eric

Update

You can now find this packaged up nicely in a rubygem as rollout-zk.

---

Zookeeper storage adapter for rollout

I've implemented a zookeeper-based storage adapter for [rollout][] that does not require any network roundtrips to check if a feature is active for a user.

perplexes

Buildpacks

Heroku uses buildpacks to compile your application into a slug that is used across dynos for scaling horizontally quickly. A slug is a tar.gz archive of your app’s repository with certain pre-deploy features baked into the filesystem. Since everything to run your application is included in the archive, scaling becomes a simple matter of transferring it to a dyno, unpacking, and running the appropriate process. This is how Heroku achieves scaling-by-moving-a-slider.

For example, the Ruby buildpack will:

• install ruby locally
• install the jvm/jruby (if you’re using it)
• install/run bundler and install your gems to Rails.root/vendor
• create your database.yml (which ends up reading from your app’s environment variables)

paneq

  if (-f $document_root/system/maintenance.html) {
    return 503;
  }

  error_page 503 @maintenance;
  location @maintenance {
    internal;
    if ($http_accept ~ json) {
      return 503 "{}";
    }

myronmarston

max_stack_frames = 500

TooManyStackFrames = Class.new(StandardError)

set_trace_func proc { |event, file, line, id, binding, classname|
  if event == "call" && caller.size >= max_stack_frames
    raise TooManyStackFrames, "Stack has exceeded #{max_stack_frames} frames"
  end
}

polarblau

$ = jQuery

$.fn.extend

  foobar: (method) ->

    settings =
      option1: true
      option2: false    
    

rwjblue

#Overview drip is an awesome command line tool that can be used to dramatically lower perceived JVM startup time. It does this by preloading an entirely new JVM process\instance and allowing you to simply use the preloaded environment. This has extraordinary results with jruby.

We reduced time to run rake environment from 13 seconds to a mere 3.5 seconds. This is actually at or near MRI 1.9.3p327 (with falcon patch) speeds!

Adding a few addition jruby options will reduce startup time even further (down to 1.69 seconds).

#Install Drip Install drip if you haven't already (see https://github.com/flatland/drip)

overplumbum

#!/bin/bash
set -ve
lxc-create -n vpn -t ubuntu

# ln -s /var/lib/lxc/vpn/config /etc/lxc/auto/vpn.conf
perl -i -ple 's/#lxc.aa_profile = unconfined/lxc.aa_profile = unconfined/' /etc/lxc/auto/vpn.conf

perl -i -ple 's/^exit 0/# exit 0/' /etc/rc.local
cat >>/etc/rc.local  <<hello
mkdir -p /var/lib/lxc/vpn/rootfs/dev/net/

raggi

class Safeware

  def initialize(app)
    @app = app
  end

  # This dup pattern is used frequently to avoid race conditions on state stored
  # inside this middleware. It's not foolproof, but if you're just using
  # single-reference instance variables (instance variables with primitive
  # values (not data structures)) then it works well.

errordeveloper

Using Unicorn with Upstart

This configuration works with Upstart on Ubuntu 12.04 LTS

The reason why it needs to be done this way (i.e. with the pre-start and post-stop stanzas), is because Upstart is unable to track whever Unicorn master process re-execs itself on hot deploys. One can use it without hot-deploys and run Unicorn in foreground also, it then only needs one exec stanza.

This presumes you are not using RVM, so no voodoo dances.

ku1ik

You don't need bundle exec, rbenv-bundler or other bundler integration scripts

You just need 2 things

Just prepend $PATH with .bundle/bin:

$ cat ~/.bashrc
...
export PATH=".bundle/bin:$PATH"

...