Ethereum full node on VPS: Geth + Lighthouse, archive vs full

To run an Ethereum full node on a VPS in 2024-2025, you will need a server with at least 4 vCPUs, 16 GB RAM (32 GB recommended), and a fast NVMe drive with a capacity of 2 TB or more for a Full node, or 12 TB or more for an Archive node. The rental cost for a suitable VPS/VDS with these specifications starts at $60-90 per month, depending on the location and type of disk subsystem.

Ethereum Architecture After The Merge: Why Do You Need Two Clients?

Running an ethereum node vps today is fundamentally different from how it worked before 2022. After the network transitioned to the Proof-of-Stake algorithm (The Merge), the node ceased to be a single binary file. Now, a full node consists of two independent layers that must work in tandem: the Execution Layer (EL) and the Consensus Layer (CL).

Execution Layer node: Transaction Processing

The Execution layer node is responsible for executing smart contracts, processing transactions, and storing the current state of the blockchain (State). This is where clients like Geth, Nethermind, Erigon, or Reth operate. This layer interacts with users via the JSON-RPC API, allowing them to send transactions and query balances. However, without a connection to the consensus layer, this layer cannot know which block chain is the canonical one.

Consensus Layer: Beacon Node and Validation

The consensus layer (Beacon Node) manages the Proof-of-Stake logic. It monitors which blocks are recognized as valid and coordinates the work of validators. Clients like Lighthouse, Prysm, Teku, or Lodestar are used for this task. Communication between the EL and CL is carried out via the Engine API using the JWT (JSON Web Token) authentication protocol. If you plan to run a staking node ethereum, you will also need a third component — the Validator Client, which signs blocks with your private keys.

Choosing Hardware Specifications for an ethereum node vps

The main bottleneck when running an Ethereum node is the disk subsystem. Regular HDDs or even budget SATA SSDs cannot handle the load due to low IOPS (input/output operations per second). An NVMe drive is critical for stable synchronization and lag-free operation. When choosing an ethereum node vps, look for the following parameters:

It is important to consider that the volume of data in the Ethereum network grows daily. If you are installing an eth archive node, the database growth rate is about 150-200 GB per month. Unlike Bitcoin, where hardware requirements are significantly lower (you can read more about this in the article Bitcoin full node on VPS: pruned vs full), Ethereum requires a constant supply of free RAM for caching the state tree (Trie).

Full Node vs Archive Node: What's the Difference and Which to Choose?

The choice of node type directly affects your infrastructure costs. For most developers and stakers, a Full Node is sufficient; it stores only the current state and the last few hundred blocks, allowing for the verification of any new transactions.

Full Node Features and Snap Sync

A Full Node uses the Snap Sync synchronization mode (in Geth). This allows you to download the current state of the network in a few hours without recalculating the entire history from 2015. Currently, such a node takes up about 1.2–1.4 TB. If you are developing a dApp or running a staking node ethereum for personal use, this is the optimal choice. Saving on disk space allows you to use faster NVMe drives on smaller VPS plans.

Why Do You Need an eth archive node?

An Eth archive node stores state "snapshots" at every block. This allows for queries like "what was the balance of this address at block 5,000,000?". Such nodes are essential for analytical services, block explorers, and DeFi protocols for historical analysis. The data volume here exceeds 14 TB, making it impossible to run on standard VPS. Dedicated servers with large disk arrays are required here. To optimize costs for such projects, many consider moving from AWS to dedicated servers, as cloud providers charge a huge premium for storage volume.

Geth + Lighthouse Combo: Installation and Configuration

The geth lighthouse combination is considered the "gold standard" due to Geth's stability and Lighthouse's performance (written in Rust). Let's look at the basic setup process on Ubuntu 22.04.

Step 1: Preparing the Environment and JWT Token

For secure communication between the execution and consensus layers, you need to create a secret key:

sudo mkdir -p /var/lib/ethereum/jwt
openssl rand -hex 32 | sudo tee /var/lib/ethereum/jwt/jwt.hex > /dev/null
sudo chmod 644 /var/lib/ethereum/jwt/jwt.hex

Step 2: Configuring Geth (Execution Layer)

Running Geth with Engine API support and cache limiting to save RAM:

geth --mainnet \
--authrpc.addr localhost \
--authrpc.port 8551 \
--authrpc.vhosts localhost \
--authrpc.jwtsecret /var/lib/ethereum/jwt/jwt.hex \
--http \
--http.api eth,net,web3 \
--cache 4096 \
--datadir /var/lib/ethereum/geth

Step 3: Configuring Lighthouse (Consensus Layer)

Lighthouse will connect to Geth via port 8551. Note that the Beacon Node requires synchronization, which can take time unless you use Checkpoint Sync.

lighthouse bn \
--network mainnet \
--execution-endpoint http://localhost:8551 \
--execution-jwt /var/lib/ethereum/jwt/jwt.hex \
--datadir /var/lib/ethereum/lighthouse \
--http \
--checkpoint-sync-url https://mainnet.checkpoint.sigp.io

Using the --checkpoint-sync-url flag allows you to synchronize the Consensus Layer almost instantly by trusting a verified network state.

Alternative Clients: Erigon and Reth

While Geth dominates the market (over 70% of nodes), client diversification is important for network health. Additionally, alternative clients can be more efficient in certain scenarios.

• Erigon: A rewritten fork of Geth optimized for data storage. Erigon is the best choice if you need an eth archive node, as it uses a flat database architecture and takes up significantly less space for archive data than standard Geth.
• Reth: A new client from Paradigm, written in Rust. It shows phenomenal synchronization and request processing speeds. Reth is actively displacing other solutions in high-load RPC services.
• Nethermind: A client in .NET, often chosen by institutional stakers for its excellent support and Enterprise features.

Custom dashboards are often used to analyze the performance of such systems. If you are already using self-hosted analytics for your projects, integrating node metrics into Grafana is a logical step for monitoring server health.

How to Make Money with an Ethereum Node: MEV and RPC Sale

Running an ethereum node vps is not just an expense, but also an opportunity to generate income. There are three main ways to monetize your own node.

Running a Validator and MEV-Boost

If you have 32 ETH, you can run a staking node ethereum. In addition to the base staking reward (~3-4% APR), you can connect MEV-Boost. This is additional software (a relay) that allows your validator to propose blocks filled with the most profitable transactions from "searchers". This can increase staking yields by 10-50% during periods of high market volatility.

Selling RPC Access

Many developers do not want to maintain their own nodes and are willing to pay for API access. You can sell access to your node directly or through decentralized protocols like Pocket Network. Your node becomes part of a global network, and you receive tokens for each processed request. High availability (Uptime) and low latency (Latency) are critical for this, which is provided by a high-quality ethereum node vps.

Arbitrage and Frontrunning

Your own node gives you an advantage in the speed of receiving information about new transactions in the mempool. If you write bots for arbitrage on DEXs (Uniswap, Curve), having a local node reduces latency by milliseconds, which is often the deciding factor for a successful trade. In such cases, an execution layer node based on Reth is often chosen for maximum performance.

Node Security and Administration

An Ethereum node is a complex software system that requires attention to security. Since the node is open to the network, it can become a target for DDoS attacks or attempts to exploit vulnerabilities in clients.

1. Firewall (UFW): Close all ports except those that are necessary. For P2P interaction, 30303 (TCP/UDP) for EL and 9000 (TCP/UDP) for CL are usually required. API ports (8545, 8551) should never be accessible from the outside without a VPN or SSH tunnel.
2. Resource Monitoring: Ethereum clients can suddenly start consuming more RAM during chain reorganizations. Set up alerts for disk and memory usage. If you store critical data, you might consider using vector databases to index specific blockchain events.
3. Updates: Clients are updated frequently. Missing a critical update before a hard fork can result in your node ending up on the wrong branch of the blockchain.

Docker Compose is convenient for managing configurations and automating deployment. This allows you to quickly switch between client versions and easily scale infrastructure.

services:
  geth:
    image: ethereum/client-go:latest
    volumes:
      - /path/to/data:/data
    ports:
      - "30303:30303"
    command: --datadir /data --http --http.addr 0.0.0.0

  lighthouse:
    image: sigp/lighthouse:latest
    volumes:
      - /path/to/data-cl:/data
    command: lighthouse bn --network mainnet --execution-endpoint http://geth:8551

Conclusions

To successfully run an Ethereum node on a VPS in the current environment, you need to choose a server with an NVMe drive of at least 2 TB and at least 16 GB of RAM, using a combination of Geth and Lighthouse for maximum stability. If your goal is deep analytics, prepare to rent a dedicated server for an Archive Node with more than 12 TB, and to earn from staking, be sure to set up MEV-boost to maximize profits.