Frederic BOLTZ

Sometimes the build-essential package is not up to date. as the the message say: the version of the buildin kernel is not the same a thé build-essential package to build...

Probably you have reached the max cpu/memory limit by default with autoscaler. Add parameters in your manifest ``` - command: - ./cluster-autoscaler - --v=3 - ... - --max-nodes-total=100 - --cores-total=0:36...

I think that you reach the max limit pods:by default 110, it's a kubelet option. Your cluster is build with kubeadm or with k3?

rke2: but vmware autoscaler can pilot only cluster build with kubeadm or k3s. So the problem is in the autoscaler process won't autoscale. So I need the full log (verbose)...

I saw this: **(no node group config)** ``` I1211 15:54:12.004657 1 externalgrpc_cloud_provider.go:115] Performing gRPC call NodeGroupForNode for node enbitkub01 - I1211 15:54:12.005043 1 pre_filtering_processor.go:57] Node enbitkub01 should not be processed...

@Forestsoft-de About your deployment i just readed that rke2 use k3s, so it's means that you written a derived method to create the cluster. Do you expect to share it...

> Hey @Fred78290 its ok enbitkub01 shouldnt be managed by autoscaler. its one of our manual provisioned node and a master node also :) I used ansible to bootstrap the...

Thx. But The reason for any scale up is that your cluster created by ansible, nodes are not tagged to be member of node group expected by autoscaler. So when...

@Forestsoft-de Yo, I have integrated rke2 support into vmware-autoscaler. A docker image: [v1.27.11-rke2](https://hub.docker.com/layers/fred78290/vsphere-autoscaler/v1.27.11-rke2/images/sha256-0e39f4b8aa6b91272e3e330b79d1ef0575cb879ba2109c4735b468b5422019f2?context=repo) is available to test. In you case I have updated README with some informations to integrate foreign...

Argh, is a gRPC error. I have updated gRPC version in vmware-autoscaler and probably it's a compatibility error. I'll investigate. Can you repost the config & deployment updated.