This repo aims at delivering deployment scripts for running prod-grade Starknet nodes.
The docker-compose.yml file embeds everything to run or deploy goerli and mainnet nodes in one command.
It uses env variables defined in the .env file that you need to create from the example.env and populate with Ethereum RPC endpoints.
cp example.env .env
# replace the value(s) of PATHFINDER_ETHEREUM_API_URL by the HTTP URL(s) pointing to your Ethereum node's endpointBy default, docker compose up will run both a mainnet and a goerli node. If you want to run only one service, you can specify it after the up:
docker compose up # run both mainnet and goerli
docker compose up starknet-mainnet # run mainnet only
docker compose up starknet-goerli # run goerli onlyTo retrieve the logs, use docker-compose logs -f.
To retrieve the URLs of the nodes, use docker compose ps.
The mainnet node runs on port 9546 while the goerli one runs on port 9545. You can check this by calling the starknet_chainId method:
curl '0.0.0.0:9545' \
-H 'content-type: application/json' \
--data-raw '{"method":"starknet_chainId","jsonrpc":"2.0","params":[],"id":0}' \
--compressed | jq .result | xxd -rp
# SN_GOERLI
curl '0.0.0.0:9546' \
-H 'content-type: application/json' \
--data-raw '{"method":"starknet_chainId","jsonrpc":"2.0","params":[],"id":0}' \
--compressed | jq .result | xxd -rp
# SN_MAIN
curl '0.0.0.0:9547' \
-H 'content-type: application/json' \
--data-raw '{"method":"starknet_chainId","jsonrpc":"2.0","params":[],"id":0}' \
--compressed | jq .result | xxd -rp
# SN_GOERLI2If you use the backup files, you can check the block position of your node:
curl <IP address> \
-H 'content-type: application/json' \
--data-raw '{"method":"starknet_syncing","jsonrpc":"2.0","params":[],"id":0}' \
--compressed | \
jq '.result | { starting_block_num,current_block_num,highest_block_num } | to_entries[] | .value' -r | \
xargs -L1 printf "%d\n" $1Docker has built-in integrations with AWS and Azure using docker context.
For these provider, the above-mentioned docker compose commands work out-of-the-box when no project-name is specified.
More details are given in the dedicated pages:
- for AWS: docs/aws/README.md
- for Azure: docs/azure/README.md
- for GCP: docs/gcp/README.md
The RPC node works with a built-in sqlite database that can be populated using daily snapshots stored in a S3 bucket.
Hence the container's command starts with a curl to the appropriate file. You can safely remove this command if you want to run the node from scratch or if you have another backup mechanism.