How to set up Erigon, Erigon’s RPC and TrueBlocks (scrape and API) as services

Update 2022-09-10 (Be ready for the Merge)

Ethereum Foundation recently announced that “It is strongly advised to stay up-to-date on the latest available versions of clients being used. This can be done by continue monitoring client repositories, client mailing lists and client social media channels for updates.”

It is thus recommended to use Erigon v2022.09.01-alpha as of now. If you have previously been on Erigon beta you will have to re-sync the DB from scratch, otherwise it should be compatible.

Further, I got an email from Aristide Pottier that in the later version of Erigon the http server is activated by default and we have to deactivate it otherwise the RPC daemon won’t be able to deploy the http API at port 8045. Aristide provided the fix,

ExecStart=/bin/erigon --datadir=mainnet --private.api.addr=localhost:9090 --http=false

A big thank you for helping this post to be up to date!

Update 2022-07-13

Erigon has had several updates since this blog post was first written. The previous problems with the alpha branch seems to be resolved, and the latest beta release is depreciated. It is recommended to run the latest alpha release. When you change from beta to alpha you will have to re-sync the DB. Further, Erigon now has a built in RPC server. The standalone RPC daemon still works, but you need to change ports from was originally written in this post. I changed the ports in the systemd service for the RPC daemon from 9090 to 9094 in order for it to work.

Resources

It is highly recommended to read the documentation of Erigon and TrueBlocks. This blog post assumes that you have Erigon and TrueBlocks installed. These previous blog posts might be of interest:

Introduction

In this blog post the Erigon archive node, the Erigon RPC daemon, TrueBlocks’ index scraper and TrueBlocks’ API are set up as Systemd services.

Erigon is a resource efficient Ethereum archive node that you can run on a standard local machine if you have 2TB or more of storage. TrueBlocks is a software that indexes the historical data of an archive node such that you can “list every appearance of an address anywhere on the chain”.

Set up Erigon and Erigon’s RPC as Systemd services

One way to set up Erigon and its RPC daemon as services on Linux is via Systemd. Create the files erigon.service and erigon-rpc.service in /etc/systemd/system/. Set up respectively file as,

[Unit]
Description=Erigon Node
After=network.target network-online.target
Wants=network-online.target

[Service]
WorkingDirectory=/media/node/erigon/erigon/
ExecStart=/bin/erigon --datadir=mainnet --private.api.addr=localhost:9090
User=magnus
Restart=always
RestartSec=5s

# Output to syslog
StandardOutput=syslog
StandardError=syslog
#Change this to find app logs in /var/log/syslog
SyslogIdentifier=erigon

[Install]
WantedBy=multi-user.target

and

[Unit]
Description=Erigon RPC Daemon
After=erigon.service

[Service]
WorkingDirectory=/media/node/erigon/erigon/
ExecStart=/bin/rpcdaemon --datadir=mainnet --txpool.api.addr=localhost:9090
    +++--private.api.addr=localhost:9090 --http.api=eth,erigon,web3,net,
    +++debug,trace,txpool --ws
User=magnus
Restart=always
RestartSec=5s

# Output to syslog
StandardOutput=syslog
StandardError=syslog
#Change this to find app logs in /var/log/syslog
SyslogIdentifier=erigon-rpc

[Install]
WantedBy=multi-user.target

Note that the “trace” option is enabled for the RPC daemon.

Now start the services with sudo systemctl start erigon.service and sudo systemctl start erigon-rpc.service. The status of the services can be viewed with sudo systemctl status erigon and sudo systemctl status erigon-rpc. The logs of the services will appear in /var/log/syslog.

There are more options that might be useful depending on how the services are to be run. E.g., depending on what network manager you are using the “After=” value might be different. This could be important if you want to run sudo systemctl enable <service> such that the service is started upon booting.

Connect TrueBlocks to Erigon

In TrueBlocks’ config file ~/.local/share/trueblocks/trueblocks.toml the “rpcProvider” value point to the RPC endpoint and it by default set to “http://localhost:8545”. Thus, if you run a local Erigon node you do not need to change this value.

To see if the connection between Erigon and TrueBlocks is correctly established type chifra status --terse the output should be without errors and look something like,

TB_CONFIG_PATH: /home/magnus/.local/share/trueblocks/
TB_CACHE_PATH:  /media/node/trueblocks/cache/mainnet/
TB_INDEX_PATH:  /media/node/trueblocks/index/unchained/mainnet/
TB_CACHE_PATH: /media/node/trueblocks/cache/mainnet/
TB_INDEX_PATH: /media/node/trueblocks/index/unchained/mainnet/
2022/01/11 16:02:01 Client:       erigon/2021.11.3/linux-amd64/go1.17.5 (archive, tracing)
2022/01/11 16:02:01 TrueBlocks:   GHC-TrueBlocks//0.23.7-alpha-4b5fbdfe3-20220111 (eskey, no
pinkey)
2022/01/11 16:02:01 Config Path:  /home/magnus/.local/share/trueblocks/
2022/01/11 16:02:01 Cache Path:   /media/node/trueblocks/cache/
2022/01/11 16:02:01 Index Path:   /media/node/trueblocks/index/unchained/
2022/01/11 16:02:01 RPC Provider: http://localhost:8545
2022/01/11 16:02:01 Progress:     13984914, 13883211, 0, 13896584

In order for TrueBlocks to connect to the archive node both Erigon and its RPC daemon need to be running. If only the RPC client is running errors such as, transport: Error while dialing dial tcp [::1]:9090: connect: connection refused will appear.

Creating the index

The index is created with the chifra scrape command that has a few options, most notably the “sleep” and “block_cnt” flags which are set to 14 and 2000 by default. When creating the index from scratch it might be a good idea to have a low sleep value and a high block count value. Later when the index is up-to-date these values can be changed. To build the index,

chifra scrape --sleep 1 --block_cnt 20000 indexer

This requires that the connection between Erigon, Erigon’s RPC daemon and TrueBlocks is well established. If not error such as This tool requires Parity traces. Add [requires]\nparity=false to $CONFIG/blockScrape.toml to turn this restriction off. Quitting... might appear, in which TrueBlocks is not able to talk to the node correctly.

If the index will be used for historical tx analysis where an always updated index is not important it is fine to run the above command in the terminal. However, if you want a continuously updated index it is possible to create a service for this as well. First create a script that runs the chifra scrape command,

#!/bin/bash
export PATH="$PATH:/media/node/trueblocks/trueblocks-core/bin"
chifra scrape --sleep 300 indexer

--sleep is set to 300 such that the script scrapes every 5 minutes. A block is created every 14 seconds on average and when the index is up to date it is unnecessary to run scrape continuously. Make the script executable with chmod +x chifra-scrape.sh. Note that Sysetemd services are not run in user login, meaning that if you have just added the TrueBlocks path to your .profile or .bashrc exit code 127 (command not found) will be thrown if you don’t export the path in the script file.

Then create a service running the script /etc/systemd/system/trueblock-scrape.service,

[Unit]
Description=TrueBlocks Scrape
After=erigon-rpc.service
Requires=erigon-rpc.service

[Service]
WorkingDirectory=/home/magnus/opt/bin/
ExecStart=/home/magnus/opt/bin/chifra-scrape.sh
User=magnus
Restart=always
RestartSec=5s

# Output to syslog
StandardOutput=syslog
StandardError=syslog
#Change this to find app logs in /var/log/syslog
SyslogIdentifier=trueblocks-chifra-scrape

[Install]
WantedBy=multi-user.target

TrueBlocks’ API

The chifra command can be used directly in the terminal and in bash scripts. However, there is also a TrueBlocks API that allows you to easily interact with TrueBlocks trough making calls to localhost. chifra serve creates a REST JSON API for chifra commands through a flame server. chifra serve --help,

Purpose:
  Serve the TrueBlocks API using the flame server.

Usage:
  chifra serve [flags]

Flags:
  -p, --port string   specify the server's port (default ":8080")
  -x, --fmt string    export format, one of [none|json|txt|csv|api]
  -v, --verbose       enable verbose (increase detail with --log_level)
  -h, --help          display this help screen

Notes:
  - To start API open terminal window and run chifra serve.

By running chifra serve in a terminal it is possible to call the API and query data, e.g.,

curl "http://localhost:8080/blocks?blocks=100"

To create a Systemd service running the API as a daemon create a script chifra-serve.sh,

#!/bin/bash
export PATH="$PATH:/media/node/trueblocks/trueblocks-core/bin"
chifra serve

Make the script executable chmod +x chifra-serve.sh and create a service,

sudo vim /etc/systemd/system/trueblock-serve.service

[Unit]
Description=TrueBlocks serve API
After=erigon-rpc.service
Requires=erigon-rpc.service

[Service]
WorkingDirectory=/home/magnus/opt/bin/
ExecStart=/home/magnus/opt/bin/chifra-serve.sh
User=magnus
Restart=always
RestartSec=5s

# Output to syslog
StandardOutput=syslog
StandardError=syslog
#Change this to find app logs in /var/log/syslog
SyslogIdentifier=trueblocks-chifra-serve

[Install]
WantedBy=multi-user.target

Management of the services

All services can be stopped at any time an restarted in the same fashion as ctrl + c on the command line. E.g., chifra scrape will contiue from the last bloom filter and Erigon will start to sync from the last block.

Useful commands: