Description
What happened:
Deploying a pod to an AKS Virtual Node, the pod is stuck in the Pending state if a container has:
env:
- name: SERVING_POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
I see the following error in the aci linux connector pod:
time="2021-07-05T22:34:00Z" level=warning msg="requeuing \"default/acitest-wv7lc\" due to failed sync"
error="failed to sync pod \"default/acitest-wv7lc\" in the provider:
unsupported fieldPath: status.podIP" key=default/acitest-wv7lc method=handleQueueItem node=virtual-node-aci-linux
operatingSystem=Linux provider=azure watchedNamespace= workerId=1
What you expected to happen:
For the pod to start up. Below is a simple repro yaml for a job that displays the behavior. You can comment out the SERVING_POD_IP variable and it starts as expected. But the real use-case for this is Knative launching on the Virtual Nodes and they require that variable be set.
How to reproduce it (as minimally and precisely as possible):
run: kubectl apply -f ./acitest.yaml
apiVersion: batch/v1
kind: Job
metadata:
name: acitest
spec:
backoffLimit: 1
template:
spec:
containers:
- name: ubuntu
image: ubuntu:latest
command: ["uname", "-a"]
resources:
limits:
cpu: 1
memory: 32Mi
requests:
cpu: 1
memory: 32Mi
env:
- name: SERVING_POD
valueFrom:
fieldRef:
apiVersion: v1
fieldPath: metadata.name
- name: SERVING_POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP
restartPolicy: Never
nodeSelector:
beta.kubernetes.io/os: linux
kubernetes.io/role: agent
type: virtual-kubelet
tolerations:
- key: virtual-kubelet.io/provider
operator: Exists
- key: node.kubernetes.io/not-ready
operator: Exists
effect: NoExecute
tolerationSeconds: 300
- key: node.kubernetes.io/unreachable
operator: Exists
effect: NoExecute
tolerationSeconds: 300
- key: node.kubernetes.io/memory-pressure
operator: Exists
effect: NoSchedule
Anything else we need to know?:
Knative in front of Virtual Nodes seems like it would be a very popular/common model. I see another couple users with the same issue here and another issue in this repository that was auto-closed: #1139
Environment:
- Kubernetes version (use
kubectl version): 1.18.4 - 1.20.7 - Size of cluster (how many worker nodes are in the cluster?): 3 node, Virtual Nodes enabled
- General description of workloads in the cluster: Mixed dependencies + dotnet core components