Comment 9 for bug 1645422

This is still an issue with Juju 2.7.8 and MAAS 2.8.2.

My occurrence is a transient MAAS failed deployment because of *reasons*, and I want Juju to retry so that I can get a working machine.

I see from bug discussion history that there is some disagreement about what retry-provisioning means or does, and I guess I'll add to the scale that to me I expected it to mean that Juju could re-use the machine slot it has in its model and either fill it with a new instance or reach out to maas and do a release+deploy dance with the instance ID it already has.

Right now nothing happens and there is zero feedback to the user.