Automating Docker pruning in Terraform Enterprise

Expected Outcome

Dangling Docker containers,volumes and images are automatically pruned daily, helping to promote a healthy OverlayFS performance.

Prerequisites

•  Linux System with crond service enabled
•  Root access

Procedure

1. SSH into the Terraform Enterprise host and switch to the root user. NOTE: If you are utilizing either a custom worker image or a custom agent image skip docker image prune -f to avoid potentially having to pull the image manually.   
   

   # Run as root 
   cat > /etc/cron.daily/docker-prune << EOF
   #!/bin/bash
   echo "Pruning dangling Docker containers"
   docker container prune -f
   echo "Pruning dangling Docker volumes"
   docker volume prune -f
   echo "Pruning dangling Docker images"
   docker image prune -f
   > EOF

2. Make the file executable.
   

   # Enable executable bit
   chmod +x /etc/cron.daily/docker-prune

3. Confirm the script will run as expected.    
   

   run-parts /etc/cron.daily
   # Example output 
   /etc/cron.daily/docker-prune:
   Pruning dangling Docker containers
   Total reclaimed space: 0B
   Pruning dangling Docker volumes
   Total reclaimed space: 0B
   Pruning dangling Docker images
   Total reclaimed space: 0B

Additional Information

• Terraform Enterprise suddenly unhealthy due to Docker disk space issues.

• Terraform Enterprise Support Bundles Are Empty