Comment 2 for bug 1755155

Hey John, thank you for taking a look.

We haven't upgraded the deployed model yet as we didn't expect the upgrade of the controller model to have this side effect of breaking the deployed model. As you can probably understand, we'd didn't want to push through an update while things are in an broken state.

Couple of quick questions:
1. How come none of the many other charms in the "openstack" model are complaining about this?
2. Is there a specific issue/fix related to your hunch? As I said earlier, we didn't see this in multiple rounds of staging which only has a few less machines/charms deployed.
3. Will the agents for nova-cloud-controller and nova-compute charms in question in the "openstack" model handle being upgraded even though they are experiencing hook failures?
4. In order to double check things, will the --dry-running flag do any useful prechecks in this case of upgrading a model?