8.1 Labeling Nodes

Does not apply if you used the ./install-single-node-post.sh installation script

Labeling a node tells Kubernetes what type of application can run on a specific node. It is a means for identifying application processing and qualifying the application as a candidate to run on a specific host system. Labeling is required only for worker nodes and not for master nodes.

You can follow the instructions in this section to label the nodes manually for both single-node and multi-node deployment. However, for single node deployment, you can alternatively use the /opt/<Identity_Intelligence_Installer>/postinstall_label_master_node.sh script to label the node automatically.

To label the nodes:

  1. Launch the CDF Management portal using the link (https://ha-address:5443) that you copied from the Configuration Complete page in Step 11.

  2. Log in with the following credentials:

    User name: admin

    Password: The password that you provided during CDF installation

  3. Select Cluster > Nodes.

  4. In Predefined Labels, click + to add labels.

  5. Specify the labels as follows:

    NOTE:Labels are case-sensitive. Ensure that you enter the values correctly.

    • fusion:yes

    • th-platform:yes

    • th-processing:yes

    • zk:yes

    • kafka:yes

  6. (Conditional) For single-node deployment, drag all the newly added labels to the worker node.

  7. (Conditional) For multi-node deployment, drag-and-drop the new labels from the predefined set to each of the worker nodes based on your workload sharing configuration.

    You may need to click Refresh to see the attached labels.

  8. For Kafka and ZooKeeper, ensure that the number of the nodes you labeled corresponds to # of Worker Nodes in the Kafka cluster and # of Worker Nodes running ZooKeeper in the Kafka cluster properties from the pre-deployment configuration page.

    In a multi-worker deployment, the default number is 3.