No ability to reuse existing Juju controller
Bug #2066542 reported by
Nobuto Murata
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Snap |
Triaged
|
Wishlist
|
Unassigned |
Bug Description
By reading the tutorial and the flow in the MAAS scenario:
https:/
it looks like there is no ability to use an existing Juju controller instead of bootstrapping a new one for Sunbeam.
A new set of controller consumes three additional nodes in MAAS. It would be nice if we have a way to specify an existing controller and create a new model for Sunbeam deployment.
Changed in snap-openstack: | |
status: | New → Triaged |
importance: | Undecided → Wishlist |
To post a comment you must log in.
I was going to suggest something like that too.
Also, in case you do "spend" three machines for that, it should be easy to put more stuff in those same machines like for example COS. Possibly other things that are not directly openstack-related.
This might imply subdividing those machines somehow (for example using lxd) instead of deploying juju directly on the host OS -- not sure if juju can deploy itself directly into a lxd of a machine it requested from maas for the bootstrap process. Actually you might be able to deploy juju to the host directly and still create more lxd containers on top of it for adding more stuff.
The rationale is that in general customers will have similarly sized machines for all nodes in the cluster and it will be a waste to spend some only with juju controllers.