Zero downtime adjustment of Replicated application settings in Active/Active deployments

Introduction

Certain Terraform Enterprise application settings, available through the replicatedctl command-line utility, may require adjustment to tune performance. For example, you can modify the capacity_concurrency setting to change the maximum number of concurrent runs allowed at any given time.

In an Active/Active deployment, you can update these settings on each node sequentially without incurring application downtime.

Procedure

Follow these steps for each node in your Active/Active configuration.

1. Establish an SSH connection to one of the nodes.

2. Gracefully drain the node from service. This command prevents the node from accepting new work.

   $ tfe-admin node-drain

3. Stop the Terraform Enterprise application on the node.

   $ replicatedctl app stop

4. Confirm the application has stopped. The State should report as stopped.

   $ watch replicatedctl app status
   ## The "State" field should display "stopped"

5. Adjust the desired setting using its key and a new value.

   $ replicatedctl app set <setting> --value <desired_value>

6. Restart the application.

   $ replicatedctl app start

7. Confirm the application has started successfully. The State should report as started.

   $ watch replicatedctl app status
   ## The "State" field should display "started"

8. Repeat this procedure for all other nodes in the Active/Active deployment.

Additional Information

• To view a list of settings that you can modify, run the following command.

  $ replicatedctl app-config export

• For more information on the tfe-admin node-drain command, refer to the article on How to stop a Terraform Enterprise active/active node.
• You can find more details on managing your installation in the official Terraform Enterprise documentation.