Registering runners

Introduced in GitLab Runner 15.0, a change to the registration request format prevents the GitLab Runner from communicating with GitLab versions lower than 14.8. You must use a Runner version that is appropriate for the GitLab version, or upgrade the GitLab application.

Runner registration is the process that links the runner with one or more GitLab instances.

You can register multiple runners on the same host machine, each with a different configuration, by repeating the register command.

Requirements

Before you register a runner, you must first:

note
When registering a runner on GitLab.com, the gitlab-ci coordinator URL is https://gitlab.com.

Docker

Prerequisites:

The following steps describe launching a short-lived gitlab-runner container to register the container you created during install. After you finish registration, the resulting configuration is written to your chosen configuration volume (for example, /srv/gitlab-runner/config) and is loaded by the runner using that configuration volume.

To register a runner using a Docker container:

  1. Run the register command based on the mount type:

    For local system volume mounts:

    docker run --rm -it -v /srv/gitlab-runner/config:/etc/gitlab-runner gitlab/gitlab-runner register
    note
    If you used a configuration volume other than /srv/gitlab-runner/config during install, be sure to update the command with the correct volume.

    For Docker volume mounts:

    docker run --rm -it -v gitlab-runner-config:/etc/gitlab-runner gitlab/gitlab-runner:latest register
  2. Enter your GitLab instance URL (also known as the gitlab-ci coordinator URL).
  3. Enter the token you obtained to register the runner.
  4. Enter a description for the runner. You can change this value later in the GitLab user interface.
  5. Enter the tags associated with the runner, separated by commas. You can change this value later in the GitLab user interface.
  6. Enter any optional maintenance note for the runner.
  7. Provide the runner executor. For most use cases, enter docker.
  8. If you entered docker as your executor, you are asked for the default image to be used for projects that do not define one in .gitlab-ci.yml.
note
Using gitlab-runner restart inside a Docker container starts a new process instead of restarting the existing process. To apply configuration changes, restart the Docker container instead.

Linux

To register a runner under Linux:

  1. Run the following command:

    sudo gitlab-runner register

    If you are behind a proxy, add an environment variable and then run the registration command:

    export HTTP_PROXY=http://yourproxyurl:3128
export HTTPS_PROXY=http://yourproxyurl:3128

sudo -E gitlab-runner register
  2. Enter your GitLab instance URL (also known as the gitlab-ci coordinator URL).
  3. Enter the token you obtained to register the runner.
  4. Enter a description for the runner. You can change this value later in the GitLab user interface.
  5. Enter the tags associated with the runner, separated by commas. You can change this value later in the GitLab user interface.
  6. Enter any optional maintenance note for the runner.
  7. Provide the runner executor. For most use cases, enter docker.
  8. If you entered docker as your executor, you are asked for the default image to be used for projects that do not define one in .gitlab-ci.yml.

macOS

note
Install Docker.app before registering a runner under macOS.

To register a runner under macOS:

  1. Run the following command:

    gitlab-runner register
  2. Enter your GitLab instance URL (also known as the gitlab-ci coordinator URL).
  3. Enter the token you obtained to register the runner.
  4. Enter a description for the runner. You can change this value later in the GitLab user interface.
  5. Enter the tags associated with the runner, separated by commas. You can change this value later in the GitLab user interface.
  6. Enter any optional maintenance note for the runner.
  7. Provide the runner executor. For most use cases, enter docker.
  8. If you entered docker as your executor, you’ll be asked for the default image to be used for projects that do not define one in .gitlab-ci.yml.

Windows

To register a runner under Windows:

  1. Run the following command:

    .\gitlab-runner.exe register
  2. Enter your GitLab instance URL (also known as the gitlab-ci coordinator URL).
  3. Enter the token you obtained to register the runner.
  4. Enter a description for the runner. You can change this value later in the GitLab user interface.
  5. Enter the tags associated with the runner, separated by commas. You can change this value later in the GitLab user interface.
  6. Enter any optional maintenance note for the runner.
  7. Provide the runner executor. For most use cases, enter docker.
  8. If you entered docker as your executor, you are asked for the default image to be used for projects that do not define one in .gitlab-ci.yml.

FreeBSD

To register a runner under FreeBSD:

  1. Run the following command:

    sudo -u gitlab-runner -H /usr/local/bin/gitlab-runner register
  2. Enter your GitLab instance URL (also known as the gitlab-ci coordinator URL).
  3. Enter the token you obtained to register the runner.
  4. Enter a description for the runner. You can change this value later in the GitLab user interface.
  5. Enter the tags associated with the runner, separated by commas. You can change this value later in the GitLab user interface.
  6. Enter any optional maintenance note for the runner.
  7. Provide the runner executor. For most use cases, enter docker.
  8. If you entered docker as your executor, you are asked for the default image to be used for projects that do not define one in .gitlab-ci.yml.

One-line registration command

If you want to use the non-interactive mode to register a runner, you can either use the register subcommands or use their equivalent environment variables.

To display a list of all the register subcommands, run the following command:

gitlab-runner register -h

Register a runner created in the UI with an authentication token

Introduced in GitLab Runner 15.10.

To register a runner using the most common options, you would do:

sudo gitlab-runner register \
  --non-interactive \
  --url "https://gitlab.com/" \
  --token "$RUNNER_TOKEN" \
  --executor "docker" \
  --docker-image alpine:latest \
  --description "docker-runner"

If you’re running the runner in a Docker container, run the register command with a structure similar to the following example:

docker run --rm -v /srv/gitlab-runner/config:/etc/gitlab-runner gitlab/gitlab-runner register \
  --non-interactive \
  --executor "docker" \
  --docker-image alpine:latest \
  --url "https://gitlab.com/" \
  --token "$RUNNER_TOKEN" \
  --description "docker-runner"

Register a runner with a registration token (deprecated)

Deprecated in GitLab 15.6.

caution
This feature was deprecated in GitLab 15.6 and is planned for removal in 17.0. This change is a breaking change. In GitLab 16.2 and later, legacy-compatible registration processing automatically ignores several deprecated parameters if they are passed during registration. For more information, see Legacy-compatible registration process.

To register a runner using the most common options, run the register command with a structure similar to the following example:

sudo gitlab-runner register \
  --non-interactive \
  --url "https://gitlab.com/" \
  --registration-token "$PROJECT_REGISTRATION_TOKEN" \
  --executor "docker" \
  --docker-image alpine:latest \
  --description "docker-runner" \
  --maintenance-note "Free-form maintainer notes about this runner" \
  --tag-list "docker,aws" \
  --run-untagged="true" \
  --locked="false" \
  --access-level="not_protected"

If you’re running the runner in a Docker container, the register command is structured similar to the following:

docker run --rm -v /srv/gitlab-runner/config:/etc/gitlab-runner gitlab/gitlab-runner register \
  --non-interactive \
  --executor "docker" \
  --docker-image alpine:latest \
  --url "https://gitlab.com/" \
  --registration-token "PROJECT_REGISTRATION_TOKEN" \
  --description "docker-runner" \
  --maintenance-note "Free-form maintainer notes about this runner" \
  --tag-list "docker,aws" \
  --run-untagged="true" \
  --locked="false" \
  --access-level="not_protected"

The --access-level parameter was added in GitLab Runner 12.0. It uses a registration API parameter introduced in GitLab 11.11. Use this parameter during registration to create a protected runner. For a protected runner, use the --access-level="ref_protected" parameter. For an unprotected runner, use --access-level="not_protected" instead or leave the value undefined. This value can later be toggled on or off in the project’s Settings > CI/CD menu.

The --maintenance-note parameter was added in GitLab Runner 14.8. You can use it to add information related to runner maintenance. The maximum allowed length is 255 characters.

Legacy-compatible registration processing

Registration tokens and several runner configuration arguments were deprecated in GitLab 15.6 and will be removed in GitLab 17.0.

To ensure minimal disruption to your automation workflow, the legacy-compatible registration process triggers if an authentication token is specified in the legacy parameter --registration-token. This process causes the following command-line parameters to be ignored. These parameters can only be configured when a runner is created in the UI or with the API.

  • --locked
  • --access-level
  • --run-untagged
  • --maximum-timeout
  • --paused
  • --tag-list
  • --maintenance-note

Check registration token error

The check registration token error message is displayed when the GitLab instance does not recognize the entered registration token. This issue can occur when the instance group or project registration token has been changed in GitLab or when the user did not correctly enter the registration token.

When this error occurs, the first step is to ask a GitLab administrator to verify that the registration token is valid.

Additionally, confirm that runner registration in the project or group is allowed by the GitLab administrator.

[[runners]] configuration template file

Introduced in GitLab Runner 12.2.

Some runner configuration settings can’t be set with environment variables or command line options.

For example:

  • Environment variables do not support slices.
  • Command line option support is intentionally unavailable for the settings for the whole Kubernetes executor volumes tree.

This is a problem for environments that are handled by any kind of automation, such as the GitLab Runner official Helm chart. In cases like these, the only solution was to manually update the config.toml file after the runner was registered. This is less than ideal, error-prone, and not reliable. Especially when more than one registration for the same GitLab Runner installation is done.

This problem can be resolved with the usage of a configuration template file.

To use a configuration template file, pass a path to the file to register with either the:

  • --template-config command line option.
  • TEMPLATE_CONFIG_FILE environment variable.

The configuration template file supports:

When --template-config or TEMPLATE_CONFIG_FILE is used, the configuration of [[runners]] entry is merged into the configuration of newly created [[runners]] entry in the regular config.toml file.

The merging is done only for options that were empty. That is:

  • Empty strings.
  • Nulls or/non existent entries.
  • Zeroes.

With this:

  • All configuration provided with command line options and/or environment variables during the register command call take precedence.
  • The template fills the gaps and adds additional settings.

Example

We register a Kubernetes-executor-based runner to some test project and see what the config.toml file looks like:

$ sudo gitlab-runner register \
     --config /tmp/test-config.toml \
     --non-interactive \
     --url "https://gitlab.com" \
     --registration-token __REDACTED__ \
     --name test-runner \
     --tag-list kubernetes,test \
     --locked \
     --paused \
     --executor kubernetes \
     --kubernetes-host http://localhost:9876/

Runtime platform                                    arch=amd64 os=linux pid=1684 revision=436955cb version=15.11.0

Registering runner... succeeded                     runner=__REDACTED__
Runner registered successfully. Feel free to start it, but if it's running already the config should be automatically reloaded!

The command above creates the following config.toml file:

concurrent = 1
check_interval = 0

[session_server]
  session_timeout = 1800

[[runners]]
  name = "test-runner"
  url = "https://gitlab.com"
  token = "__REDACTED__"
  executor = "kubernetes"
  [runners.cache]
    [runners.cache.s3]
    [runners.cache.gcs]
  [runners.kubernetes]
    host = "http://localhost:9876/"
    bearer_token_overwrite_allowed = false
    image = ""
    namespace = ""
    namespace_overwrite_allowed = ""
    privileged = false
    service_account_overwrite_allowed = ""
    pod_labels_overwrite_allowed = ""
    pod_annotations_overwrite_allowed = ""
    [runners.kubernetes.volumes]

We can see the basic configuration created from the provided command line options:

  • Runner credentials (URL and token).
  • The executor specified.
  • The default, empty section runners.kubernetes with only the one option provided during the registration filled out.

Normally one would have to set few more options to make the Kubernetes executor usable, but the above is enough for the purpose of our example.

Let’s now assume that we have to configure an emptyDir volume for our Kubernetes executor. There is no way to add this while registering with neither environment variables nor command line options. We would have to manually append something like this to the end of the file:

[[runners.kubernetes.volumes.empty_dir]]
  name = "empty_dir"
  mount_path = "/path/to/empty_dir"
  medium = "Memory"

Because TOML doesn’t require proper indentation (it relies on entries ordering), we could just append the required changes to the end of the file. ​ However, this becomes tricky when more [[runners]] sections are being registered within one config.toml file. The assumption that the new one is always at the end is risky.

With GitLab Runner 12.2, this becomes much easier using the --template-config flag.

$ cat > /tmp/test-config.template.toml << EOF
[[runners]]
  [runners.kubernetes]
    [runners.kubernetes.volumes]
      [[runners.kubernetes.volumes.empty_dir]]
        name = "empty_dir"
        mount_path = "/path/to/empty_dir"
        medium = "Memory"
EOF

Having the file, we can now try to register the runner again, but this time adding the --template-config /tmp/test-config.template.toml option. Apart from this change, the rest of registration command is exactly the same:

$ sudo gitlab-runner register \
     --config /tmp/test-config.toml \
     --template-config /tmp/test-config.template.toml \
     --non-interactive \
     --url "https://gitlab.com" \
     --registration-token __REDACTED__ \
     --name test-runner \
     --tag-list kubernetes,test \
     --locked \
     --paused \
     --executor kubernetes \
     --kubernetes-host http://localhost:9876/

Runtime platform                                    arch=amd64 os=linux pid=8798 revision=436955cb version=15.11.0

Registering runner... succeeded                     runner=__REDACTED__
Merging configuration from template file
Runner registered successfully. Feel free to start it, but if it's running already the config should be automatically reloaded!

As we can see, there is a little change in the output of the registration command. We can see a Merging configuration from template file line.

Now let’s see what the configuration file looks like after using the template:

concurrent = 1
check_interval = 0

[session_server]
  session_timeout = 1800

[[runners]]
  name = "test-runner"
  url = "https://gitlab.com"
  token = "__REDACTED__"
  executor = "kubernetes"
  [runners.cache]
    [runners.cache.s3]
    [runners.cache.gcs]
  [runners.kubernetes]
    host = "http://localhost:9876/"
    bearer_token_overwrite_allowed = false
    image = ""
    namespace = ""
    namespace_overwrite_allowed = ""
    privileged = false
    service_account_overwrite_allowed = ""
    pod_labels_overwrite_allowed = ""
    pod_annotations_overwrite_allowed = ""
    [runners.kubernetes.volumes]

      [[runners.kubernetes.volumes.empty_dir]]
        name = "empty_dir"
        mount_path = "/path/to/empty_dir"
        medium = "Memory"

We can see, that the configuration is almost the same as it was previously. The only change is that it now has the [[runners.kubernetes.volumes.empty_dir]] entry with its options at the end of the file. It’s added to the [[runners]] entry that was created by the registration. And because the whole file is saved with the same mechanism, we also have proper indentation.

If the configuration template includes a settings, and the same setting is passed to the register command, the one passed to the register command takes precedence over the one specified inside of the configuration template.

$ cat > /tmp/test-config.template.toml << EOF
[[runners]]
  executor = "docker"
EOF

$ sudo gitlab-runner register \
     --config /tmp/test-config.toml \
     --template-config /tmp/test-config.template.toml \
     --non-interactive \
     --url "https://gitlab.com" \
     --registration-token __REDACTED__ \
     --name test-runner \
     --tag-list shell,test \
     --locked \
     --paused \
     --executor shell

Runtime platform                                    arch=amd64 os=linux pid=12359 revision=436955cb version=15.11.0

Registering runner... succeeded                     runner=__REDACTED__
Merging configuration from template file
Runner registered successfully. Feel free to start it, but if it's running already the config should be automatically reloaded!

As we can see, the registration command is specifying the shell executor, while the template contains the docker one. Let’s see what is the final configuration content:

concurrent = 1
check_interval = 0

[session_server]
  session_timeout = 1800

[[runners]]
  name = "test-runner"
  url = "https://gitlab.com"
  token = "__REDACTED__"
  executor = "shell"
  [runners.cache]
    [runners.cache.s3]
    [runners.cache.gcs]

The configuration set with the register command options took priority and was chosen to be placed in the final configuration.