Managing the Hadoop service on datanodes
After you configure the Hadoop service, you can use the server.sh script to manage the service.
Procedure
You can choose to manage the service on an individual datanode
or manage all of the service instances together.
- To manage the service on an individual datanode, use the LA user
that you created when you configured the service to log in to a Hadoop datanode
server.
- Run the <LA_SERVICE_HOME>/bin/server.sh script
with one of the following parameters:
- Start
- Starts the service on the Hadoop datanode server.
- Stop
- Stops the service on the Hadoop datanode server.
- Status
- Retrieves the status for the Hadoop datanode server.
- Run the <LA_SERVICE_HOME>/bin/server.sh script
with one of the following parameters:
- To manage all of the instances, select one datanode to act
as a LA_Service_Controller_Node. This will manage
the service on all of the datanodes.
- (Optional) Create password-less SSH for the LA user from this datanode to all of the datanodes, including this datanode, in the Hadoop cluster.
- Use the LA user to login to the LA_Service_Controller_Node datanode.
- Run the <LA_SERVICE_HOME>/bin/server.sh script
with one of the following parameters:
- clusterStart
- Starts the service on each Hadoop datanode server.
- clusterStop
- Stops the service on each Hadoop datanode server.
- clusterStatus
- Retrieves the status for each Hadoop datanode server.
If you do not configure password-less SSH connections during the configuration, you are prompted for the password for each datanode server.