Jacob Gyllenstierna jgyllen

My notes from updating to El Capitan, as a Ruby/Rails developer working in VirtualBox/Vagrant.

VirtualBox and Vagrant

Our dev environment currently requires us to use older versions of VirtualBox (4.3) and Vagrant (1.7.1).

It seems the new "SIP" protection messes with that version of VirtualBox.

I tried upgrading VirtualBox (to 5.0.6) and Vagrant (to 1.7.4). Those seemed to run fine on El Capitan, but we rely on Chef v. 10 and Vagrant doesn't seem to get along with it (error: "invalid option: --force-formatter").

Falsehoods programmers believe about prices

You can store a price in a floating point variable.
All currencies are subdivided in 1/100th units (like US dollar/cents, euro/eurocents etc.).
All currencies are subdivided in decimal units (like dinar/fils)
All currencies currently in circulation are subdivided in decimal units. (to exclude shillings, pennies) (counter-example: MGA)
All currencies are subdivided. (counter-examples: KRW, COP, JPY... Or subdivisions can be deprecated.)
Prices can't have more precision than the smaller sub-unit of the currency. (e.g. gas prices)
For any currency you can have a price of 1. (ZWL)
Every country has its own currency. (EUR is the best example, but also Franc CFA, etc.)

Zero downtime deploys with unicorn + nginx + runit + rvm + chef

Below are the actual files we use in one of our latest production applications at Agora Games to achieve zero downtime deploys with unicorn. You've probably already read the GitHub blog post on Unicorn and would like to try zero downtime deploys for your application. I hope these files and notes help. I am happy to update these files or these notes if there are comments/questions. YMMV (of course).

Other application notes:

Our application uses MongoDB, so we don't have database migrations to worry about as with MySQL or postgresql. That does not mean that we won't have to worry about issues with the database with indexes being built in MongoDB or what have you.
We use capistrano for deployment.

Salient points for each file:

	#!/bin/bash

	# Usage: slackpost <token> <channel> <message>

	# Enter the name of your slack host here - the thing that appears in your URL:
	# https://slackhost.slack.com/
	slackhost=PUT_YOUR_HOST_HERE

	token=$1

	Latency Comparison Numbers (~2012)
	----------------------------------
	L1 cache reference 0.5 ns
	Branch mispredict 5 ns
	L2 cache reference 7 ns 14x L1 cache
	Mutex lock/unlock 25 ns
	Main memory reference 100 ns 20x L2 cache, 200x L1 cache
	Compress 1K bytes with Zippy 3,000 ns 3 us
	Send 1K bytes over 1 Gbps network 10,000 ns 10 us
	Read 4K randomly from SSD* 150,000 ns 150 us ~1GB/sec SSD

	class Settings < Settingslogic
	source "#{Rails.root}/config/application.yml"

	if Rails.env == 'development'
	namespace 'development-' + Socket.gethostname.gsub(/\W/, '-').gsub(/[-]{2,}/, '-')
	else
	namespace Rails.env
	end

	# load local settings

	# inspired by http://ariejan.net/2010/08/23/resque-how-to-requeue-failed-jobs

	# retry all failed Resque jobs except the ones that have already been retried
	# This is, for instance, useful if you have already retried some jobs via the web interface.
	Resque::Failure.count.times do \|i\|
	Resque::Failure.requeue(i) unless Resque::Failure.all(i, 1)['retried_at'].present?
	end

	# retry all :)
	Resque::Failure.count.times do \|i\|

	# When resque workers are no longer around, but records of them are still sitting around in Redis...

	def remove_resque_workers worker, ips
	ips.each do \|ip\|
	eval("
	def worker.to_s
	'#{ip}:*'
	end
	")
	worker.unregister_worker

	{ "keys": ["home"], "command": "move_to", "args": {"to": "bol"} },
	{ "keys": ["end"], "command": "move_to", "args": {"to": "eol"} }

	#trace mongo in RPM
	Mongo::Collection.class_eval do
	include NewRelic::Agent::MethodTracer
	add_method_tracer :find
	add_method_tracer :find_one
	add_method_tracer :find_and_modify
	add_method_tracer :update
	add_method_tracer :insert
	add_method_tracer :save
	add_method_tracer :remove