Start or Stop The Fusion SQL Service

Starting the Fusion SQL service

The Fusion SQL service is an optional service that must be started manually.

Give these commands from the bin directory below the Fusion home directory, for example, /opt/fusion/latest.x (on Unix) or C:\lucidworks\fusion\latest.x (on Windows).

On Unix:

When starting the Fusion SQL service, the best practice is to also start the Spark master and Spark worker services:

./fusion start spark-master spark-worker sql

On Windows:

When starting the Fusion SQL service, the best practice is to also start the Spark master and Spark worker services:

spark-master.cmd start
spark-worker.cmd start
sql.cmd start

Stopping the Fusion SQL service

The Fusion SQL service is an optional service that must be stopped manually.

Give these commands from the bin directory below the Fusion home directory, for example, /opt/fusion/latest.x (on Unix) or C:\lucidworks\fusion\latest.x (on Windows).

On Unix:

When stopping the Fusion SQL service, the best practice is to also stop the Spark master and Spark worker services:

./fusion stop sql spark-worker spark-master

On Windows:

When starting the Fusion SQL service, the best practice is to also start the Spark master and Spark worker services:

sql.cmd stop
spark-worker.cmd stop
spark-master.cmd stop

Updating the group.default definition

If you plan to run the Fusion SQL service for production, we recommend updating the group.default definition in the file conf/fusion.cors (fusion.properties in Fusion 4.x) (on Unix) or conf\fusion.cors (fusion.properties in Fusion 4.x) (on Windows) to include the spark-master, spark-worker, and sql services:

group.default = zookeeper, solr, api, connectors-classic, connectors-rpc, proxy, webapps, admin-ui, log-shipper, spark-master, spark-worker, sql

Verifying that the Fusion SQL service started

Verify the Fusion SQL service application started. Give these commands from the bin directory below the Fusion home directory.

On Unix:

./sql status

On Windows:

sql.cmd status

Alternatively, check the Spark UI, for example:

Check Spark UI