Leaf Node

Leaf node mode allows a knot server to run on a local machine while maintaining access to templates and resources configured within the cluster. This setup is ideal for scenarios where performance is critical or when accessing resources that are only available in a local configuration.

By default, support for leaf mode is enabled. However, it can be disabled by setting server.cluster.allow_leaf_nodes to false in the knot.toml configuration file.

Cluster with Leaf Node
Cluster with Leaf Node

Configuring the Leaf Server

To configure a leaf server, follow these steps:

1. Set Up the Local Server

Begin by setting up a standalone knot server on the local workstation. Follow the Standalone Server Setup Guide, but do not start the server yet. Once the setup is complete, return to this guide.


2. Create a Personal Access Token

A personal access token is required for the leaf server to connect to the cluster. To create one:

  1. Log in to the knot server that the leaf will connect to.
  2. Click on API Tokens in the menu. If this option is unavailable, request permission from your site administrator.
  3. Click New Token and provide a name for the token (e.g., Leaf Test) to identify it.
    New Token Form
    New Token Form
  4. Click Create Token to generate the access token.
    List of Available Tokens
    List of Available Tokens

3. Configure the Leaf Node

Add the token and server information to the knot.toml configuration file of the leaf node:

[server.origin]
server = "https://knot-server.internal"
token = "dp2IlH_wQWIA_ad_UApHjAGYLZMcvn4vsH383N-AB2I="

Configuration Parameters

  • server: The URL of the knot server that the leaf will connect to.
  • token: The API token generated in the previous step. Click the token in the list to copy it to your clipboard for easy pasting.

4. Start the Leaf Server

Start the knot server on the local machine using the following command:

knot server --config knot.toml

Once the server is running, open the web interface of the local instance at https://knot.internal:3000 (as per the standalone guide). Log in using your email address and password from the cluster.


Tokens that have not been used for access in over two weeks are automatically removed from the system.