Skip to content

Latest commit

 

History

History
109 lines (75 loc) · 2.91 KB

README.md

File metadata and controls

109 lines (75 loc) · 2.91 KB
Raw

Chain Gang

Chain Gang is a small in-process Node.js queue. It ensures a limit to the number of simultaneous tasks in action at any time.

INSTALL

npm install chain-gang

The npm install process is only verified to work in npm 1.0.x.

USAGE

First, set up a Chain Gang, and specify the number of workers.

var chainGang = require('chain-gang')
var chain = chainGang.create({workers: 3})

Now, write a function to do some work.

var task = function(job) {
  // do some work
  var err;
  try {
    // do some work
  catch(e) {
    err = e
  }
  job.finish(err, 1, 2, 3) // and call this when finished
}

You can add it add it to the Chain Gang and have it completed when there is an open slot.

chain.add(task);

You can also give the task a unique identifier. If you have multiple requests for the same task to complete, this ensures that the actual work is only done a single time.

chain.add(task, 'bar')
chain.add(task, 'bar')

Finally, you can specify the callback to run after the task completes.

chain.add(task, 'bar', function(err, a, b, c) {
  // this optional callback is called with err, and any other arguments
  // sent to worker.finish() above.
})

The Chain Gang instance also has a few higher level events so you can watch the status of jobs entering and exiting the chain.

chain.on('add', function(name) {
  console.log(name, "has been queued.")
})

chain.on('starting', function(name) {
  console.log(name, "has started running.")
})

chain.on('finished', function(err, name) {
  console.log(name, "has finished.  Error:", err)
})

chain.on('empty', function() {
  console.log('queue is empty')
})

chain.on('timeout', function(job) {
  console.log(job.name, 'timed out')
})

Use Case

Let's say you have an expensive child process to run when requests come in. If traffic is really busy, you don't want to spin up 50 of these at once. Also, you don't want identical child processes running. If multiple requests want the same thing, assume they'll all get notified when it's finished.

NOT TODO

  • Persistence
  • Single Process Durability

If you need anything more, use a real queue (like Resque) backed by a persistent data store.

Development

Run this to compile coffeescript to javascript as you go:

make dev

Hopefully tests are green on Travis-CI.

Note on Patches/Pull Requests

  • Fork the project.
  • Make your feature addition or bug fix.
  • Add tests for it. This is important so I don't break it in a future version unintentionally.
  • Commit, do not mess with version or history. (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)
  • Send me a pull request. Bonus points for topic branches.

Copyright

Copyright (c) 2010 rick. See LICENSE for details.