[chef] Re: Re: [chef-dev] Re: Re: Re: Re: Re: Re: Nodejs community CK mess


Chronological Thread 
  • From: Lamont Granquist < >
  • To:
  • Subject: [chef] Re: Re: [chef-dev] Re: Re: Re: Re: Re: Re: Nodejs community CK mess
  • Date: Mon, 21 Jul 2014 11:16:16 -0700


On 7/20/14, 9:46 PM, Julian C. Dunn wrote:
In reality, only people who want to share cookbooks among their peers
inside a company are likely to run their own Supermarket. And those
won't be public.
Last time we had this discussion someone (ranjib?) wanted to fork the community cookbooks to produce an Amazon-Linux-centric set of cookbooks. That's the way i would expect to see different sets of cookbooks arise. Its not just because one author had a difference of opinion over a single cookbook, but because there was a broader need that wasn't getting met well by the community cookbooks.

And we've got a huge problem in scoping the community cookbooks that I already see occurring. They try to be everything to everyone. So all the ubuntus and redhats and (ideally) solaris and everything else, with binary installs and source installs crossed into every way an admin could want to do everything, and the result it the "don't look at the community cookbooks they're not good examples" problem because there's so much complexity caused by every use case. Looking at the FreeBSD ports use case as the closest model they have a considerably easier problem because everything is getting compiled from source and everything is getting installed onto FreeBSD by definition. I'd like to see a narrower focus on the ChefInc set of cookbooks, but that also means that other areas of focus will necessarily be left outside of the scope of those cookbooks.

BUT, it sounds to me like if you solve the internal supermarket case that I think you'll solve the public supermarket case as well. The case of having an internal supermarket where your internal chef clients can point to either the community versions or the overridden internal cookbooks of the same name should solve the problems that I'm thinking of with multiple public supermarkets as well. So, I'm pretty happy seeing the discussion start based around the internal use of supermarket and the problems that need to get resolved there. That's the actual technical problem that needs to get solved in either case.



Archive powered by MHonArc 2.6.16.

§