On 06/04/14 23:27, Steve Baker wrote: > On 05/04/14 04:47, Tomas Sedovic wrote: >> Hi All, >> <snip> >> > The maintenance burden of merge.py can be gradually reduced if features > in it can be deleted when they are no longer needed. At some point in > this process merge.py will need to accept HOT templates, and risk of > breakage during this changeover would be reduced the smaller merge.py is. > > How about this for the task order? > 1. remove OpenStack::ImageBuilder::Elements support from merge.py > 2. move to software-config based templates > 3. remove the following from merge.py > 3.1. merging params and resources > 3.2. FileInclude > 3.3. OpenStack::Role > 4. port tripleo templates and merge.py to HOT > 5. use some HOT replacement for Merge::Map, delete Merge::Map from tripleo > 6. move to resource providers/scaling groups for scaling > 7. rm -f merge.py
I like this. Clint's already working on #2. I can tackle #1 and help review & test the software config changes. We can deal with the rest afterwards. One note on 3.1: until we switch to provider resources or get_file, we can't drop the "merging params and resources" feature. We can drop FileInclude, OpenStack::Role and deep merge (e.g. joining `notCompute0Config` from `overcloud-source.yaml` and `swift-source.yaml` example from my email), but we'll have to keep the functionality of putting multiple templates together for a bit longer. That said, I don't think switching to provider resources is going to be a drastic change. > > > _______________________________________________ > OpenStack-dev mailing list > OpenStack-dev@lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > _______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
