This message was deleted.

Both clusters same k8s version? https://kubernetes.io/docs/reference/using-api/deprecation-guide/#customresourcedefinition-v122

I think I found a likely cause. It wasn't until rancher-backup v2.1.0-rc1 that the CRDs were saved with apiextensions.k8s.io/v1 instead of v1beta1. You can see the commit here. We have inherited a rancher installation on k3s 1.18.20 and we're trying to migrate to EKS 1.22. Based on what I see, the newest version of rancher-backup that I can run on the k3s cluster is 2.0.1 because changes in 2.1.0 are incompatible with 1.18.x. If we had finished this even 2 weeks ago we could have created a new EKS 1.21 and we likely could have restored the backup stored with apiVersion apiextensions.k8s.io/v1beta1. It looks like we just need to commit to upgrading the k3s cluster we're trying to move out of to 1.20 and then we can attempt a backup using rancher-backup 2.1.x series which exports all of the CRDs that are currently throwing errors with apiVersion apiextensions.k8s.io/v1 which will be restorable on the EKS 1.22 cluster we're moving in to. I know this is a messy picture and that there are risks lurking that mean we might still not be out of the woods, but that's the unfortunate lay of the land we have.