Use the Java provided local JMX authentication method, which stores credentials and controls access using a local file. Only use this method to require authentication to run utilities such as nodetool when DSE Unified Authenitcation is disabled on the DSE.
Generally, JMX settings are inserted into the cassandra-env.sh file. However, you can also specify them on the command line:
Set the JMX remote authenticate to true for remote and/or local:
Connections are tested to see if they are local, change the first instance to enable authentication on local connections and the second instance (in the else statement) to enable remote.
On DSE nodes where you want to disable access, set the JMX remote authenticate to false for remote and/or local:
Disabling local may cause unexpected behavior, such as preventing dsetool from running on the node.
On nodes that allow access, set the path to the credentials file:
Ensure that the path is accessible to the user who runs cassandra.
Create a file that contains a user name and password on each line and save it to the location entered in the previous step.
To limit the types of actions a user can perform, create a jmxremote.access file, uncomment the remote access option, and specify the path in the following setting
If all nodes on the cluster were updated, perform a rolling restart; otherwise restart only the affected nodes.
Verify that authentication is working by running a
nodetoolcommand with credentials:
nodetool -u sallyjones -pw 123DSEadmin status
The results should display.
Datacenter: DataStax ===================== Status=Up/Down |/ State=Normal/Leaving/Joining/Moving -- Address Load Tokens Owns Host ID Rack UN 10.200.182.180 316.76 KiB 1 ? 5ca115f6-250a-4964-9a52-c10926031f1b rack1 UN 10.200.182.181 446.76 KiB 1 ? 74a44407-5e26-43d4-83dc-aae9fe35c2f4 rack1 Datacenter: Solr ================ Status=Up/Down |/ State=Normal/Leaving/Joining/Moving -- Address Load Tokens Owns Host ID Rack UN 10.200.182.183 368.38 KiB 1 ? d59d912c-dcc9-469f-8ae1-1c14313e16b1 rack1 Note: Non-system keyspaces don't have the same replication settings, effective ownership information is meaningless
Repeat the configuration on each node in the cluster.