docker-cardano-node

Builds a Cardano full node from source on Debian. This image attempts to keep interface compatibility with inputoutput/cardano-node, but may diverge slightly, particularly with any Nix-specific paths.

We found that the learning curve for using Nix is too high for many in the community who need to run a full node for various reasons but don't want to get too heavy into learning a new operating system just to extend a container image. This image uses FHS paths and installs into /usr/local while taking advantage of multi-stage image builds to reduce final image size over the official Nix-based distribution.

Running a Cardano Node

The container image has some differing behaviors depending on how it's invoked. There is a NETWORK environment variable which can be used as a short cut to starting nodes on a specific named Cardano network with a default configuration.

We recommend using this method for running the container if you are not familiar with running a node.

For nodes on the preprod and mainnet networks, the container will start fetching the latest Mithril snapshot if an empty data directory is detected. This snapshot will be used to bootstrap the node into an operating state much more rapidly than doing a sync from the beginning of the chain. This behavior can be disabled in the advanced path below.

Using NETWORK (Recommended)

Using the NETWORK environment variable causes the entrypoint script to use the simplified path. The only configurable option in this path is the NETWORK environment variable itself.

This is in keeping with the IOHK image behavior.

To run a Cardano full node on preprod:

docker run --detach \
  --name cardano-node \
  -v node-data:/data/db \
  -v node-ipc:/ipc \
  -e NETWORK=preprod \
  -p 3001:3001 \
  ghcr.io/blinklabs-io/cardano-node

Node logs can be followed:

docker logs -f cardano-node

The node can be monitoring using nview from within the container. We run nview from within the same container as our running node to get access to information about the node process.

docker exec -ti cardano-node nview

Running the Cardano CLI

To run a Cardano CLI command, we use the node-ipc volume from our node.

docker run --rm -ti \
  -e NETWORK=preprod \
  -v node-ipc:/ipc \
  ghcr.io/blinklabs-io/cardano-node cli <command>

This can be added to a shell alias:

alias cardano-cli="docker run --rm -ti \
  -e NETWORK=preprod \
  -v node-ipc:/ipc \
  ghcr.io/blinklabs-io/cardano-node cli"

Now, you can use cardano-cli commands as you would, normally.

cardano-cli query tip --network-magic 1

Or, for a node running with NETWORK=mainnet:

cardano-cli query tip --mainnet

Using run (Recommended for SPO/Advanced)

Using the run argument to the image causes the entrypoint script to use the fully configurable code path. Do not set the NETWORK environment variable.

To run a Cardano full node on mainnet with minimal configuration and defaults:

docker run --detach \
  --name cardano-node \
  -v node-data:/opt/cardano/data \
  -v node-ipc:/opt/cardano/ipc \
  -p 3001:3001 \
  ghcr.io/blinklabs-io/cardano-node run

NOTE The container paths for persist disks are different in this mode

The above maps port 3001 on the host to 3001 on the container, suitable for use as a mainnet relay node. We use docker volumes for our persistent data storage and IPC (interprocess communication) so they live beyond the lifetime of the container. To mount local host paths instead of using Docker's built in volumes, use the local path on the left side of a -v argument.

An example of a more advanced configuration, running a block producer for mainnet:

docker run --detach \
  --name cardano-node \
  --restart unless-stopped \
  -e CARDANO_BLOCK_PRODUCER=true \
  -e CARDANO_SHELLEY_KES_KEY=/opt/cardano/config/keys/kes.skey \
  -e CARDANO_SHELLEY_OPERATIONAL_CERTIFICATE=/opt/cardano/config/keys/node.cert \
  -e CARDANO_SHELLEY_VRF_KEY=/opt/cardano/config/keys/vrf.skey \
  -v /src/cardano/node-keys:/opt/cardano/config/keys \
  -v /srv/cardano/node-db:/opt/cardano/data \
  -v /srv/cardano/node-ipc:/opt/cardano/ipc \
  -p 3001:3001 \
  -p 12798:12798 \
  ghcr.io/blinklabs-io/cardano-node run

The above uses Docker's built in supervisor to restart a container which fails for any reason. This will also cause the container to automatically restart after a host reboot, so long as Docker is configured to start on boot. We set variables to configure a block producer and pass the paths to the 3 keys we need. Our node's persistent data and client communication socket are mapped to /src/cardano/node-db and /src/cardano/node-ipc on the host, respectively. This allows for running applications directly on the host which may need access to these. We also map /src/cardano/node-keys on the host to a path within the container to support running as a block producer. Last, we add mapping the host's port 12798 to the container 12798, which is the port for exposing the node's metrics in Prometheus format, for monitoring.

This mode of operation allows configuring multiple facets of the node using environment variables, described below.

Node logs can be followed:

docker logs -f -n 500 cardano-node

Adding the -n 500 to the logs command limits the logs to the last 500 lines before following.

Configuration using environment variables

The power in using run is being able to configure the node's behavior to provide the correct type of service.

This behavior can be changed via the following environment variables:

• CARDANO_CONFIG_BASE
  • A directory which contains configuration files (default: /opt/cardano/config)
  • This variable is used as the base for other configuration variable default
• CARDANO_BIND_ADDR
  • IP address to bind for listening (default: 0.0.0.0)
• CARDANO_BLOCK_PRODUCER
  • Set to true for a block producing node (default: false)
  • Requires key files and node certificates to be present to start
• CARDANO_CONFIG
  • Full path to the Cardano node configuration (default: ${CARDANO_CONFIG_BASE}/mainnet/config.json)
  • Use your own configuration file to modify the node behavior fully
• CARDANO_DATABASE_PATH
  • A directory which contains the ledger database files (default: /opt/cardano/data)
  • This is the location for persistent data storage for the ledger
• CARDANO_PORT
  • TCP port to bind for listening (default: 3001)
• CARDANO_RTS_OPTS
  • Controls the Cardano node's Haskell runtime (default: -N2 -A64m -I0 -qg -qb --disable-delayed-os-memory-return)
  • This allows tuning the node for specific use cases or resource contraints
• CARDANO_SOCKET_PATH
  • UNIX socket path for listening (default: /opt/cardano/ipc/socket)
  • This socket speaks Ouroboros NtC and is used by client software
• CARDANO_TOPOLOGY
  • Full path to the Cardano node topology (default: ${CARDANO_CONFIG_BASE}/mainnet-topology.json)

Running a block producer

To run a block producing node, you should, at minimum, configure your topology to recommended settings and have only your own relays listed. You will need to set CARDANO_BLOCK_PRODUCER to true and provide the appropriate key files and operational certificate.

• CARDANO_SHELLEY_KES_KEY
  • Stake pool hot key, authenticates (default: ${CARDANO_CONFIG_BASE}/keys/kes.skey)
• CARDANO_SHELLEY_VRF_KEY
  • Stake pool signing key, verifies (default: ${CARDANO_CONFIG_BASE}/keys/vrf.skey)
• CARDANO_SHELLEY_OPERATIONAL_CERTIFICATE
  • Stake pool identity certificate (default: ${CARDANO_CONFIG_BASE}/keys/node.cert

Controlling Mithril snapshots

If the container does not find a protocolMagicId file within the CARDANO_DATABASE_PATH location, it will initiate Mithril snapshot downloads for preprod and mainnet networks. This can be disabled by setting RESTORE_SNAPSHOT to false.

• AGGREGATOR_ENDPOINT
  • Mithril Aggregator URL (default: https://aggregator.release-${CARDANO_NETWORK}.api.mithril.network/aggregator)
• GENESIS_VERIFICATION_KEY
  • Network specific Genesis Verification Key (default: reads file at: ${CARDANO_CONFIG_BASE}/${CARDANO_NETWORK}/genesis.vkey)
• SNAPSHOT_DIGEST
  • Digest identifier to fetch (default: latest)

Running the Cardano CLI

To run a Cardano CLI command, we use the node-ipc volume from our node.

docker run --rm -ti \
  -v node-ipc:/opt/cardano/ipc \
  ghcr.io/blinklabs-io/cardano-node cli <command>

This can be added to a shell alias:

alias cardano-cli="docker run --rm -ti \
  -v node-ipc:/opt/cardano/ipc \
  ghcr.io/blinklabs-io/cardano-node cli"

Now, you can use cardano-cli commands as you would, normally.

cardano-cli query tip --mainnet