Avalanche CLI is a command line tool that gives developers access to everything Avalanche. This release specializes in helping developers develop and test subnets.

The tool has been tested on Linux and Mac. Windows is currently not supported.

To download a binary for the latest release, run:

curl -sSfL https://raw.githubusercontent.com/ava-labs/avalanche-cli/main/scripts/install.sh | sh -s

The binary will be installed inside the ~/bin directory.

To add the binary to your path, run

export PATH=~/bin:$PATH

To add it to your path permanently, add an export command to your shell initialization script (ex: .bashrc).

To download the binary into a specific directory, run:

curl -sSfL https://raw.githubusercontent.com/ava-labs/avalanche-cli/main/scripts/install.sh | sh -s -- -b <relative directory>

After installing, launch your own custom subnet:

avalanche subnet create <subnetName>
avalanche subnet deploy <subnetName>

Shut down your local deployment with:

avalanche network stop

Restart your local deployment (from where you left off) with:

avalanche network start

• Creation of Subnet-EVM, SpacesVM, and custom virtual machine subnet configurations
• Precompile integration and configuration
• Local deployment of subnets for development and rapid prototyping
• Fuji Testnet and Avalanche Mainnet deployment of subnets
• Ledger support
• Avalanche Package Manager Integration

You can provide a global node config to edit the way your local avalanchego nodes perform under the hood. To provide such a config, you need to create an avalanche-cli config file. By default, a config file is read in from $HOME/.avalanche-cli.json. If none exists, no error will occur. To provide a config from a custom location, run any command with the flag --config <pathToConfig>.

To specify the global node config, provide it as a body for the node-config key. Ex:

{
  "network-peer-list-gossip-frequency":"250ms",
  "network-max-reconnect-delay":"1s",
  "public-ip":"127.0.0.1",
  "health-check-frequency":"2s",
  "api-admin-enabled":true,
  "api-ipcs-enabled":true,
  "index-enabled":true
}

You may wish to deploy your subnet on a cloud instance and access it remotely. If you'd like to do so, use this as your node config:

{
  "node-config": {
    "http-host": "0.0.0.0"
  }
}

To build Avalanche-CLI, you'll first need to install golang. Follow the instructions here: https://go.dev/doc/install.

Once golang is installed, run:

./scripts/build.sh

The binary will be called ./bin/avalanche.

To make Avalanche CLI work in a docker container, add this

{
  "ipv6": true,
  "fixed-cidr-v6": "fd00::/80"
}

to /etc/docker/daemon.json on the host, then restart the docker service. This is because ipv6 is used to resolve local bootstrap IPs, and it is not enabled on a docker container by default.

To run our suite of end-to-end tests, you'll need to install Node-JS and yarn. You can follow instructions to do that here and here.

To run the tests, execute the following command from the repo's root directory:

./scripts/run.e2e.sh

Network snapshots are used by the CLI in order to keep track of blockchain state, and to improve performance of local deployments.

They are the main way to persist subnets, blockchains, and blockchain operations, among different executions of the tool.

Three different kinds of snapshots are used:

• The bootstrap snapshot is provided as the starting network state. It is never modified by CLI usage. Designed for fast deploys. Enables full reset of the blockchain state.
• The default snapshot is the main way to keep track of blockchain state. Used by default in the tools. It is initialized from the bootstrap snapshot, and after that is updated from CLI operations.
• custom snapshots can be specified by the user, to save and restore particular states. Only changed if explicitely asked to do so.

Usage of local networks:

• The local network will be started in the background only if it is not already running
• If the network is not running, both network start and subnet deploy will start it from the default snapshot. subnet deploy will also do the deploy on the started network.
• If the network is running, network start will do nothing, and subnet deploy will use the running one to do the deploy.
• The local network will run until calling network stop, network clean, or until machine reboot

How the CLI commands affect the default snapshot:

• First call of network start or subnet deploy will initialize default snapshot from the bootstrap snapshot
• Subsequent calls to subnet deploy do not change the snapshot, only the running network
• network stop persist the running network into the default snapshot
• network clean copy again the bootstrap snapshot into the default snapshot, doing a reset of the state

So typically a user will want to do the deploy she needs, change the blockchain state in a specific way, and after that execute network stop to preserve all the state. In a different session, network start or subnet deploy will recover that state.

How the CLI commands affect the custom snapshots:

• network stop can be given an optional snapshot name. This will then be used instead of the default one to save the state
• network start can be given an optional snapshot name. This will then be used instead of the default one to save the state
• subnet deploy will take a running network if it is available, so there is a need to use network start previously to do deploys, if wanting to use custom snapshots
• network clean does not change custom snapshots

So typically a user who wants to use a custom snapshot will do the deploy she needs, change the blockchain state in a specific way, and after that execute network stop with --snapshot-name flag to preserve all the state into the desired snapshot. In a different session, network start with --snapshot-name flag will be called to load that specific snapshot, and after that subnet deploy can be used on top of it. Notice that you need to continue giving --snapshot-name flag to those commands if you continue saving/restoring to it, if not, default snapshot will be used.

• ~/.avalanche-cli/snapshot will contain all saved snapshots, which can for example be used to pass work around

More detailed information on how to use Avalanche CLI can be found at here.