Skip to content

Latest commit

 

History

History
154 lines (124 loc) · 3.36 KB

telescope.md

File metadata and controls

154 lines (124 loc) · 3.36 KB
Raw

Ansible O'reilly Book will feature a cow!

Who is telescope

Participation media - big voting tv shows use a realtime participation method. e.g.

american idol
the voice

Sometimes it's overnight sometimes it's like 5 minutes They work with all the big media channels: abc fox cbs bbc disney mtv

Ryan seacrest says vote and america does it!

Background

Moved from bare metal to AWS Why choose ansible for automation How ansible solved our challenges

First love

6 racks across 2 buildings 2 sysadmins responsible for hardware and operations no env for environment to play :( hardware was restricting ability to grow long lead time to build/increase totla capacit

Maybe we should look for a new love

  • Hard to justify building up a datacenter
  • Telescope's case required signigicant burst capacity
  • business needs change rapidly

Lots of other options

  • puppet
  • chef
  • ansible
  • bconfig2
  • salt
  • cfengine
  • vagrant cobbler

AWS

Strong APIs

Can you make a build in one step?

quote from: Joel Spolsky, "the joel test: 12 steps to better code"

Puppet

Too Needy!

Chef

Hard to understand & complicated!

Ansible

Launch, Push, Scale(up,down) Tight AWS integration ec2-module - instances in 10 minutes load-balancers elastic IPs!

Enter Ian!

Ian Plans for boxes to go down at any time.

Playbook

  1. Create Instances
  • Install things
  • checkout code and build
  • pull artifacts in maven
Internal debate

Compile code at build or pull down a pre-build version

Monitoring

Ansible templates give flexibility with infelxible tools e.g. Zabbix for monitoring

Tags

Tags have been useful with: ppTier+environment use

Dynamic Inventory File

http://docs.ansible.com/intro_dynamic_inventory.html

Often a user of a configuration management system will want to keep inventory in a different software system. Ansible provides a basic text-based system as described in Inventory but what if you want to use something else?

What's left?

  • Chronos as a replacement for Chron
  • zabbix module for better automations
    • scaling down
    • adding JMX interfaces
    • eventually express all items/triggers from a dictionary
  • try out other providers
  • onboarding is easier since the architecture is clearly defined!

modules are fun and easy to write

http://docs.ansible.com/modules.html

Questions

Q: How to do you manage resources on AWS? Deletion? A: Security groups used Scaling down is interesting

Q: Inventory? A: We use ec2.py

Q: How long does a deploy take? A: About 30 minutes! Build process is slowest part of the deploy That's why pre-created artifacts may work well.

Q: Do you use AMI's A: Nah

Q: How do you test? A: Still Working on it. Verification tests: everything is up and working Benchmarks wanted: identify problems in the wild. No Benchmarks piece yet. They use Grinder. test strategies in ansible docs: http://docs.ansible.com/test_strategies.html

Q: Zero Downtime Deploy? A: They update one server at a time If any more that 10% of the builds fail, they deploy will be halted

Q: How do you deal with multiple zones? A: We don't want to manage AMI's We can't catch spikes with autoscale. We don't do a lot of autoscaling because spikes are clearly define. We do 2 definitions in 2 zones in a dc

Q: Continuous Integration? A: bamboo and ansible