> I really like #1 and #2. I'm not sure #3 is necessary since core devs / QA have historically tended to use AWS for testing
I agree that #3 might not be practical as an explicit goal, but if we accept it in the form of a constraint on our deployment architecture I think it will benefit us and the community. The self-hosting packaging is essentially a single server deployment. If we design the deployment to scale from 1-n nodes we get most of the way to an easy self hosting setup. The rest is just (optionally) wrapping up the config and multi-process complexity into a single config file and a parent process to manage all the others. The tidy packaging is the only bit that doesn’t directly benefit development because in dev we tend to want things unpacked for maximum flexibility. I think the bits that wrap everything up nicely should be a separate project but everything to make it possible to run on a single machine is already being done in fxa-dev (with the caveat that is has some non-essential Mozilla specific details that can be factored out) _______________________________________________ Dev-fxacct mailing list [email protected] https://mail.mozilla.org/listinfo/dev-fxacct
