Hey John, We, the Berkshelf developers, actually strongly suggest against using roles and attempt to avoid data bags unless they really make sense (organization wide data). We also advocate for not versioning environments or nodes in a chef repo since both of those pieces are manipulated by our orchestrator. We hope to open this up pretty soon to the community. With Berkshelf and motherbrain we're trying to do away with the idea of a Chef Repo and instead provide tools which populate and manipulate the data in a Chef server. This will add an easier to use high level of abstraction on top of the typical Chef workflow which will hopefully lower the learning curve of getting into Chef. I think a Chef repo containing data not managed by a higher level abstraction is a good in between goal. Just make sure you immediately evict those cookbooks from it ;) Jamie Winsor @resetexistence
|
Archive powered by MHonArc 2.6.16.