You can change the default IBM Z Database Assistant Spark settings to improve
performance.
About this task
You can choose to use more than 2 spark worker threads for system assessments so that they
complete quicker. You can also choose to update the Spark port settings. Follow this process to
change the Spark settings manually to override the default settings.
Procedure
To change the IBM Z Database Assistant Spark
worker thread settings:
-
Locate the assessconfig.json file in the
$ZDBA_HOME/config/ directory.
- To enable system assessment tasks to complete more quickly, increase the number of cores
that Spark can use from "2" to a higher number (for example, "4"):
"ASSESSMENT_SPARK_MAX_CORES": "4"
To let Spark use all the available cores for its processing, set the value of
"ASSESSMENT_SPARK_MAX_CORES" to "0" as follows:
"ASSESSMENT_SPARK_MAX_CORES": "0"
To update the Spark port settings, including the starting port in the range of ports
used by Spark and the maximum number of ports in the range:
- To specify the starting port for Spark, change the value of
"ASSESSMENT_SPARK_DRIVER_PORT_FOR_LOCAL" from "random" to a specific network port:
"ASSESSMENT_SPARK_DRIVER_PORT_FOR_LOCAL": "random"
- To specify the size of the range of ports from the default of 16, change the following
value:
"ASSESSMENT_SPARK_DRIVER_PORT_RETRIES_FOR_LOCAL": "16"