On 20/2/2024 12:43 am, Mohammed Naser wrote:
Thanks for your reply Jake.
However, merging a driver into Magnum does ship the message that it is the default.
Possibly what I am trying to do is to correct the message. Magnum had multiple heat drivers, going from fedora atomic to coreos to fedora coreos. IIRC, at some point, there was also talk of a heat Ubuntu, which if contributed and found working, could have been accepted. (sadly that did not materialise). So, in my opinion, there never was a 'default' driver.
Also, I can’t imagine a world where there’s two CAPI drivers inside of Magnum.
Why not? People move on. CAPI implementation differs. On this point there was someone asking if they want to implement a CAPI driver without a control cluster, would it be accepted? The answer is Yes.
I just think everything should stay out of tree till an implementation stands out to be “the one” then we can decide.
I think this may be the crux of the issue, there is no objective standard to determining 'the one'. I cannot, in good conscious, deny a commit from the community because another organisation have developed something. (This is my personal opinion, if OpenStack TC wants to rule on this I am willing to accept that) If the driver was 'official' from the ClusterAPI sub-project, then the conversation may be different. We may be able to treat it like third party drivers in Cinder, and point interested contributors to the effort made by the vendor. Regards, Jake