Autoscaler may be able to do what I need and I hadn't notice it before. :-)
Is autoscaler likely going to be the best recommended way to manage orchestrator from k8s?
I was in search of something that might be named
cluster-api-provider-xenorchestra
but I don't know that that exists yet? Maybe I should make a forum post instead.
I'm unsure if xen can handle the full cluster-api I have only really used the most basic features, I want to run and automatically rebuild/scale vm nodes of talos that will automatically register themselves into the network, but i need a way to tell kubernetes to tell xen what to do.
The hope being that I could say, automatically destroy 8 unused 2-core nodes (or automatically destroy/shrink cluster consistently), so that I could quickly spin up a 16 vcpu node for a different purpose like a specific batch load triggers or something.
Hi! Is this functional? I saw placeholder and limited commits and the discussion here and so was unsure:
https://xcp-ng.org/forum/topic/1950/kubernetes-autoscaler-xen-api-xo-api https://xcp-ng.org/forum/topic/5939/how-to-kubernetes-on-xcp-ng-csi/6
Autoscaler may be able to do what I need and I hadn't notice it before. :-) Is autoscaler likely going to be the best recommended way to manage orchestrator from k8s?
I was in search of something that might be named
cluster-api-provider-xenorchestra
but I don't know that that exists yet? Maybe I should make a forum post instead.
Example: https://github.com/kubernetes-sigs/cluster-api-provider-vsphere https://github.com/openshift/cluster-api-provider-libvirt https://cluster-api.sigs.k8s.io/ http://packages.shapeblue.com/cluster-api-provider-cloudstack/images/xen/
I'm unsure if xen can handle the full cluster-api I have only really used the most basic features, I want to run and automatically rebuild/scale vm nodes of talos that will automatically register themselves into the network, but i need a way to tell kubernetes to tell xen what to do.
The hope being that I could say, automatically destroy 8 unused 2-core nodes (or automatically destroy/shrink cluster consistently), so that I could quickly spin up a 16 vcpu node for a different purpose like a specific batch load triggers or something.