Skip to content

exonum/exonum-launcher

Repository files navigation

Exonum Dynamic Service Launcher

Status: Travis Build Status Quality Gate Status

Exonum Dynamic Service Launcher is a tool to manage the lifecycle of the Exonum services.

This tool is capable of deploying artifacts, managing running services, and changing the configuration of the Exonum blockchain.

Usage

usage: exonum_launcher [-h] -i INPUT [-r RUNTIMES [RUNTIMES ...]]
                       [--runtime-parsers RUNTIME_PARSERS [RUNTIME_PARSERS ...]]
                       [--instance-parsers INSTANCE_PARSERS [INSTANCE_PARSERS ...]]

Exonum service launcher

optional arguments:
  -h, --help            show this help message and exit
  -i INPUT, --input INPUT
                        A path to yaml input for service initialization
  -r RUNTIMES [RUNTIMES ...], --runtimes RUNTIMES [RUNTIMES ...]
                        Additional runtimes, e.g. `--runtimes java=1 python=2
                        wasm=3`
  --runtime-parsers RUNTIME_PARSERS [RUNTIME_PARSERS ...]
                        Runtime spec parsers, e.g. `--runtime-parsers
                        python=your_module.YourRuntimeSpecLoader` Values will
                        be imported and treated like SpecLoader, so ensure
                        that module with loader is in `sys.path`.
  --instance-parsers INSTANCE_PARSERS [INSTANCE_PARSERS ...]
                        Instance spec parsers, e.g. `--runtime-parsers
                        python=your_module.YourInstanceSpecLoader` Values will
                        be imported and treated like InstanceSpecLoader, so
                        ensure that module with loader is in `sys.path`.

So, if you want to run exonum-launcher with Rust runtime only and without custom artifact spec loaders, you can just use:

python3 -m exonum_launcher -i sample.yml

If you want to use exonum-launcher with Python runtime and Python runtime spec loader, the command will be:

python3 -m exonum_launcher --runtimes python=2 --runtime-parsers python=exonum_launcher.runtimes.python.PythonSpecLoader -i sample.yml

Example of expected yaml file:

networks:
  - host: "127.0.0.1"
    ssl: false
    public-api-port: 8080
    private-api-port: 8081

deadline_height: 10000

artifacts:
  cryptocurrency:
    runtime: rust
    name: "exonum-cryptocurrency-advanced:0.1.0"
    action: deploy
  
  # Example of artifact that should not be deployed
  example_artifact:
    runtime: rust
    name: "exonum-cryptocurrency-advanced:0.1.0"
    
instances:
  xnm-token:
    artifact: cryptocurrency
    action: start
    config: []
  nnm-token:
    artifact: "cryptocurrency"
    action: start
    config: []
  some-instance:
    # Since we will not deploy `example_artifact`, it is assumed that it is already deployed
    artifact: "example_artifact"
    action: start
    config:
      val_a: "123"
      val_b: 345

action field in the artifacts section can be one of the following:

  • deploy - to deploy an artifact.
  • unload - to unload deployed artifact.
  • none - to do nothing (default action).

action field in the instances section can be one of the following:

  • start - to start a new instance (default action);
  • config - to change a configuration of existing service (only in simple supervisor mode!);
  • stop - to stop a running service.
  • freeze - to freeze a running service.
  • resume - to resume a frozen or stopped service.

Important: if you have more than one validator in the network, ensure that connection data (networks section of the config) is specified for every validator.

Deploy&init process requires requests to be sent to each validator, so don't expect that transaction broadcast mechanism will work here.

If supervisor works in the simple mode, it's also possible to change consensus config by providing a consensus field in the config, for example:

consensus:
    validator_keys:
      - ["1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a1a",
         "2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b2b"]
    first_round_timeout: 100
    status_timeout: 100
    peers_timeout: 100
    txs_block_limit: 100
    max_message_len: 100
    min_propose_timeout: 100
    max_propose_timeout: 100
    propose_timeout_threshold: 100

Plugins

You can define custom runtimes and plugins in the config (so you won't have to provide them from command line):

runtimes:
  python: 2

plugins:
  runtime:
    python: "exonum_launcher.runtimes.python.PythonSpecLoader"
  artifact: {}

See samples folder for more examples.

Install

pip install exonum-launcher --no-binary=protobuf

License

Apache 2.0 - see LICENSE for more information.