Running a Local Multi-node Network
This guide describes how to run a local network with multiple validator nodes and validator fullnodes. You will use the Aptos Forge CLI for this.
The method described in this guide should be used only for test deployments of multi-node local networks. Do not use this guide for deploying in production environments. Currently this is the only guide for multi-node deployments.
For deploying a local network with a single node, see Running Local Testnet and Local testnet development flow.
Before you proceed
Make sure you cloned the Aptos source GitHub repo by following these steps:
- Clone the Aptos repo.
git clone https://github.com/aptos-labs/aptos-core.git
cd
into aptos-core directory.
cd aptos-core
- Run the scripts/dev_setup.sh Bash script as shown below. This will prepare your developer environment.
./scripts/dev_setup.sh
- Update your current shell environment.
source ~/.cargo/env
With your development environment ready, now you can proceed below.
Running multiple validators
To deploy multiple local validators, run:
cargo run -p forge-cli \
-- \
--suite "run_forever" \
--num-validators 4 test local-swarm
This will start a local network of 4 validators, each running in their own process. The network will run forever unless you manually terminate it.
The terminal output will display the locations of the validator files (for example, the genesis files, logs, node configurations, etc.) and the commands that were run to start each node. The process id (PID) of each node and server addresses (e.g., REST APIs) are also displayed when it starts. For example, if you run the above command you should see:
...
2022-09-01T15:41:27.228289Z [main] INFO crates/aptos-genesis/src/builder.rs:462 Building genesis with 4 validators. Directory of output: "/private/var/folders/dx/c0l2rrkn0656gfx6v5_dy_p80000gn/T/.tmpq9uPMJ"
...
2022-09-01T15:41:28.090606Z [main] INFO testsuite/forge/src/backend/local/swarm.rs:207 The root (or mint) key for the swarm is: 0xf9f...
...
2022-09-01T15:41:28.094800Z [main] INFO testsuite/forge/src/backend/local/node.rs:129 Started node 0 (PID: 78939) with command: ".../aptos-core/target/debug/aptos-node" "-f" "/private/var/folders/dx/c0l2rrkn0656gfx6v5_dy_p80000gn/T/.tmpq9uPMJ/0/node.yaml"
2022-09-01T15:41:28.094825Z [main] INFO testsuite/forge/src/backend/local/node.rs:137 Node 0: REST API is listening at: http://127.0.0.1:64566
2022-09-01T15:41:28.094838Z [main] INFO testsuite/forge/src/backend/local/node.rs:142 Node 0: Inspection service is listening at http://127.0.0.1:64568
...
Using the information from this output, you can stop a single node and restart
it. For example, to stop and restart the node 0
, execute the below commands:
kill -9 <Node 0 PID>
cargo run -p aptos-node \
-- \
-f <Location to the node 0 configuration file displayed above>
Faucet and minting
In addition, the terminal output also displays the root (or mint) key for the network. This allows you to run a local faucet and start minting test tokens in the network. For this, simply run the faucet command using the mint key and point it to the REST API of one of the nodes, for example, node 0
:
cargo run --bin aptos-faucet \
-- \
-c TESTING \
--mint-key <Root/mint key displayed above> \
-s <URL for node 0 REST API> \
-p 8081
The above command will run a faucet locally, listening on port 8081
. Using this faucet, you could then mint tokens to your test accounts, for example:
curl -X POST http://127.0.0.1:8081/mint\?amount\=<amount to mint>\&pub_key\=<public key to mint tokens to>\&return_txns\=true
01000000000000000000000000000000dd05a600000000000001e001a11ceb0b01000...
Validator fullnodes
To also run validator fullnodes inside the network, use the --num-validator-fullnodes
flag. For example:
cargo run -p forge-cli \
-- \
--suite "run_forever" \
--num-validators 3 \
--num-validator-fullnodes 1 test local-swarm
Additional usage
To see all tool usage options, run:
cargo run -p forge-cli --help