Erigon is an implementation of Ethereum (aka "Ethereum client"), on the efficiency frontier, written in Go.
NB! In-depth links are marked by the microscope sign (🔬)
Disclaimer: this software is currently a tech preview. We will do our best to keep it stable and make no breaking changes but we don't guarantee anything. Things can and will break.
🔬 Alpha/Beta versions difference: here
-
For an Archive node of Ethereum Mainnet we recommend >=3TB storage space: 1.8TB state (as of March 2022), 200GB temp files (can symlink or mount folder
<datadir>/etl-tmp
to another disk). Ethereum Mainnet Full node (see--prune*
flags): 400Gb (April 2022). -
Goerli Full node (see
--prune*
flags): 189GB on Beta, 114GB on Alpha (April 2022). -
BSC Archive: 7TB. BSC Full: 1TB.
-
Polygon Mainnet Archive: 5TB. Polygon Mumbai Archive: 1TB.
SSD or NVMe. Do not recommend HDD - on HDD Erigon will always stay N blocks behind chain tip, but not fall behind. Bear in mind that SSD performance deteriorates when close to capacity.
RAM: >=16GB, 64-bit architecture, Golang version >= 1.18, GCC 10+
🔬 more details on disk storage here and here.
git clone --recurse-submodules -j8 https://github.com/ledgerwatch/erigon.git
cd erigon
make erigon
./build/bin/erigon
Default --snapshots
for mainnet
, goerli
, bsc
. Other networks now have default --snapshots=false
. Increase download speed by flag --torrent.download.rate=20mb
. 🔬 See Downloader docs
Use --datadir
to choose where to store data.
Use --chain=bor-mainnet
for Polygon Mainnet and --chain=mumbai
for Polygon Mumbai.
Running make help
will list and describe the convenience commands available in the Makefile
Erigon by default is "all in one binary" solution, but it's possible start TxPool as separated processes. Same true about: JSON RPC layer (RPCDaemon), p2p layer (Sentry), history download layer (Downloader), consensus. Don't start services as separated processes unless you have clear reason for it: resource limiting, scale, replace by your own implementation, security. How to start Erigon's services as separated processes, see in docker-compose.yml.
There is an optional stage that can be enabled through flags:
--watch-the-burn
, Enable WatchTheBurn stage which keeps track of ETH issuance and is required to useerigon_watchTheBurn
.
If you would like to give Erigon a try, but do not have spare 2TB on your drive, a good option is to start syncing one of the public testnets, Görli. It syncs much quicker, and does not take so much disk space:
git clone --recurse-submodules -j8 https://github.com/ledgerwatch/erigon.git
cd erigon
make erigon
./build/bin/erigon --datadir=<your_datadir> --chain=goerli
Please note the --datadir
option that allows you to store Erigon files in a non-default location, in this example,
in goerli
subdirectory of the current directory. Name of the directory --datadir
does not have to match the name of
the chain in --chain
.
Disclaimer: Not supported/tested for Polygon Network (In Progress)
Support only remote-miners.
- To enable, add
--mine --miner.etherbase=...
or--mine --miner.miner.sigkey=...
flags. - Other supported options:
--miner.extradata
,--miner.notify
,--miner.gaslimit
,--miner.gasprice
,--miner.gastarget
- JSON-RPC supports methods: eth_coinbase , eth_hashrate, eth_mining, eth_getWork, eth_submitWork, eth_submitHashrate
- JSON-RPC supports websocket methods: newPendingTransaction
- TODO:
- we don't broadcast mined blocks to p2p-network yet, but it's easy to accomplish
- eth_newPendingTransactionFilter
- eth_newBlockFilter
- eth_newFilter
- websocket Logs
🔬 Detailed mining explanation is here.
Windows users may run erigon in 3 possible ways:
-
Build executable binaries natively for Windows using provided
wmake.ps1
PowerShell script. Usage syntax is the same asmake
command so you have to run.\wmake.ps1 [-target] <targetname>
. Example:.\wmake.ps1 erigon
builds erigon executable. All binaries are placed in.\build\bin\
subfolder. There are some requirements for a successful native build on windows :- Git for Windows must be installed. If you're cloning this repository is very likely you already have it
- GO Programming Language must be installed. Minimum required version is 1.18
- GNU CC Compiler at least version 10 (is highly suggested that you install
chocolatey
package manager - see following point) - If you need to build MDBX tools (i.e.
.\wmake.ps1 db-tools
) then Chocolatey package manager for Windows must be installed. By Chocolatey you need to install the following components :cmake
,make
,mingw
bychoco install cmake make mingw
. Make sure Windows System "Path" variable has: C:\ProgramData\chocolatey\lib\mingw\tools\install\mingw64\bin
Important note about Anti-Viruses During MinGW's compiler detection phase some temporary executables are generated to test compiler capabilities. It's been reported some anti-virus programs detect those files as possibly infected by
Win64/Kryptic.CIS
trojan horse (or a variant of it). Although those are false positives we have no control over 100+ vendors of security products for Windows and their respective detection algorythms and we understand this might make your experience with Windows builds uncomfortable. To workaround the issue you might either set exclusions for your antivirus specifically forbuild\bin\mdbx\CMakeFiles
sub-folder of the cloned repo or you can run erigon using the following other two options -
Use Docker : see docker-compose.yml
-
Use WSL (Windows Subsystem for Linux) strictly on version 2. Under this option you can build Erigon just as you would on a regular Linux distribution. You can point your data also to any of the mounted Windows partitions ( eg.
/mnt/c/[...]
,/mnt/d/[...]
etc) but in such case be advised performance is impacted: this is due to the fact those mount points useDrvFS
which is a network file system and, additionally, MDBX locks the db for exclusive access which implies only one process at a time can access data. This has consequences on the running ofrpcdaemon
which has to be configured as Remote DB even if it is executed on the very same computer. If instead your data is hosted on the native Linux filesystem non limitations apply. Please also note the default WSL2 environment has its own IP address which does not match the one of the network interface of Windows host: take this into account when configuring NAT for port 30303 on your router.
Erigon can be used as an Execution Layer (EL) for Consensus Layer clients (CL). Default configuration is OK. CL
relies on availability of receipts – don't prune them: don't add character r
to --prune
flag. However, old receipts
are not needed for CL and you can safely prune them with --prune.r.before=<old block number>
in combination with --prune htc
.
If your CL client is on a different device, add --authrpc.addr 0.0.0.0
(Engine API listens on localhost by default)
as well as --authrpc.vhosts <CL host>
.
In order to establish a secure connection between the Consensus Layer and the Execution Layer, a JWT secret key is automatically generated.
The JWT secret key will be present in the datadir by default under the name of jwt.hex
and its path can be specified with the flag --authrpc.jwtsecret
.
This piece of info needs to be specified in the Consensus Layer as well in order to establish connection successfully. More information can be found here.
Once Erigon is running, you need to point your CL client to <erigon address>:8551
,
where <erigon address>
is either localhost
or the IP address of the device running Erigon, and also point to the JWT secret path created by Erigon.
Define 6 flags to avoid conflicts: --datadir --port --http.port --authrpc.port --torrent.port --private.api.addr
. Example of multiple chains on the same machine:
# mainnet
./build/bin/erigon --datadir="<your_mainnet_data_path>" --chain=mainnet --port=30303 --http.port=8545 --authrpc.port=8551 --torrent.port=42069 --private.api.addr=127.0.0.1:9090 --http --ws --http.api=eth,debug,net,trace,web3,erigon
# rinkeby
./build/bin/erigon --datadir="<your_rinkeby_data_path>" --chain=rinkeby --port=30304 --http.port=8546 --authrpc.port=8552 --torrent.port=42068 --private.api.addr=127.0.0.1:9091 --http --ws --http.api=eth,debug,net,trace,web3,erigon
Quote your path if it has spaces.
🔬 Detailed explanation is DEV_CHAIN.
🔬 See more
detailed overview of functionality and current limitations. It
is being updated on recurring basis.
Flat KV storage. Erigon uses a key-value database and storing accounts and storage in a simple way.
🔬 See our detailed DB walkthrough here.
Preprocessing. For some operations, Erigon uses temporary files to preprocess data before inserting it into the main DB. That reduces write amplification and DB inserts are orders of magnitude quicker.
🔬 See our detailed ETL explanation here.
Plain state.
Single accounts/state trie. Erigon uses a single Merkle trie for both accounts and the storage.
Erigon uses a rearchitected full sync algorithm from Go-Ethereum that is split into "stages".
🔬 See more detailed explanation in the Staged Sync Readme
It uses the same network primitives and is compatible with regular go-ethereum nodes that are using full sync, you do not need any special sync capabilities for Erigon to sync.
When reimagining the full sync, with focus on batching data together and minimize DB overwrites. That makes it possible to sync Ethereum mainnet in under 2 days if you have a fast enough network connection and an SSD drive.
Examples of stages are:
-
Downloading headers;
-
Downloading block bodies;
-
Recovering senders' addresses;
-
Executing blocks;
-
Validating root hashes and building intermediate hashes for the state Merkle trie;
-
[...]
Most of Erigon's components (sentry, txpool, snapshots downloader, can work inside Erigon and as independent process.
To enable built-in RPC server: --http
and --ws
(sharing same port with http)
Run RPCDaemon as separated process: this daemon can use local DB (with running Erigon or on snapshot of a database) or remote DB (run on another server). 🔬 See RPC-Daemon docs
This works regardless of whether RPC daemon is on the same computer with Erigon, or on a different one. They use TPC socket connection to pass data between them. To use this mode, run Erigon in one terminal window
make erigon
./build/bin/erigon --private.api.addr=localhost:9090 --http=false
make rpcdaemon
./build/bin/rpcdaemon --private.api.addr=localhost:9090 --http.api=eth,erigon,web3,net,debug,trace,txpool
9090
erigon, 9091
sentry, 9092
consensus engine, 9093
torrent downloader, 9094
transactions pool
Supported JSON-RPC calls (eth, debug , net, web3):
For a details on the implementation status of each command, see this table.
Docker allows for building and running Erigon via containers. This alleviates the need for installing build dependencies onto the host OS.
User UID/GID need to be synchronized between the host OS and container so files are written with correct permission.
You may wish to setup a dedicated user/group on the host OS, in which case the following make
targets are available.
# create "erigon" user
make user_linux
# or
make user_macos
There is a .env.example
file in the root of the repo.
DOCKER_UID
- The UID of the docker userDOCKER_GID
- The GID of the docker userXDG_DATA_HOME
- The data directory which will be mounted to the docker containers
If not specified, the UID/GID will use the current user.
A good choice for XDG_DATA_HOME
is to use the ~erigon/.ethereum
directory created by helper targets make user_linux
or make user_macos
.
In all cases, XDG_DATA_HOME
(specified or default) must be writeable by the user UID/GID in docker, which will be determined by the DOCKER_UID
and DOCKER_GID
at build time.
If a build or service startup is failing due to permissions, check that all the directories, UID, and GID controlled by these environment variables are correct.
Next command starts: Erigon on port 30303, rpcdaemon on port 8545, prometheus on port 9090, and grafana on port 3000.
#
# Will mount ~/.local/share/erigon to /home/erigon/.local/share/erigon inside container
#
make docker-compose
#
# or
#
# if you want to use a custom data directory
# or, if you want to use different uid/gid for a dedicated user
#
# To solve this, pass in the uid/gid parameters into the container.
#
# DOCKER_UID: the user id
# DOCKER_GID: the group id
# XDG_DATA_HOME: the data directory (default: ~/.local/share)
#
# Note: /preferred/data/folder must be read/writeable on host OS by user with UID/GID given
# if you followed above instructions
#
# Note: uid/gid syntax below will automatically use uid/gid of running user so this syntax
# is intended to be ran via the dedicated user setup earlier
#
DOCKER_UID=$(id -u) DOCKER_GID=$(id -g) XDG_DATA_HOME=/preferred/data/folder DOCKER_BUILDKIT=1 COMPOSE_DOCKER_CLI_BUILD=1 make docker-compose
#
# if you want to run the docker, but you are not logged in as the $ERIGON_USER
# then you'll need to adjust the syntax above to grab the correct uid/gid
#
# To run the command via another user, use
#
ERIGON_USER=erigon
sudo -u ${ERIGON_USER} DOCKER_UID=$(id -u ${ERIGON_USER}) DOCKER_GID=$(id -g ${ERIGON_USER}) XDG_DATA_HOME=~${ERIGON_USER}/.ethereum DOCKER_BUILDKIT=1 COMPOSE_DOCKER_CLI_BUILD=1 make docker-compose
Makefile creates the initial directories for erigon, prometheus and grafana. The PID namespace is shared between erigon and rpcdaemon which is required to open Erigon's DB from another process (RPCDaemon local-mode). See: https://github.com/ledgerwatch/erigon/pull/2392/files
If your docker installation requires the docker daemon to run as root (which is by default), you will need to prefix
the command above with sudo
. However, it is sometimes recommended running docker (and therefore its containers) as a
non-root user for security reasons. For more information about how to do this, refer to
this article.
Windows support for docker-compose is not ready yet. Please help us with .ps1 port.
docker-compose up prometheus grafana
, detailed docs.
Disabled by default. To enable see ./build/bin/erigon --help
for flags --prune
The ./docs
directory includes a lot of useful but outdated documentation. For code located
in the ./cmd
directory, their respective documentation can be found in ./cmd/*/README.md
.
A more recent collation of developments and happenings in Erigon can be found in the
Erigon Blog.
- Baseline (ext4 SSD): 16Gb RAM sync takes 6 days, 32Gb - 5 days, 64Gb - 4 days
- +1 day on "zfs compression=off". +2 days on "zfs compression=on" (2x compression ratio). +3 days on btrfs.
- -1 day on NVMe
Detailed explanation: ./docs/programmers_guide/db_faq.md
Port | Protocol | Purpose | Expose |
---|---|---|---|
30303 | TCP & UDP | eth/66 or 67 peering | Public |
9090 | TCP | gRPC Connections | Private |
42069 | TCP & UDP | Snap sync (Bittorrent) | Public |
6060 | TCP | Metrics or Pprof | Private |
8551 | TCP | Engine API (JWT auth) | Private |
Typically, 30303 is exposed to the internet to allow incoming peering connections. 9090 is exposed only internally for rpcdaemon or other connections, (e.g. rpcdaemon -> erigon). Port 8551 (JWT authenticated) is exposed only internally for Engine API JSON-RPC queries from the Consensus Layer node.
Port | Protocol | Purpose | Expose |
---|---|---|---|
8545 | TCP | HTTP & WebSockets | Private |
Typically, 8545 is exposed only internally for JSON-RPC queries. Both HTTP and WebSocket connections are on the same port.
Port | Protocol | Purpose | Expose |
---|---|---|---|
30303 | TCP & UDP | Peering | Public |
9091 | TCP | gRPC Connections | Private |
Typically, a sentry process will run one eth/xx protocol (e.g. eth/66) and will be exposed to the internet on 30303. Port 9091 is for internal gRCP connections (e.g erigon -> sentry).
Port | Protocol | Purpose | Expose |
---|---|---|---|
6060 | TCP | pprof | Private |
6060 | TCP | metrics | Private |
Optional flags can be enabled that enable pprof or metrics (or both) - however, they both run on 6060 by default, so
you'll have to change one if you want to run both at the same time. use --help
with the binary for more info.
Reserved for future use: gRPC ports: 9092
consensus engine, 9093
snapshot downloader, 9094
TxPool
- Get stack trace:
kill -SIGUSR1 <pid>
, get trace and stop:kill -6 <pid>
- Get CPU profiling: add
--pprof flag
rungo tool pprof -png http://127.0.0.1:6060/debug/pprof/profile\?seconds\=20 > cpu.png
- Get RAM profiling: add
--pprof flag
rungo tool pprof -inuse_space -png http://127.0.0.1:6060/debug/pprof/heap > mem.png
🔬 Detailed explanation is here.
Docker uses user erigon with UID/GID 1000 (for security reasons). You can see this user being created in the Dockerfile. Can fix by giving a host's user ownership of the folder, where the host's user UID/GID is the same as the docker's user UID/GID (1000). More details in post
https://github.com/mathMakesArt/Erigon-on-RPi-4
The main discussions are happening on our Discord server. To get an invite, send an email to tg [at] torquem.ch
with
your name, occupation, a brief explanation of why you want to join the Discord, and how you heard about Erigon.
Send an email to security [at] torquem.ch
.
Core contributors (in alpabetical order of first names):
-
Alex Sharov (AskAlexSharov)
-
Alexey Akhunov (@realLedgerwatch)
-
Andrea Lanfranchi(@AndreaLanfranchi)
-
Andrew Ashikhmin (yperbasis)
-
Artem Vorotnikov (vorot93)
-
Boris Petrov (b00ris)
-
Eugene Danilenko (JekaMas)
-
Igor Mandrigin (@mandrigin)
-
Giulio Rebuffo (Giulio2002)
-
Thomas Jay Rush (@tjayrush)
Thanks to:
-
All contributors of Erigon
-
All contributors of Go-Ethereum
-
Our special respect and graditude is to the core team of Go-Ethereum. Keep up the great job!
Happy testing! 🥤
Erigon's internal DB (MDBX) using MemoryMap
- when OS does manage all read, write, cache
operations instead of
Application
(linux
, windows)
htop
on column res
shows memory of "App + OS used to hold page cache for given App", but it's not informative,
because if htop
says that app using 90% of memory you still can run 3 more instances of app on the same machine -
because most of that 90%
is "OS pages cache".
OS automatically free this cache any time it needs memory. Smaller "page cache size" may not impact performance of
Erigon at all.
Next tools show correct memory usage of Erigon:
vmmap -summary PID | grep -i "Physical footprint"
. Withoutgrep
you can see detailssection MALLOC ZONE column Resident Size
shows App memory usage,section REGION TYPE column Resident Size
shows OS pages cache size.
Prometheus
dashboard shows memory of Go app without OS pages cache (make prometheus
, open in browserlocalhost:3000
, credentialsadmin/admin
)cat /proc/<PID>/smaps
Erigon uses ~4Gb of RAM during genesis sync and ~1Gb during normal work. OS pages cache can utilize unlimited amount of memory.
Warning: Multiple instances of Erigon on same machine will touch Disk concurrently, it impacts performance - one of main Erigon optimisations: "reduce Disk random access". "Blocks Execution stage" still does much random reads - this is reason why it's slowest stage. We do not recommend run multiple genesis syncs on same Disk. If genesis sync passed, then it's fine to run multiple Erigon on same Disk.
Please read erigontech#1516 (comment) In short: network-disks are bad for blocks execution - because blocks execution reading data from db non-parallel non-batched way.
For example: btrfs's autodefrag option - may increase write IO 100x times
Gnome Tracker - detecting miners and kill them.
For anyone else that was getting the BuildKit error when trying to start Erigon the old way you can use the below...
XDG_DATA_HOME=/preferred/data/folder DOCKER_BUILDKIT=1 COMPOSE_DOCKER_CLI_BUILD=1 make docker-compose