Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.firebolt.io/llms.txt

Use this file to discover all available pages before exploring further.

You can use Docker Compose to create a multi-node Firebolt Core deployment. This is useful if you have multiple hosts that are already on the same network (for example, your home or office LAN); if you have remote machines available on Kubernetes, see Deploy using Kubernetes instead. For the purpose of this example, we will assume that the hostnames of these machines are example-node-a and example-node-b; you can use an arbitrary number of nodes, but beware that you cannot add them later after the cluster has been started.

​
Configuration

On both example-node-a and example-node-b you will need to clone the repository: 
git clone https://github.com/firebolt-db/firebolt-core
And then create the following config.json file. 
{
    "nodes": [
        {
            "host": "example-node-a"
        },
        {
            "host": "example-node-b"
        }
    ]
}
Docker Compose deployments use local directory bind mounts for persistence. This means the local filesystem is the only copy of your data — there is no object storage backing. This setup is suitable for development and testing, but is not recommended for production. For production deployments, use the Kubernetes deployment with object storage configured. See Data durability for details.
Persistence is already enabled by default on each of the nodes through the mounted local directory firebolt-core-data, which you will need to create and assign correct permissions: 
mkdir -p ./firebolt-core-data
sudo chown -R 1111:1111 ./firebolt-core-data
If your hosts are running macOS, use before each docker compose command: 
CORE_USER=root docker compose ...

​
Node 0 start

Run: 
docker compose up
On the cloned repository directory on example-node-a; this will start the first Firebolt Core node. Please note that attempting to send queries to its HTTP endpoint will return an HTTP 500 error since the cluster is not yet healthy - this is expected because the second node is not yet up and running.

​
Node 1 start

Run: 
NODE=1 docker compose -f compose.yaml up
On the cloned repository directory on example-node-b; this will start the second Firebolt Core node.

​
Further nodes

For any further node repeat the steps for node 1 and use an incremental NODE (e.g. 2 and so on).

​
Querying

You can now query each node independently, directly utilizing their resources: 
curl -s "http://example-node-a:3473" --data-binary "select 42";
curl -s "http://example-node-b:3473" --data-binary "select 42";
fb-cli is also present within each of the Docker containers, you can invoke it with: 
docker exec -ti firebolt-core-node-0 fbcli

​
Web UI

Each of the hosts will also have a web UI available on port 9100.