Every summit there's at least one big thing that we somehow avoided talking about as a group. We had fantastic discussions but we somehow forgot to talk about this.
I'd like to get this done _soon_. To be clear, the plan has always been (I forgot this and had to be reminded, it has been so long): * Make the undercloud a scaled-out seed (so point Galera and Rabbit at the seed's Galera/Rabbit) * Once undercloud is up, shutdown seed and remove references to it. The blocker for this is that if we then want to do an update of the undercloud, we can't touch Heat during said operation, because Heat would lose state when the undercloud node it is running on got rebooted. We have two ways to deal with that now: * Re-issue the update since we have update retry now * Use tripleo-ansible from outside of the undercloud itself or in a way in which it actively tries not to disrupt itself. I'll be working on this first by submitting a spec but I wanted to make sure any discussion of these points happens before I run off and try to write something. _______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev