How-to manage /etc/resolv.conf in Terraform Enterprise build worker for Terraform CLI

Introduction

Expected Outcome

Terraform runs in Terraform Enterprise respect the supplied DNS options in /etc/resolv.conf.

Prerequisites

• The permission to execute docker command in Terraform Enterprise host machine.
• The access to the Replicated console on https://<<TFE FQDN>>:8800 or the sudo permission to execute replicatedctl command.

Use Case

• This article aims to provide an example use case to manage DNS options that are suitable for Terraform CLI which is executed inside an isolated docker container. The approach can be used for troubleshooting DNS resolution issues or as a permanent solution, however, some DNS options may have performance implications that would require extra consideration prior to applying the settings. 

Procedure

• Launch a terminal console and ssh into the Terraform Enterprise host machine.

• Create a temporary directory /var/tmp/sample and navigate into the directory.

• Create a custom initialization script init_custom_worker.sh, below is the sample for using the sed command to append DNS options, substitute the placeholder <<DNS OPTIONS>> with the actual DNS options.

  #!/bin/bash
  
  sed 's/\boptions\b/&  <<DNS OPTIONS>> /' /etc/resolv.conf

• Create the Dockerfile with the sample following content which will extend the default worker image hashicorp/build-worker:now with the custom initialization script:

  FROM hashicorp/build-worker:now
  COPY init_custom_worker.sh /usr/local/bin/
  

• Execute command docker build -t hashicorp/build-worker:new . to build a new docker image with the tag hashicorp/build-worker:new, the tag name is only for an example.

• Update Terraform Enterprise to use the custom worker image via the following options:

  • • Command Line Interface

• • As you should already be in the ssh terminal in Terraform Enterprise host machine from the previous steps, executing the following commands will update the build worker image:

    replicatedctl app-config set tbw_image --value "custom_image"
    replicatedctl app-config set custom_image_tag --value "hashicorp/build-worker:new"

  • Restart Terraform Enterprise application to reload the application settings

    • Stop the application with the command

      replicatedctl app stop

      .
    • Watch the status of the application with the command

      watch replicatedctl app status

      .
    • Wait until the State changed to stopped and the IsTransitioning is false
    • Start the application with the command

      replicatedctl app start

• • • Graphical User Interface
      • • Navigate to the Replicated console at the following page, https://<<TFE FQDN>>:8800/settings#worker_image
        • Select the Provide the location of a custom image option
        • Enter the tag name of the new docker image
        • Click the Save button at the bottom of the page to save the settings and restart the application by clicking on the Restart Now button.

Additional Information

• For further clarification/assistance, please contact support at support@hashicorp.com or submit a ticket via our support portal.