Running Layer 2 nodes¶

As explained in the Node Types section there are various L2 technologies to scale the Ethereum blockchain and lower the transaction fees.

It is important to keep both L1 and L2 nodes as decentralized as possible and that basically means run nodes.

Important

Unless otherwise specified, all commands should be run under the ethereum user account.

Quick Navigation¶

Fuel Network - Modular execution layer with optimistic rollups (see fuel-network)
Arbitrum - Optimistic rollup with high throughput (see arbitrum)
Optimism - Optimistic rollup scaling solution (see optimism)
Starknet - ZK-rollup with validity proofs (see starknet)
EthRex L2 - Minimalist Rust-based L2 (see ethrex-l2)

Hardware Requirements¶

L2 Solution	Minimum Storage	Recommended Storage	Special Notes
Fuel Network	500 GB	1 TB	Requires synced L1 node
Arbitrum	1 TB	2 TB	Requires L1 node access
Optimism/Base	1 TB	2 TB	Requires L1 node access
Starknet	500 GB	1 TB	Standalone operation
EthRex L2	500 GB	1 TB	Requires L1 node access

Note

If you have an Ethereum on ARM image installed prior to June 2023 you need to install the clients manually. Otherwise you can skip this step

sudo apt-get update
sudo apt-get install arbitrum-nitro starknet-juno optimism-op-geth optimism-op-node fuel-network

Fuel Network Production Ready Rust¶

Official Site: https://fuel.network/

Fuel Network is a modular execution layer designed to provide high-performance, scalable smart contract execution for blockchain applications. It leverages advanced technologies like optimistic rollups and UTXO-based design to achieve low fees, fast transactions, and secure interoperability with Ethereum.

In order to run a Fuel node you need to:

Run and sync and Ethereum mainnet node
Install the fuel-network package
Start the Fuel Systemd Service

Sync an Ethereum node.

You can choose from any Consensus and Execution clients available. See our section “Running L1 Clients”

Note

The Fuel package is configured to run with a local L1 node running and synced. So first, you need to run and sync an Ethereum Node. In case you want to run a Fuel node alone you need to configure the flag --relayer in /etc/ethereum/fuel/fuel.conf file.

Installation

Run the apt command:

sudo apt-get update && sudo apt-get install fuel-network

Start the Fuel Core Service

sudo systemctl start fuel

You can check out the logs by running:

sudo journalctl -u fuel -f

EthRex L2 Alpha Rust¶

Official Site: https://github.com/lambdaclass/ethrex

EthRex is a minimalist, modular execution layer written in Rust. It is designed to be highly performant and easy to modify.

Installation

Run the apt command:

sudo apt-get update && sudo apt-get install ethrex-l2

Configuration

The main configuration file is located at /etc/ethereum/ethrex-l2.conf. You should edit this file to configure your L1 connection and other parameters:

ARG="--eth.rpc-url https://sepolia.infura.io/v3/<YOUR_INFURA_KEY>
--l1.on-chain-proposer-address 0x1111111111111111111111111111111111111111
--l1.bridge-address 0x2222222222222222222222222222222222222222
--committer.l1-private-key 0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
--proof-coordinator.l1-private-key 0xbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
--block-producer.coinbase-address 0x3333333333333333333333333333333333333333
--http.addr 0.0.0.0
--http.port 1729
--metrics
--metrics.port 9092
--datadir /home/ethereum/.ethrex-l2"

Setup Note

Ensure you have a valid L1 RPC URL (Sepolia is used in this example) and the correct contract addresses.

Start the Sequencer

Once configured, start the Ethrex L2 Sequencer service:

sudo systemctl start ethrex-l2
sudo journalctl -u ethrex-l2 -f

Example output:

INFO ethrex_l2::sequencer: Connected to L1 RPC https://sepolia.infura.io/v3/...
INFO ethrex_l2::p2p: Connected peers: 45
INFO ethrex_l2::commit: New batch committed to L1 block #XXXXXXX
INFO ethrex_l2::sequencer: Block 0xabc123… produced (gas_used=8,100,000)

Start the Prover

The Ethrex L2 Prover generates and submits validity proofs for each batch committed by the sequencer.

sudo systemctl start ethrex-l2-prover
sudo journalctl -u ethrex-l2-prover -f

By default it connects to the local proof coordinator (tcp://127.0.0.1:3900) and uses the exec backend. For production, you can switch to SP1 or RISC0 backends in /etc/ethereum/ethrex-l2-prover.conf.

Verify that everything is running

Check the sequencer RPC:

curl http://localhost:1729 \
     -H 'content-type: application/json' \
     -d '{"jsonrpc":"2.0","method":"eth_blockNumber","id":"1","params":[]}'

Expected result:

{"id":"1","jsonrpc":"2.0","result":"0x5"}

The value should increase every few seconds as new L2 blocks are produced.

Manage services

sudo systemctl stop ethrex-l2
sudo systemctl stop ethrex-l2-prover
sudo systemctl enable ethrex-l2 ethrex-l2-prover
sudo journalctl -u ethrex-l2* -f

Notes

Default JWT secret path: /etc/ethereum/jwtsecret
Validium mode (no state diffs on L1): add –validium in /etc/ethereum/ethrex-l2.conf
Both services use the ethereum user
Data stored at /home/ethereum/.ethrex-l2/

Quick start

sudo apt-get update && sudo apt-get install ethrex-l2
sudo systemctl start ethrex-l2
sudo systemctl start ethrex-l2-prover
curl http://localhost:1729 -H 'content-type: application/json' \
     -d '{"jsonrpc":"2.0","method":"eth_blockNumber","id":"1","params":[]}'

Congrats, your Ethrex L2 Sequencer and Prover are now running on your Ethereum on ARM device.

Arbitrum Production Ready Go¶

Official Site: https://arbitrum.io/

Arbitrum uses a technology called Optimistic Rollups to bundle multiple transactions into a single proof that is submitted to the Ethereum mainnet (Layer 1). By moving much of the transaction processing and computation off-chain, Arbitrum reduces congestion and gas fees on the Ethereum network, while maintaining a high level of security and decentralization.

The Arbitrum Nitro client is available.

Note

You need a L1 node to connect to in order to run an Arbitrum node.

First step is to set the IP for your L1 Ethereum node:

sudo sed -i "s/setip/192.168.0.10/" /etc/ethereum/nitro.conf

Replace 192.168.0.10 with your actual L1 node IP address.

We need to download and decompress the initial snapshot in order to initialize the database. Run:

mkdir -p /home/ethereum/.arbitrum/arb1/nitro
cd /home/ethereum/.arbitrum/arb1/nitro
wget -O - https://snapshot.arbitrum.foundation/arb1/nitro-pruned.tar | tar -xvf -

Warning

The snapshot is very large (2+ TB). Use screen or tmux to prevent disconnection.

Once finished, start the Nitro client service and wait for the client to get in sync:

sudo systemctl start nitro
sudo journalctl -u nitro -f

The Arbitrum node is up and running.

Starknet Production Ready Rust¶

Official Site: https://www.starknet.io/

StarkNet is a Layer 2 scaling solution for the Ethereum blockchain, designed to improve scalability, transaction throughput, and efficiency using a technology called Zero-Knowledge (ZK) Rollups. This approach allows StarkNet to bundle multiple transactions together, process them off-chain, and then submit a proof of their validity to the Ethereum mainnet (Layer 1).

Available Clients¶

There are 2 production-ready clients available for running a Starknet node:

Juno - Golang implementation: A full Starknet node implementation written in Go by Nethermind. Juno is known for its stability, performance, and lower resource requirements. It’s the recommended choice for most users.
Madara - Substrate-based sequencer: A high-performance Starknet sequencer built on Substrate framework by the Madara Alliance. Madara is designed for advanced use cases and can be configured as both a full node and a sequencer.

Juno¶

Juno is a Golang Starknet full-node implementation by Nethermind.

Features: - Full node synchronization - Lower resource requirements - Stable and well-tested - Active development and support

Installation and Setup:

Start the Juno client:

sudo systemctl start juno
sudo journalctl -u juno -f

Juno will sync with the Starknet network and provide RPC access once synchronized.

Madara¶

Madara is a high-performance Starknet sequencer built on Substrate.

Features: - Substrate-based architecture - Can run as full node or sequencer - High performance and scalability - Modular design for customization

Installation and Setup:

Start the Madara client:

sudo systemctl start madara
sudo journalctl -u madara -f

Configuration¶

The default configuration file is located at /etc/ethereum/madara.conf. You can edit this file to change the default parameters.

# /etc/ethereum/madara.conf
MADARA_OPTS="--base-path /var/lib/madara --rpc-port 9944"

After changing the configuration, restart the service:

sudo systemctl restart madara

Client Comparison¶

Feature	Juno	Madara
Language	Go	Rust (Substrate)
Resource Usage	Lower	Moderate
Use Case	Full node	Full node / Sequencer
Maturity	Stable	Production Ready
Recommended For	Most users	Advanced setups

Optimism / Base Production Ready Go¶

Official Sites: https://www.optimism.io/ | https://base.org/

Optimism is a Layer 2 scaling solution for Ethereum that uses Optimistic Rollups to increase scalability. Base (developed by Coinbase) is built on the same Optimism stack.

Architecture¶

An Optimism/Base node consists of two components:

Execution Client - Processes transactions (op-geth or op-reth)
Consensus Client - Op-Node (rollup node that derives L2 chain from L1)

Important

You need access to a synced Ethereum L1 node (both Execution and Consensus layers).

Quick Start¶

Installation:

sudo apt-get update
sudo apt-get install optimism-op-geth optimism-op-node

Configure L1 connection in /etc/ethereum/op-node.conf:

ARGS="--l1=http://YOUR_L1_IP:8545 \
    --l1.beacon=http://YOUR_L1_IP:5052 \
    ..."

Start the services:

sudo systemctl start op-geth
sudo systemctl start op-node

Linea¶

Linea is a developer-ready zkEVM rollup for scaling Ethereum dApps. It uses a Dual-Layer Architecture:

Execution Layer (EL): Besu (packaged as linea-besu).
Consensus Layer (CL): Maru (packaged as maru). Maru replaces the Clique PoA consensus with QBFT and coordinates with the Ethereum mainnet.

Important

Resource Requirements

RAM: 16GB Recommended.
Disk Space: - Full Node: ~510 GB (Growing ~9 MB/day). - Archive Node: ~4 TB (Growing ~3 GB/day).

Installation:

# Install Execution Client (Besu)
sudo apt install linea-besu

# Install Consensus Client (Maru) - Required for consensus
sudo apt install maru

Profiles & Services:

The package provides pre-configured systemd services for different profiles:

Basic Profile (Follower Node): - Mainnet: linea.service - Sepolia: linea-sepolia.service
Advanced Profile (Plugins enabled, needed for linea_estimateGas): - Mainnet: linea-advanced.service - Sepolia: linea-advanced-sepolia.service

Maru Configuration:

Maru is required for the consensus layer. Ensure it is configured for the same network as Besu:

Mainnet: maru.service (Default)
Sepolia: maru-sepolia.service

For deeper configuration details, see the package pages:

Linea
Maru

Running Layer 2 nodes¶

Quick Navigation¶

Hardware Requirements¶

Fuel Network Production Ready Rust¶

EthRex L2 Alpha Rust¶

Arbitrum Production Ready Go¶

Starknet Production Ready Rust¶

Available Clients¶

Juno¶

Madara¶

Configuration¶

Client Comparison¶

Optimism / Base Production Ready Go¶

Architecture¶

Quick Start¶

Linea¶

Troubleshooting¶