Maintenance intent deployment

Deploying a maintenance intent causes the system to deploy configuration files to the affected nodes, which ultimately results in the fabric changes embodied in the intent.

Before you can deploy your maintenance intent, you must have saved the maintenance intent and generated its configuration.

Deploying the maintenance intent involves two procedures:

  • You must add the maintenance intent to the deployment pipeline, so the intent can take its place in the list of planned deployments of fabric, workload, and maintenance intents for the current region.

    Adding a maintenance intent to the pipeline triggers preparatory actions that the system performs behind the scenes. These actions must finish before the system allows you to proceed with deployment.

    For example, if your maintenance intent is replacing a node, adding the intent to the deployment pipeline causes the system to stop the DHCP client application on the existing node to release its current IP address. Only after this is complete can you deploy the maintenance intent from the deployment pipeline.

  • From the deployment pipeline, you must then manually deploy the intent. This signals to the system that it can proceed with the deployment of configuration files to all participating nodes. This deployment may not occur immediately; the system resolves each active deployment in sequence, waiting until one is complete before proceeding to the next.

Adding a maintenance intent to the deployment pipeline

To add the maintenance intent to the region's deployment pipeline, do the following:
  1. Open the list of maintenance intents.
  2. Select the maintenance intent you want to deploy from the displayed list.
  3. Click the More actions icon ( ) at the right edge of the maintenance intent's row, and select Open from the drop-down list.
  4. Click the Add Intent to Deployment Pipeline icon ( ).
  5. Click the ADD TO PIPELINE button.

Progress of a deployed maintenance intent

When a deployed maintenance intent is in progress, whether for a software update or node replacement, the system locks the affected fabric maintenance. No other operations can be carried out on the locked fabric until the maintenance activity is finished.

The deployed maintenance intent progresses through a series of states, depending on the type of maintenance being performed.

For a software update:

  • Deploy drain policies (1/3): the system diverts traffic away from the node in anticipation of the software update or hardware replacement.

    If the maintenance intent fails at this stage, you must discard the fabric intent changes manually. The details of any failure can be found in the Maintenance Error/Event log.

    If this step is successful, the system automatically proceeds to step 2/3.

  • ZTP (2/3): After the maintenance intent's configuration change has been made upon the node, the system rediscovers the updated node using Zero-touch Provisioning (ZTP). The end result of this state is the post-maintenance node returned to a ready state. If this step is successful, the system automatically proceeds to step 3/3.
  • Deploy undrain policies (3/3): the traffic previously diverted away from the node is restored.

    If the maintenance intent fails at this stage, you must re-deploy the affected fabric intent manually. The details of any failure can be found in the Maintenance Error/Event log.

For a node replacement:

  • Deploy drain policies (1/3): the system diverts traffic away from the node in anticipation of the software update or hardware replacement.

    If this step is successful, the system automatically proceeds to step 2/3.

  • Releasing DHCP IP (2/3): The node’s previous IP address is released from DHCP.If this step is successful, the system automatically proceeds to step 3/3.
  • Deploy undrain policies (3/3): the traffic previously diverted away from the node is restored.

    If the maintenance intent fails at this stage, you will need to re-deploy the affected fabric intent manually. The details of any failure can be found in the Maintenance Error/Event log.

If any issues arise during deployment, they appear as alerts in the alert panel at the right side of the page. In the maintenance intent Design view, the system also highlights them as deployment issues in the status bar, by adding a red circle to the fabrics affected by the deployment error, and with entries in the Events Log.

Deploying a maintenance intent from the deployment pipeline

The maintenance intent remains in the deployment pipeline until you tell the system to proceed with the deployment.

Adding a maintenance intent to the pipeline triggers preparatory actions that the system performs behind the scenes. These actions must finish before the system allows you to proceed with deployment.

When you are ready to proceed with deployment, do the following:

  1. In the Fabric Services System UI, click the menu.
  2. Select Maintenance Intents.
  3. Select the row corresponding to the maintenance intent you want to deploy.
  4. Click the icon at the far-right edge of the row and select Deployment Pipeline from the actions list.
  5. Find your maintenance intent in the deployment pipeline list.
  6. At the right edge of the row click the More actions icon ( ).
  7. From the resulting actions list, select Deploy.