On Thu, Apr 27, 2017 at 09:39:51AM +0300, Pavlo Shchelokovskyy wrote: > HI all, > from some conversations I had during Pike PTG and recently in IRC I > understood that the need for ironic resources in heat has arisen again. I > remember back when it was proposed for the first time there was some > opposition from ironic community (although I personally find it reasonable > to have those) but AFAIU this is no longer the case. > I would gladly revive old Steven Hardy patches [0] (have them starred on > Gerrit since then :) ) and make it happen if there are no objections. > I also see that the spec itself to this regard has been recently > re-proposed [1], so if someone is already working on those, I'm > volunteering to help with it with my both ironic and heat hats on :)
Feel free to revive my patches, or reuse any parts of them which may be useful. FWIW (as I think I mentioned in the spec review and some previous threads on this topic), I reached the conclusion that the really interesting thing about driving ironic via heat is not so much registering the inventory, but driving the workflow around the various node states. I personally thing mistral is a better fit for that, so although I'm fine with some Ironic heat resources landing, I think it'd be great if we could define the workflow to e.g do introspection, or deploy an image to some node via mistral, instead of hard-coded via a heat resource. I started that here, but never had time to finish it, any help welcome! :) https://review.openstack.org/#/c/313048/ Also, Giulio has started working on a Heat resource that can run mistral workflows, which could be combined with a workflow similar to the above to deploy nodes via heat. https://review.openstack.org/#/c/420664/ Perhaps those are some other ideas we can consider while also looking at any heat native plugins which may be useful. Thanks! Steve __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: [email protected]?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
