Node Configurations

Node configurations enable you to customize a node associated with a cluster. Compared to operations performed via the cloud-init script these configurations are managed on the WLM server increasing flexibility and security. Node configurations are supported for PBS, Grid Engine and Slurm clusters.

You can define custom PBS resources, Grid Engine and Slurm complexes for cloud nodes and use scripts with predefined variables to configure cloud nodes. Use node configurations to deploy cloud node manually or automate cloud scaling for PBS Professional, automate cloud scaling for Grid Engine and automate cloud scaling for Slurm.
Warning: Deleting a node configuration, cluster, or node which is in use will cause automations to fail. Do not delete a node configuration, cluster, or node without ensuring all automations and deployments using them are removed.

To create a node configuration:

  1. In NavOps, click Clusters.
  2. Click the name of the cluster to open the cluster overview.
  3. Click Node configurations.
  4. You can add a head node or Mom (execution) node configuration.
  5. Provide a Name and a Description.
  6. For Grid Engine, enter the names of the host groups for the node configuration. The names should start with '@'. If no host groups are entered the configuration is applied to @allhosts.
  7. The required custom resources (complexes for Grid Engine) cloud_identifier and cloud_deployed_by are added automatically.
  8. For additional resources or Grid Engine complexes, click +Add resource.
  9. Enter the Key and Value. For example for Grid Engine complexes, you can specify navops_job_class=RADIOSS,slots=4,cloud_job=true.
  10. You can choose to add a script to run on deploying navops agent in NavOps.
  11. You can choose the default onboarding script or customize the script according to your requirements.
    Warning: If you switch between default and custom while editing the script, your changes are not saved. Do not remove the #<RESOURCES># section in the script.
    You can access following per node variables:
    • HOSTNAME: one (or more) hostnames, comma-separated.
    • PRIVATE_IP: one (or more) private IP addresses, comma-separated (same order as HOSTNAME).
    • CLOUD_IDENTIFIER: node identifier.
    You can access these common variables:
    • INSTANCE_COST: the cost/burn-rate of the instance, in $/h.
    • INSTANCE_TYPE: the instance type name for the deployed node.
    • PREEMPTIBLE: "true" if node is preemptible, "false" otherwise.
    • NODE_CLASS_UID: the UID of the node class that is used for the deployment.
    • NODE_CLASS_NAME: the node class name that is used for the deployment.
    • CLOUD_ENVIRONMENT_UID: the cloud environment UID where the node is deployed.
    • CLOUD_ENVIRONMENT_NAME: the cloud environment name where the node is deployed.
    • CLOUD_CREDENTIAL_UID: the cloud environment UID where the node is deployed.
    • CLOUD_CREDENTIAL_NAME: the cloud credentials used to connect to the cloud to deploy the node.
    • DEPLOYED_BY: the entity that triggers the deployment of the node (can be an automation or a user).
    • LOW_LATENCY_NETWORK: id of a low latency network grouping all nodes, may be empty
    • CORES: number of cores
    • VIRTUAL_CORES: number of virtual cores
  12. For the removal script, you can choose the default one or customize it.
  13. Click Save.
Use the node configuration for scaling and manual deployment. You can add Automations and enable them for the cluster.