node-amaster

Just a (cluster) master process for Node.js.

Install

npm install amaster

Usage

node node_modules/amaster [ path/to/config.json ] [ stop | reload | kill | clean ]

Default config: config.json.

Config

• log - path to log file. Default: null, stdout.

• errorLog - path to error. Default equals to log.

• pid - path to pid file. Required for stop reload kill clean commands.

• consoleMode - if set to true, process will not be daemonized. Default: false.

• cwd - working directory. Default: current dir.

• killTimeout - time in ms of waiting process termination before kill. Default: 5000.

• killCheckInterval - time in ms of checking interval for stop and kill commands. Default: 200.

• forkRateTimeout - time in ms to check fork rate. Default: 1000.

• forkRateLimit - max allowed fork count in forkRateTimeout interval. Default: null, workers count.

• user - worker user. Default: null, current user.

• group - worker group. Default: null, current group.

• worker - worker module. Default: worker (worker.js in cwd directory). Must export function:

    module.exports = function(options) { /* ... */ }

• workers - count of workers or workers options. Default: 1.

  You can specify array of options objects for each worker, these options will override user, group, worker or any additional options for every worker.

    {
        "worker": "myworker.js",
        "host": "localhost",
        "workers": [
            { "port": 8001 },
            { "port": 8002, "worker": "anotherworker.js" }
        ]
    }

  You can specify object with length property and any numeric properties (like an array) to override options in some workers, and use main options in others.

    {
        "worker": "myworker.js",
        "host": "localhost",
        "port": 8000,
        "workers": {
            "length": 5,
            "4": { "port": 8001, "worker": "anotherworker.js" }
        }
    }

  You can specify count of workers, in this case all workers will have same options.

    {
        "worker": "myworker.js",
        "host": "localhost",
        "port": 8000,
        "workers": 4
    }