vmi start fails to start when both pods and vms label is opted in #140
Description
What happened:
when trying to run vm - getting the following error:
failed to create virtual machine pod: admission webhook "mutatepods.kubemacpool.io" denied the request: failed to allocate requested mac address
What you expected to happen:
the mutatepods should ignore this pod since it is controlled by kubevirt (aka this mac is already given to the vm instance). running the vm should be successful.
How to reproduce it (as minimally and precisely as possible):
- apply namespace with opt-in lables of vms and pods:
apiVersion: v1
kind: Namespace
metadata:
labels:
mutatepods.kubemacpool.io: allocateForAll
mutatevirtualmachines.kubemacpool.io: allocateForAll
name: kmp-opt-in-ns
- apply this NAD:
apiVersion: "k8s.cni.cncf.io/v1"
kind: NetworkAttachmentDefinition
metadata:
name: "kmp-opt-br"
namespace: "kmp-opt-in-ns"
annotations:
k8s.v1.cni.cncf.io/resourceName: "bridge.network.kubevirt.io/kmp-opt-br"
spec:
config: '{
"cniVersion": "0.3.1",
"name": "kmp-opt-br",
"plugins": [{
"type": "cnv-bridge",
"bridge": "kmp-opt-br"
}]
}'
- apply this vm:
apiVersion: kubevirt.io/v1alpha3
kind: VirtualMachine
metadata:
labels:
special: kmp-opt-vm
name: kmp-opt-vm
namespace: kmp-opt-in-ns
spec:
running: false
template:
metadata:
labels:
kubevirt.io/vm: vm-fedora
spec:
domain:
devices:
disks:
- disk:
bus: virtio
name: containerdisk
- disk:
bus: virtio
name: cloudinitdisk
interfaces:
- name: default
masquerade: {}
- name: kmp-opt-br
bridge: {}
rng: {}
machine:
type: ""
resources:
requests:
memory: 1024M
terminationGracePeriodSeconds: 0
networks:
- name: default
pod: {}
- multus:
networkName: kmp-opt-br
name: kmp-opt-br
volumes:
- containerDisk:
image: quay.io/redhat/cnv-tests-fedora-staging:31
name: containerdisk
- cloudInitNoCloud:
userData: |-
#cloud-config
password: fedora
chpasswd: { expire: False }
name: cloudinitdisk
- start vm
virtctl start kmp-opt-vm
Anything else we need to know?:
Note that this is not related to the opt-in functionality. on the contrary - the opt-in functionality gives us a workaround where we only use the vm label and then can avoid this.
Environment:
KUBEVIRT_PROVIDER=k8s-1.17