Labels are not synced when using --target-namespace, --fake-nodes=false and --sync-all-nodes=false

[olljanat]

It looks to be that combination of --target-namespace , --fake-nodes=false and --sync-all-nodes=true works fine but when I change to --sync-all-nodes=false then those nodes gets created without any labels at all.

[FabianKramm]

@olljanat thanks for creating this issue! Seems like there is a bug where vcluster cannot find the current pod because of the --target-namespace option which results in the node controller failing over and over and not updating the labels. We'll fix this for the next version

[olljanat]

Labels looks to be there now on v0.5.0-alpha.4 but I just noticed that a lot of other details are missing too.

Here is some comparisons with and without --target-namespace

Wide view:

$ kubectl get nodes -o wide
NAME            STATUS    ROLES    AGE   VERSION   INTERNAL-IP   EXTERNAL-IP   OS-IMAGE    KERNEL-VERSION   CONTAINER-RUNTIME
k8s-test        Unknown   <none>   95s             <none>        <none>        <unknown>   <unknown>        <unknown>

$ kubectl get nodes -o wide
NAME            STATUS   ROLES    AGE   VERSION          INTERNAL-IP    EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION     CONTAINER-RUNTIME
k8s-test        Ready    <none>   3s    v1.22.3+rke2r1   10.3.154.243   <none>        Ubuntu 20.04.3 LTS   5.4.0-88-generic   containerd://1.5.7-k3s2

YAML diff (image list was longer but other why full diff):

$ diff -u node-ok.yaml node_not_ok.yaml | more
--- node-ok.yaml        2021-12-01 15:10:51.733590463 +0000
+++ node_not_ok.yaml    2021-12-01 15:09:47.473682924 +0000
@@ -12,7 +12,7 @@
     rke2.io/node-config-hash: S5ZCQYHBQFTCWE257O3XJG2KSNASLRR2YUPTB2B5JWIIKSQYH5TA====
     rke2.io/node-env: '{}'
     volumes.kubernetes.io/controller-managed-attach-detach: "true"
-  creationTimestamp: "2021-12-01T15:08:43Z"
+  creationTimestamp: "2021-12-01T15:05:31Z"
   labels:
     beta.kubernetes.io/arch: amd64
     beta.kubernetes.io/instance-type: rke2
@@ -26,185 +26,24 @@
   name: k8s-test-1
-  resourceVersion: "354"
-  uid: 8a13a066-39cc-4308-aa46-c742cb1efa27
+  resourceVersion: "345"
+  uid: 0f869e7d-4584-4a4d-8300-d6a6926901e9
 spec:
-  podCIDR: 10.6.2.0/24
-  podCIDRs:
-  - 10.6.2.0/24
-  providerID: rke2://k8s-test-1
+  taints:
+  - effect: NoSchedule
+    key: node.kubernetes.io/not-ready
 status:
-  addresses:
-  - address: 10.3.154.243
-    type: InternalIP
-  allocatable:
-    cpu: 4120m
-    ephemeral-storage: "49255941901"
-    hugepages-1Gi: "0"
-    hugepages-2Mi: "0"
-    memory: 16159560Ki
-    pods: "110"
-  capacity:
-    cpu: "6"
-    ephemeral-storage: 50633164Ki
-    hugepages-1Gi: "0"
-    hugepages-2Mi: "0"
-    memory: 16343880Ki
-    pods: "110"
-  conditions:
-  - lastHeartbeatTime: "2021-11-30T14:17:24Z"
-    lastTransitionTime: "2021-11-30T14:17:24Z"
-    message: Calico is running on this node
-    reason: CalicoIsUp
-    status: "False"
-    type: NetworkUnavailable
-  - lastHeartbeatTime: "2021-12-01T15:08:12Z"
-    lastTransitionTime: "2021-11-30T14:17:08Z"
-    message: kubelet has sufficient memory available
-    reason: KubeletHasSufficientMemory
-    status: "False"
-    type: MemoryPressure
-  - lastHeartbeatTime: "2021-12-01T15:08:12Z"
-    lastTransitionTime: "2021-11-30T14:17:08Z"
-    message: kubelet has no disk pressure
-    reason: KubeletHasNoDiskPressure
-    status: "False"
-    type: DiskPressure
-  - lastHeartbeatTime: "2021-12-01T15:08:12Z"
-    lastTransitionTime: "2021-11-30T14:17:08Z"
-    message: kubelet has sufficient PID available
-    reason: KubeletHasSufficientPID
-    status: "False"
-    type: PIDPressure
-  - lastHeartbeatTime: "2021-12-01T15:08:12Z"
-    lastTransitionTime: "2021-11-30T14:17:29Z"
-    message: kubelet is posting ready status. AppArmor enabled
-    reason: KubeletReady
-    status: "True"
-    type: Ready
   daemonEndpoints:
     kubeletEndpoint:
-      Port: 10250
-  images:
-  - names:
-    - docker.io/rancher/nginx-ingress-controller@sha256:a2b197240b16b72ad784f5ec98b10d9d0b9d94ea6b60f207bae556fa475a9f03
-    - docker.io/rancher/nginx-ingress-controller:nginx-1.0.2-hardened1
-    sizeBytes: 228330259
+      Port: 0
   nodeInfo:
-    architecture: amd64
-    bootID: e5270bf2-3738-4e8c-9321-f4239469f7e7
-    containerRuntimeVersion: containerd://1.5.7-k3s2
-    kernelVersion: 5.4.0-88-generic
-    kubeProxyVersion: v1.22.3+rke2r1
-    kubeletVersion: v1.22.3+rke2r1
-    machineID: 99a0dff27e7b4ce182acef3e441a7aa7
-    operatingSystem: linux
-    osImage: Ubuntu 20.04.3 LTS
-    systemUUID: 99a0dff2-7e7b-4ce1-82ac-ef3e441a7aa7
-  volumesAttached:
-  - devicePath: ""
-    name: kubernetes.io/csi/driver.longhorn.io^pvc-193b53c4-58cb-438d-b8ad-3e833d410798
-  - devicePath: ""
-    name: kubernetes.io/csi/driver.longhorn.io^pvc-9b877180-7fb5-4fa1-a71a-5d3bc251036a
-  volumesInUse:
-  - kubernetes.io/csi/driver.longhorn.io^pvc-193b53c4-58cb-438d-b8ad-3e833d410798
-  - kubernetes.io/csi/driver.longhorn.io^pvc-9b877180-7fb5-4fa1-a71a-5d3bc251036a
+    architecture: ""
+    bootID: ""
+    containerRuntimeVersion: ""
+    kernelVersion: ""
+    kubeProxyVersion: ""
+    kubeletVersion: ""
+    machineID: ""
+    operatingSystem: ""
+    osImage: ""
+    systemUUID: ""

[FabianKramm]

@olljanat yes you are correct, but this should be fixed with v0.5.0-alpha.5

[olljanat]

Confirmed that works correctly with v0.5.0-alpha.5. Thanks 👍