Bu sayfanın çevirisine yardım edin

🌏

Henüz çevirmediğimiz için bu sayfayı İngilizce olarak görüntülüyorsunuz. Bu içeriği çevirmemize yardım edin.

Bu sayfa tamamlanmadı. Konuyla ilgili bir uzmansanız, lütfen bu sayfayı düzenleyin ve bilgilerinizle doldurunuz.

Spin up your own Ethereum node

Son düzenleme: , Invalid DateTime
Edit page

Running your own node provides you various benefits, opens new possibilities, and helps to support the ecosystem. This page will guide you through spinning up your own node and taking part in validating Ethereum transactions.

Prerequisites

You should understand what an Ethereum node is and why you might want to run a client. This is covered in Nodes and clients.

Choosing an approach

The first step in spinning up your node is choosing your approach. You have to choose the client (the software), the environment, and the parameters you want to start with. See all the available Mainnet clients.

Client settings

Client implementations enable different sync modes and various other options. Sync modes represent different methods of downloading and validating blockchain data. Before starting the node, you should decide what network and sync mode to use. The most important things to consider is the disk space and sync time client will need.

All features and options can be found in documentation of each client. Various client configurations can be set by executing client with corresponding flags. You can get more information on flags from EthHub or client documentation. For testing purposes, you might prefer running client on one of testnet networks. See overview of supported networks.

Environment and hardware

Local or cloud

Ethereum clients are able to run on consumer grade computers and don't require special hardware like mining for example. Therefore you have various options for deploying based on your needs. To simplify let's think about running a node on both a local physical machine and a cloud server:

  • Cloud
    • Providers offer high server uptime, static public IP addresses
    • Getting dedicated or virtual server can be more comfortable than building your own
    • Trade off is trusting a third party - server provider
    • Because of required storage size for full node, price of a rented server might get high
  • Own hardware
    • More trustless and sovereign approach
    • One time investment
    • An option to buy preconfigured machines
    • You have to physically prepare, maintain, and potentially troubleshoot the machine

Both options have different advantages summed up above. If you are looking for a cloud solution, in addition to many traditional cloud computing providers, there are also services focused on deploying nodes, for example:

Hardware

However, a censorship-resistant, decentralized network should not rely on cloud providers. It's healthier for the ecosystem if you run your own node on hardware. The easiest options are preconfigured machines like:

Check the minimum and recommended disk space requirements for each client and sync mod. Generally, modest computing power should be enough. The problem is usually drive speed. During initial sync, Ethereum clients perform a lot of read/write operations, therefore SSD is strongly recommended. A client might not even be able to sync current state on HDD and get stuck a few blocks behind Mainnet. You can run most of the clients on a single board computer with ARM. You can also use the Ethbian operating system for Raspberry Pi 4. This lets you [run a client by flashing the SD card. Based on your software and the hardware choices, the initial synchronization time and storage requirements may vary. Be sure to check sync times and storage requirements. Also make sure your internet connection is not limited by a bandwidth cap. It's recommended to use an unmetered connection since initial sync and data broadcasted to the network could exceed your limit.

Operating system

All clients support major operating systems - Linux, MacOS, Windows. This means you can run nodes on regular desktop or server machines with the operating system (OS) that suits you the best. Make sure your OS is up to date to avoid potential issues and security vulnerabilities.

Spinning up the node

Getting the client software

First, download your preferred client software

You can simply download an executable application or installation package which suits your operating system and architecture. Always verify signatures and checksums of downloaded packages. Some clients also offer repositories for easier installation and updates. If you prefer, you can build from source. All of the clients are open source so you can build them from source code with the proper compiler.

Executable binaries for stable Mainnet client implementations can be downloaded from their release pages:

Starting the client

Before starting Ethereum client software, perform a last check that your environment is ready. For example, make sure:

  • There is enough disk space considering chosen network and sync mode.
  • Memory and CPU is not halted by other programs.
  • Operating system is updated to latest version.
  • System has correct time and date.
  • Your router and firewall accept connections on listening ports. By default Ethereum clients use a listener (TCP) port and a discovery (UDP) port, both on 30303 by default.

Run your client on a testnet first to help make sure everything is working correctly. Running a Geth light node should help. You need to declare any client settings that aren't default at the start. You can use flags or the config file to declare your preferred configuration. Check out your client's documentation for the specifics Client execution will initiate its core functions, chosen endpoints, and start looking for peers. After successfully discovering peers, the client starts synchronization. Current blockchain data will be available once the client is successfully synced to the current state.

Using the client

Clients offer RPC API endpoints that you can use to control the client and interact with the Ethereum network in various ways:

  • Manually calling them with a suitable protocol (e.g. using curl)
  • Attaching a provided console (e.g. geth attach)
  • Implementing them in applications

Different clients have different implementations of the RPC endpoints. But there is a standard JSON-RPC which you can use with every client. For an overview read the JSON-RPC docs. Applications that need information from the Ethereum network can use this RPC. For example, popular wallet MetaMask lets you run a local blockchain instance and connect to it.

Reaching RPC

The default port of JSON-RPC is 8545 but you can modify the ports of local endpoints in the config file. By default, the RPC interface is only reachable on the localhost of your computer. To make it remotely accessible, you might want to expose it to the public by changing the address to 0.0.0.0. This will make it reachable over local and public IP addresses. In most cases you'll also need to set up port forwarding on your router.

You should do this with caution as this will let anyone on the internet control your node. Malicious actors could access your node to bring down your system or steal your funds if you're using your client as a wallet.

A way around this is to prevent potentially harmful RPC methods from being modifiable. For example, with geth, you can declare modifiable methods with a flag: --http.api web3,eth,txpool.

You can also host access to your RPC interface by pointing service of web server, like Nginx, to your client's local address and port.

The most privacy-preserving and simple way to set up a publicly reachable endpoint, you can host it on your own Tor onion service. This will let you reach the RPC outside your local network without a static public IP address or opened ports. To do this:

  • Install tor
  • Edit torrc config to enable hidden service with address of your client's RPC address and port
  • Restart tor service

Once you restart Tor, you'll get hidden service keys and a hostname in your desired directory. From then, your RPC will be reachable on a .onion hostname.

Operating the node

You should regularly monitor your node to make sure it's running properly. You may need to do occasional maintenance.

Keeping node online

Your node doesn't have to be online nonstop but you should keep it online as much as possible to keep it in sync with the network. You can shut it down to restart it but keep in mind that:

  • Shutting down can take up to a few minutes if the recent state is still being written on disk.
  • Forced shut downs can damage the database.
  • Your client will go out of sync with the network and will need to resync when you restart it.

This doesn't apply on Eth2 validator nodes. Taking your node offline will affect all services dependent on it. If you are running a node for staking purposes you should try to minimize downtime as much as possible.

Creating client service

Consider creating a service to run your client automatically on startup. For example on Linux servers, good practice would be creating a service that executes the client with proper config, under user with limited privileges and automatically restarts.

Updating client

You need to keep your client software up-to-date with the latest security patches, features, and EIPs. Especially before hard forks, make sure you are running the correct client version.

Running additional services

Running your own node lets you use services that require direct access to Ethereum client RPC. These are services built on top of Ethereum like layer 2 solutions, Eth2 clients, and other Ethereum infrastructure.

Monitoring the node

"To properly monitor your node, consider collecting metrics. Clients provide metrics endpoints so you can get comprehensive data about your node. Use tools like InfluxDB or Prometheus to create databases which you can turn into visualizations and charts in software like Grafana. There are many setups for using this software and different Grafana dashboards for you to visualise your node and the network as a whole. As part of your monitoring, make sure to keep an eye on your machine's performance. During your node's initial sync, the client software may be very heavy on CPU and RAM. In addition to Grafana, you can use the tools your OS offers like htop or uptime to do this.

Further reading