Skip to main content
Archive node vs. historical proofsThis guide covers running a full archive node, which serves every historical-state RPC call (such as eth_getBalance or eth_call) at any block and requires terabytes of storage. Celo also supports a narrower historical proofs feature for serving historical data without keeping full archive state; a dedicated guide for it is forthcoming.
op-geth is being deprecatedThese instructions use op-geth, which is supported only until the Ethereum Glamsterdam hardfork. Celo is transitioning to op-reth as the primary execution client. See End of Support for op-geth for the migration timeline; archive-node guidance for op-reth will follow once a Celo-compatible release is published.

​
Overview

To run an L2 archive node, you need to start the L2 execution client in archive mode. This allows the node to accept RPC requests that require archive data for blocks created after the L2 transition. For historical data from before the L2 transition, you can configure your node to forward those requests to a legacy Celo L1 archive node that contains the historical blockchain state.

​
Instructions

PrerequisitesThese instructions assume you already have
  1. A migrated full node datadir that has been synced to the migration block. See Migrating an L1 Node if you do not have this.
  2. A non-migrated Celo L1 archive node datadir. Do not attempt to migrate an archive datadir.
Please ensure neither datadir is being used by a running node before proceeding.
  1. Pull the latest version of celo-l2-node-docker-compose and cd into the root of the project. 
    git clone https://github.com/celo-org/celo-l2-node-docker-compose.git
cd celo-l2-node-docker-compose
  2. Configure your .env file.

    ​
    Copy default configurations

    The celo-l2-node-docker-compose repo contains a <network>.env file for each Celo network (celo-sepolia, mainnet). Start by copying the default configuration for the appropriate network. 
    export NETWORK=<celo-sepolia or mainnet>
cp $NETWORK.env .env

    ​
    Configure sync mode

    By default, celo-l2-node-docker-compose will start your node with snap sync. While archive nodes can technically run with snap sync, they will only store archive data from the point that snap sync completes. This will leave a gap in the archive data after the hardfork, so we recommend running archive nodes with full sync and a migrated pre-hardfork datadir. To use full sync, configure .env as follows: 
    OP_GETH__SYNCMODE=full
DATADIR_PATH=<path to a migrated L1 full node datadir>

    ​
    Configure node type

    To enable archive mode, configure .env as follows: 
    NODE_TYPE=archive

    ​
    Configure Historical RPC Service

    To handle RPC requests for pre-hardfork state and execution, an L2 archive node proxies to a legacy archive node or “Historical RPC Service”. There are two ways to configure a Historical RPC Service for your archive node:
    1. Supply a pre-hardfork archive datadir and let celo-l2-node-docker-compose start a legacy archive node. To do this configure .env as follows: 
      HISTORICAL_RPC_DATADIR_PATH=<path to your pre-hardfork archive datadir>
      When you start your L2 node, a legacy archive node will also start using the pre-hardfork archive datadir. Your L2 node will be configured to use the legacy archive node as its Historical RPC Service.
    2. Start the legacy archive node yourself and configure .env as follows: 
      OP_GETH__HISTORICAL_RPC=<RPC endpoint of a running legacy archive node>
      This will cause any value you set for HISTORICAL_RPC_DATADIR_PATH to be ignored. The tool will not start a legacy archive node when it starts your L2 archive node. If you choose to run your own legacy archive node, you should do so with different flags than before the hardfork, as the node will no longer be syncing blocks or communicating with other nodes. To see how we recommend re-starting a legacy archive node as a Historical RPC Service, see this script.

    ​
    Configure P2P for external network access

    Network ConfigurationIf the following options are not configured correctly, your node will not be discoverable or reachable to other nodes on the network. This is likely to impair your node’s ability to stay reliably connected to and synced with the network.
    • OP_NODE__P2P_ADVERTISE_IP - Specifies the public IP to be shared via discovery so that other nodes can connect to your node. If unset op-node other nodes on the network will not be able to discover and connect to your node.
    • PORT__OP_NODE_P2P - Specifies the port to be shared via discovery so that other nodes can connect to your node. Defaults to 9222.
    • OP_GETH__NAT - Controls how op-geth determines its public IP that is shared via the discovery mechanism. If the public IP is not correctly configured then other nodes on the network will not be able to discover and connect to your node. The default value of any will try to automatically determine the public IP, but the most reliable approach is to explicitly set the public IP using extip:<your-public-ip>. Other acceptable values are (any|none|upnp|pmp|pmp:<IP>|extip:<IP>|stun:<IP:PORT>).
    • PORT__OP_GETH_P2P - Specifies the port to be shared via discovery so that other nodes can connect to your node. Defaults to 30303.
  3. Start the node(s). 
    docker-compose up -d --build
  4. Check the progress of your L2 archive node as it syncs. 
    docker-compose logs -n 50 -f op-geth
    This will display and follow the last 50 lines of logs. In a syncing node, you would expect to see Syncing beacon headers downloaded=... where the downloaded number is increasing and later lines such as "Syncing: chain download in progress","synced":"21.07%" where the percentage is increasing. Once the percentage reaches 100%, the node should be synced.
  5. Check that node is fully synced. Once the node is fully synced, you can validate that it’s following the network by fetching the current block number via the RPC API and seeing that it’s increasing as expected. 
    cast block-number --rpc-url http://localhost:9993
    Note that until fully synced, the RPC API will return 0 for the head block number.
  6. Try querying historical state to test archive functionality. 
    cast balance --block <pre-migration-block-number> <address> --rpc-url http://localhost:9993