No database provider would publish a new release without providing a linux installer or installation docs. Lately, some are even providing working VirtualBox and VMWare VMs. Will a Chef cookbook ever become just expected, as part of the mix for each new software release?
This to me, is a highly intriguing idea. Let's examine:
Pros
-
Would allow a somewhat plug and play setup for people to bring into their existing Chef architecture
-
Gives you a baseline on how to write your own cookbook for plug and play
-
Gives you a techniques on any technical challenges to installation (autoscaling)
-
More visible a construction compared to a sandbox
-
Would it solve the EULA checkbox problem with the license agreement?
Cons
-
Applications are going to be company specific, so might not cover all tuneables.
-
To be platform agnostic it might have the community cookbook problem.
-
Issues with assumption that it will work without tuning will hinder adoption.
-
No UI on tuning makes it harder for non-chefs to get it.
-
It's not a nice little pre-built sandbox.
-
Creates one more thing to go wrong, for the vendor to have to document and support.
-
It implies that you have to have versions for Salt, Ansible, Puppet or even CFEngine, which stretches a companies resources and or creates an arms race.
Or ...
Or shops could just engage third party shops such as my own to provide same. This would at once take them out of the loop for responsibility, and give them a higher chance of successfully configured installs, in their own user community.