PhanLe1010

I see. Thank you for the update. > So, in the case of a failed migration, would need to ensure the workload (VMI) fully scaled down before we could try...

> Step 1 is only to make volume get out of a weird state (migrationNodeID == new and currentMigrationNodeID == new, but nodeID == "")? Sorry for cutting the line...

`vol_data.json: no such file or directory` error seems to be related to the upstream Kubernetes issue https://github.com/kubernetes/kubernetes/issues/101378 Which Kubernetes version are you using?

> minio deployment is changed creating a new replicaset and a new pod Sorry, I don't quite understand how minio is used together with Longhorn here. Longhorn doesn't ship minio...

Do you mean the reproduce steps are? 1. Install Longhorn and create a minio deployment that use PVC provisioned by Longhorn 2. Update Longhorn to newer version using kubectl apply...

> i can only repro by re-running kubeadm init We are not very familiar this command so not sure what is the expected behavior

Could you provide your env info @ricardomaraschini > ## Environment > > - Longhorn version: > - Installation method (e.g. Rancher Catalog App/Helm/Kubectl): > - Kubernetes distro (e.g. RKE/K3s/EKS/OpenShift) and...