Ethereum node

Steps to run an Ethereum node on WSL 2 that stores its blockchain data on a Windows mounted disk.

Items

  • 1TB SSD

  • USB 3.0 to 2.5” SATA cable

  • WSL 2

  • geth

  • Web3.py

geth

In wsl follow geth’s installation guide for Ubuntu.

Enable geth's personal package archive:
$ sudo add-apt-repository -y ppa:ethereum/ethereum

$ sudo apt-get update
$ sudo apt-get install ethereum

geth the go-ethereum command line interface is now available on your system in /usr/bin/ along with

  • abigen – package smart contracts

  • bootnode – find peers in private networks

  • clef – manage account operations

  • evm – debug EVM opcodes

  • puppeth – assemble and maintain private networks

  • rlpdump – recursive length prefix utility

More here.

/mnt/e

Configure geth to store blockchain data on the formatted SSD. The disk is located in the /mnt/ directory from wsl.

Move your node's data directory from wsl to windows.
$ mv ~/.ethereum/ /mnt/e/

Create a symlink in ~ targeting the new data dir on windows.
$ ln -s /mnt/e/.ethereum ~/.ethereum

Update geth's config file.
$ geth dumpconfig > mainnet.toml
$ vi mainnet.toml

Under [Node] edit:
Datadir = "/mnt/e/.ethereum"

Ready

Start syncing your node to the Ethereum mainnet.

$ geth --config mainnet.toml --ipcdisable --http

Without specifying the sync mode our node is the default type: Snap. Ethereum nodes communicate by JSON-RPC APIs, and geth supports multiple endpoints to transport RPC calls.

  • HTTP

  • WebSocket

  • Unix Domain Socket

For our node, we disabled the IPC server (Unix socket) and use an HTTP server listening on localhost:8545 (default) to talk to the node and, when synced, effectively the whole Ethereum network.

Final sync stats

There be no graphs or metrics here. I only synced in the evenings, starting and stopping the process by hand. 509 GB of block data (and counting) later we are synced. Power consumption for the initial download was ~30 watts greater than running the node while synced.

State

Watts*

Syncing

160

Synced

130

Idle PC

80

*Estimates are from a P3 Kill A Watt meter.

PGE’s monthly energy rates for my location:

Charge

$ per kWh

Usage

$0.06690

Transmission

$0.00243

Distrubtion

$0.04694

TOTAL

$0.11627

Assuming continuous uptime: 1 month * 30 days * 24 hours * 130 watts * 1 kWh / 1000 watts * $0.11627 = $10.88

Include a list of adjustments that add ~7.59% to the final amount.

$10.88 * 1.0759 = $11.71 monthly cost to run a node

Web3.py

We can interact with our local node using the Web3.py python library.

from web3 import Web3

web3 = Web3(Web3.HTTPProvider('http://127.0.0.1:8545'))
assert(web3.isConnected() == True)

d = web3.eth.syncing
print(d['currentBlock'])
print(d['highestBlock'])

14588491 // CurrentBlock is the point at which synchronisation has presently reached.
14588595 // HighestBlock is the latest known block number.

Acknowledgments

“Btw I’m using the 1tb ssd you gave me. Much thx I owe u one (tb) ;)”

“Yeee! Glad it’s of good use now. 💾”