Problem
After upgrading Terraform Enterprise, the instance may fail to load, and its underlying Replicated services enter a continuous stop-and-start loop.
Cause
The Replicated daemon may pull an incorrect container image during the upgrade process. When the image tags for replicated/replicated point to two different versions, the system fails to select the most recent image and enters a restart loop.
To verify if this is the cause, check the Docker images on the Terraform Enterprise instance. The output should not show conflicting tags for the current version.
$ docker images | grep replicated/replicated
Solutions
Solution 1: Manually Retag Replicated Docker Images
This solution involves manually retagging the Docker images for Replicated to ensure they point to the correct current version, then restarting the services.
Manually update the Docker tags for the three Replicated images.
$ docker tag quay.io/replicated/replicated:current replicated/replicated:current
$ docker tag quay.io/replicated/replicated-ui:current replicated/replicated-ui:current
$ docker tag quay.io/replicated/replicated-operator:current replicated/replicated-operator:current
Restart the Replicated services to apply the changes. This command typically requires root privileges.
# systemctl restart replicated replicated-ui replicated-operator
Outcome
After you retag the images and restart the services, the Replicated daemon will use the correct image version, allowing the Terraform Enterprise upgrade to complete successfully.
Additional Information
- For more details on the upgrade process, refer to the official Terraform Enterprise upgrade documentation.
- For information on managing Replicated, consult the relevant Replicated documentation.