Integrating Gitlab and Anka Build Cloud

Instructions on how to use Gitlab with Anka Build Cloud


If you are using Gitlab CI as your CI tool, Veertu provides and maintains a Gitlab CI Runner for Anka Build. The runner connects Gitlab to anka controller and builds your project upon the VM . You can run the runner on any machine.

Prepare VM Template:

  1. Start or create an Anka VM. Install git and your build/test dependencies
  2. Suspend the VM
  3. If you haven't changed your vm default networking, configure port forwarding for ssh: anka modify $VM_NAME add port-forwarding –guest-port 2–host-port 0 ssh
  4. Push the VM to the registry.

Now you have a base image to use for GitLab CI.

There are two options to activate the runner.

Option 1 - Gitlab-anka runner on docker

1. Go to your project settingsCI/CDrunner, and you will see this screen.

gitlab agent info

2. Run docker container with the “anka-gitlab runner” image, with the following command. Url and Token should be taken from the location of the above snapshot.

Docker run command:

 docker run -t asafg6/gitlab-anka-runner --executor anka
 --url http://YOUR_GITLAB_HOST --registration-token ** --ssh-user   
 VM_USER --ssh-password VM_PASSWORD --anka-controller-address
 http://ANKA_CLOUD_ADDRESS --anka-image-id IMAGE_ID --name my-anka-runner 
  • anka-image-id - This is the UUID of the VM template you want to use.
  • ssh-user - ssh user created in the VM template. You can use the default user ‘anka’ that is created
  • ssh-password- ssh user password created in the VM template.

More Anka optional parameters:

* `anka-tag`- value use a specific tag.
* `anka-node`- id value run on a specific node.
* `anka-priority` - value override the task's default priority.
* `anka-group-id` - value run on a specific node group.
* `anka-keep-alive-on-error` - Keep the VM alive in case of an error for debugging purposes.

3. Refresh the gitlab browser page, your runner should show under “the runner activated for this project”

gitlab agent info

*if the runner is on an AWS instance you will need to allow the AWS instance public IP in its own inbound rules- so the runner will be able to connect to the instance.

Last modified February 26, 2020 : anka-buildkite log doc (0938613) by danielkar1