Wikimedia Release Engineering Team/Train development environment/2020-06-18

From mediawiki.org

2020-06-18[edit]

As Always[edit]

TODOS[edit]

  • Yes Done TODO: Lars to create image for other to try
  • Yes Done TODO: Jeena to demo packer
  • Yes Done TODO: thcipriani to set next meeting
  • Yes Done TODO: Mukunda investigate cloud-init
  • Yes Done TODO: Mukunda + Lars to sync up on building a deployment server (puppet, etc.)
  • Yes Done TODO: Mukunda cloudinit to run puppet on vmdb2 image
  • Yes Done TODO: Mukunda + Lars pairing
  • TODO: Brennen to test Windows if possible

Notes[edit]

  • Config Management Options
    • Mukunda saves us
    • SRE saves us
    • We do something different
    • We could do something akin to mediawiki-vagrant: equivalent, but different
    • Would like to use a shim with ops puppet
  • Subteams
    • Puppet team to setup deployment-server-like VM
    • VM team setup a deploymet-server-like
      • GOAL: Run all the Tuesday steps of train without external access
      • Get a list of packages on the deployment server from deployment-deploy01.deployment-prep.eqiad.wmflabs:
   sudo cat /var/lib/puppet/client_data/catalog/deployment-deploy01.deployment-prep.eqiad.wmflabs.json | jq -r '.resources|.[] | select(.type=="Package").title'
  • TODO: Mukunda and Tyler to make puppet work or die trying
  • TODO: demo vmdb2 w/cloudinit + scap prep
  • TODO: lars to push WIP patch for train-dev
  • TODO: jeena + brennen to try to get running vm from train-dev repo
    • GOAL: run scap-prep inside the image
    • STRETCHGOAL: scap patch
  • TODO: tyler poke about ^ in 1:1s