Github user aledsage commented on the issue:
https://github.com/apache/brooklyn-server/pull/879
Regarding a primaryTransitioning sensor...
I agree that long-term it may not be the right way to indicate it, but I
don't think we're going to solve the rerun-failed-tasks in time for 1.0.0.
Separating the effectors into `ComputePrimaryEffector` and
`MakePrimaryEffector` would help for some of what you say: you could call
`makePrimary` again, without the second time being a no-op.
I think for making it clear to the user in v1.0.0, it would be most useful
to show a "primaryTransitioning" sensor that says "oldPrimary -> newPrimary".
If we fail half way through, it will be clearer what was happening
(particularly if we rebind, so the user can't explore the recent task history).
We wouldn't want someone to program against it (just use it for visual
inspection), but then the policy is `@Beta` so that should be fine.
---
