive 0.13.0 with HIVE-6498 and HIVE-6360

...