On Tue, Jul 5, 2022 at 10:48 AM Houston Putman wrote:
> That, or maybe the distributed update processor checks the state of the
> cluster on a failure and if the replica no longer exists, it acts
> accordingly.
>
That is what we're exploring.
> [Jan] I think it is better for the project to evolve and fix this
Glad to hear it; sorry for the confusion if I misunderstood your concerns!
Well in that case it sounds like there's general support for the idea
of broader changes to the v2 API, and no categorical objections
(albeit a few concer
