Rolling Updates

• Service Update default: replace each replica, one at a time
• We can customize a lot of that update process
• Before going live in production, test this a lot
• Different apps handle session and reconnection differently
• Few apps do this well, even with orchestrator help

There will be quite a lot of hands on and demo examples for which we need:

New Testing Tool: httping

• Test a HTTP(s) connection similar to how ping works
• Shows HTTP response code, colors, CLI GUI and more
• https://hub.docker.com/r/bretfisher/httping/
• We'll use container image from Bret Fisher, but also avail on
  • brew install httping
  • apt-get install httping
  • no easy build for Windows
• docker run bretfisher/httping localhost works everywhere

New Testing App: browncoat

• Simple web app with settings for acting badly
• https://hub.docker.com/r/bretfisher/browncoat/
• By default, no healthcheck, but functions correctly
• Different image tags and envvars for changing behavior
• Purpose: test rolling updates, rollbacks, and healthchecks

Requirements for the sub-section examples

• Created a 3-node (or more) Swarm cluster

• Created the swarm Visualizer service

  • docker run -it -d -p 5000:8080 -v /var/run/docker.sock:/var/run/docker.sock dockersamples/visualizer

• Visualizer looks like this:

External links

• https://docs.docker.com/engine/swarm/swarm-tutorial/rolling-update/