Sqoop pool size determines the number of deployed Sqoop jobs that you can run concurrently in the Hadoop environment.

By default, Sqoop pool size is set to 100. You can disable the Hadoop batch execution pool by setting the value of the Maximum Hadoop Batch Execution Pool Size property to -1. This forces the Data Integration Service to control Sqoop concurrency by treating each Sqoop job equally.

The following table describes the Maximum Hadoop Batch Execution Pool Size property:

Property	Description	Reccomended Value
Maximum Hadoop Batch Execution Pool Size	Maximum number of deployed jobs that can run concurrently in the Hadoop environment. The Data Integration Service moves Hadoop mapping jobs from the queue to the Hadoop job pool when enough resources are available. Default: 100.	-1

See the "Best Practices for Highly Concurrent Workloads" section of the Data Integration Service topic.

To allow concurrent Sqoop jobs to establish connections to the database, ensure that the database supports concurrent database connections.

Mass ingestion uses Sqoop jobs to ingest data from relational tables to Hive or HDFS targets on the Hadoop cluster. By default, Sqoop jobs spawn 4 tasks. Each task establishes one connection to the relational database.

To reconfigure the number of tasks in a Sqoop job, configure the following Sqoop argument in the JDBC connection:

Argument	Description
-num-mappers -m	Number of mappers (tasks) to run concurrently. Default is 4 when Sqoop jobs run on the Spark engine.

The polling time determines how often the Mass Ingestion tool updates the status of ingestion jobs.

By default, the polling time is 30 seconds. You can decrease the polling time to increase the number of requests and refresh the ingestion job status more frequently.

To configure the polling time, set the following custom property on the Mass Ingestion Service:

POLLING_TIME=<time in seconds>