How to stop Terraform Enterprise active/active node

Introduction

This guide describes the procedure for gracefully stopping a single node in a Terraform Enterprise active/active deployment.

Expected Outcome

After completing this procedure, you will have gracefully shut down a single node in the cluster, allowing it to finish any in-progress work before stopping.

Prerequisites

• A Terraform Enterprise instance running in a Replicated active/active configuration.
• SSH access to the node you intend to stop.

Procedure

1. Log in to the target node in the active/active cluster using SSH.

2. Run the tfe-admin node-drain command to gracefully stop the node.

   # tfe-admin node-drain

   This command performs the following actions:

   • Quiesces the current node and removes it from service.
   • Allows current work to complete before stopping the application.
   • Prevents the node from picking up new jobs from the Redis queue.

   Note the following behaviors of this command:

   • It only affects the local node (localhost) where it is executed. You cannot run it on one node to drain another.
   • There is no command to reverse the drain process. You must restart the node to restore it to service.

Additional Information

• For more information on active/active administration, refer to the Node Drain documentation.