randomcamel

Chef is a mighty set of tools that can automate and verify your infrastructure, but its full power is a lot to digest. Most of us learn new tools by trying them out for an hour or two, and that's usually how we make a decision. Today I want to show how you can get real value out of Chef real fast, maybe on a Friday afternoon with a couple of hours set aside for experimenting, by replacing a shell script with a list of Chef declarations. Then we can run that list using chef-apply, which executes a Chef runlist in a single file.

A Brief Note

Before we get started, it's worth mentioning that even though I'm going to pretend Chef is a scripting language, a Chef runlist describes the desired final state of a system, rather than providing a list of actions to execute. In other words, each Chef resource triggers code like this:

if system state already matches what the resource describes
then
 don't do anything

randomcamel

#!/usr/bin/env ruby

require 'json'
require 'logger'
log = Logger.new(STDOUT)

TEST_MODE = ARGV[0] == "--test"

fields = %w{ current_humidity current_temp_f aqi pressure }

randomcamel

# I didn't write this, but it's too good not to share.
# See also: https://www.mcsweeneys.net/articles/heres-how-time-works-now.
march () {
	perl -e 'use Date::Parse; use POSIX; my @t = localtime; print strftime ("%a Mar ", @t) . int (1 + 0.5 + ((str2time (strftime ("%Y-%m-%d 3:00", @t)) - str2time ("2020-03-01 3:00")) /(60*60*24))) . strftime (" %X %Z 2020\n", @t);'
}

randomcamel

docker-sh () {
	dockershell=${2:-bash}
	cmd="docker exec -it $1 /bin/$dockershell"
	echo "running '$cmd' (-i -> interactive, -t -> allocate a tty)"
	eval $cmd
}

randomcamel

Keybase proof

I hereby claim:

• I am randomcamel on github.
• I am cdoherty (https://keybase.io/cdoherty) on keybase.
• I have a public key ASAFza9EFloxNSHqh-7qr53alNEQhr8gDAdEakMs6ELSDwo

To claim this, I am signing this object:

randomcamel

brew () {
  if [[ "x$1" == "xcask" && "x$2" == "xsearch" ]]; then
    echo '`brew cask search` is broken, translating for you...'
    shift
    /usr/local/bin/brew $*
   else
    /usr/local/bin/brew $*
  fi
}

randomcamel

#!/bin/sh

# uses rsync's hard-link feature to mimic the function and structure of Apple's Time Machine.

# my server's name is 'shiny', which I evidently used to mount on /backup.
TARGET=/backup/shiny

date=`date "+%Y-%m-%dT%H:%M:%S"`
rsync --exclude=.AppleDouble -aCPv --link-dest=$TARGET/current $HOME $TARGET/back-$date
cd $TARGET

randomcamel

1. know how much roadmap-project work people can do in a week. commonly it's 40-60% of their time. call that PR (for Project Ratio).

2. make a spreadsheet: [Story, Description, People, Weeks, Person-Weeks].

3. break the work down into comprehensible pieces, where you feel confident saying things like "this will take 2 people, working full-time, 3 weeks to implement." in particular, break down milestones that deliver value; where, if the business told you to do something else, that would be okay, and the work done wasn't wasted.

4. fill in your spreadsheet columns appropriately (say 2 people, 3 weeks, 6 Person-Weeks).

5. sum up the Person-Weeks. ("People" and "Weeks" are for later.)

randomcamel

This is a bit of history and a bit of a quirk in timing...

The history first...

Vagrant has all manner of provisioners effectively supporting any CM tool as well as supporting plugins to add more. When I started on Packer, I basically copied this model (nit: provisioners aren't pluggable anymore cause no one ever used that functionality but the interfaces are very similar and Packer also supports all manner of CM tool).

randomcamel

# Install prereqs with Homebrew:
#
# uninstall moreutils if installed:
# `brew uninstall moreutils`
#
# brew install wget
# brew install poppler
# brew install parallel
# brew install moreutils --without-parallel
set -e