This repo contains automation used to maintain a fleet of robots using neuromorphic sensors and their supporting infrastructure. It is built on top of the attribute-driven API cookbooks maintained by Meta/Facebook. For more information about this style of writing automation code, refer to the Philosophy ,Facebook Cookbooks Suite README and Compile Time vs Run Time, and APIs.
To use these cookbooks, make sure that boxcutter_ohai and boxcutter_init
cookbooks are first in the Chef run-list, then followed by whatever
cookbooks you need for your config. The first two cookbooks in the run-list
set up all the automatic and global attribute defaults used in the
automation.
"recipe[boxcutter_ohai],recipe[boxcutter_init],..."Also make sure that all the cookbooks in Facebook's chef-cookbooks
and this boxcutter-chef-cookbooks are in the cookbook path:
cookbook_path [ 'ssh://[email protected]:facebook/chef-cookbooks.git',
'ssh://[email protected]:boxcutter/boxcutter-chef-cookbooks.git' ]For the most part we use the kitchen-dokken driver, as it is by far the
fastest Test kitchen driver. In some instances, we also use the
kitchen-digitalocean and kitchen-aws drivers to test automation that
can't run inside Docker, like the Docker automation itself or automation
involving running Docker containers.
For kitchen-digitalocean:
export DIGITALOCEAN_ACCESS_TOKEN=<do_access_token>
curl \
-X GET https://api.digitalocean.com/v2/account/keys \
-H "Authorization: Bearer $DIGITALOCEAN_ACCESS_TOKEN"
export KITCHEN_YAML=kitchen_digitalocean.yml
export DIGITALOCEAN_SSH_KEY_IDS=41887654,41887653
- @jaymzh for open sourcing https://github.com/socallinuxexpo/scale-chef which has a great example for automating
chefctl(among other things) - @jaymzh and @bwann for showing me the Facebook Chef way