Revert "Merge pull request #3312 from consideRatio/pr/test-upgrade-ku…

…be-scheduler"

This reverts commit 30a68f1, reversing
changes made to fe1251c.

.github/workflows/test-chart.yaml

@@ -160,7 +160,7 @@ jobs:
           # information from
           # https://hub.jupyter.org/helm-chart/info.json
           #
-          - k3s-channel: v1.26
+          - k3s-channel: v1.25
             test: upgrade
             upgrade-from: stable
             upgrade-from-extra-args: >-
@@ -173,7 +173,7 @@ jobs:
               --set hub.db.type=sqlite-pvc
               --set singleuser.storage.type=dynamic
             create-k8s-test-resources: true
-          - k3s-channel: v1.25
+          - k3s-channel: v1.24
             test: upgrade
             upgrade-from: dev
             upgrade-from-extra-args: >-
@@ -183,7 +183,7 @@ jobs:
             local-chart-extra-args: >-
               --set hub.db.type=sqlite-pvc
               --set singleuser.storage.type=dynamic
-          - k3s-channel: v1.24
+          - k3s-channel: v1.23
             test: upgrade
             # We're testing hub.db.upgrade with PostgreSQL so this version must be old
             # enough to require a DB upgrade

jupyterhub/Chart.yaml

@@ -8,7 +8,7 @@ keywords: [jupyter, jupyterhub, z2jh]
 home: https://z2jh.jupyter.org
 sources: [https://github.com/jupyterhub/zero-to-jupyterhub-k8s]
 icon: https://hub.jupyter.org/helm-chart/images/hublogo.svg
-kubeVersion: ">=1.24.0-0"
+kubeVersion: ">=1.23.0-0"
 maintainers:
   # Since it is a requirement of Artifact Hub to have specific maintainers
   # listed, we have added some below, but in practice the entire JupyterHub team

jupyterhub/templates/scheduling/user-scheduler/configmap.yaml

@@ -10,14 +10,26 @@ data:
     This is configuration of a k8s official kube-scheduler binary running in the
     user-scheduler.
 
+    The config version and kube-scheduler binary version has a fallback for k8s
+    clusters versioned v1.23 or lower because:
+
+    - v1 / v1beta3 config requires kube-scheduler binary >=1.25 / >=1.23
+    - kube-scheduler binary >=1.25 requires storage.k8s.io/v1/CSIStorageCapacity
+      available first in k8s >=1.24
+
     ref: https://kubernetes.io/docs/reference/scheduling/config/
     ref: https://kubernetes.io/docs/reference/config-api/kube-scheduler-config.v1/
+    ref: https://kubernetes.io/docs/reference/config-api/kube-scheduler-config.v1beta3/
   */}}
   config.yaml: |
+    {{- if semverCompare ">=1.24.0-0" .Capabilities.KubeVersion.Version }}
     apiVersion: kubescheduler.config.k8s.io/v1
+    {{- else }}
+    apiVersion: kubescheduler.config.k8s.io/v1beta3
+    {{- end }}
     kind: KubeSchedulerConfiguration
     leaderElection:
-      resourceLock: leases
+      resourceLock: endpointsleases
       resourceName: {{ include "jupyterhub.user-scheduler-lock.fullname" . }}
       resourceNamespace: "{{ .Release.Namespace }}"
     profiles:

jupyterhub/templates/scheduling/user-scheduler/deployment.yaml

@@ -50,7 +50,17 @@ spec:
       {{- end }}
       containers:
         - name: kube-scheduler
+          {{- if semverCompare ">=1.24.0-0" .Capabilities.KubeVersion.Version }}
           image: {{ .Values.scheduling.userScheduler.image.name }}:{{ .Values.scheduling.userScheduler.image.tag }}
+          {{- else }}
+          # WARNING: The tag of this image is hardcoded, and the
+          #          "scheduling.userScheduler.image.tag" configuration of the
+          #          Helm chart that generated this resource manifest isn't
+          #          respected. If you install the Helm chart in a k8s cluster
+          #          versioned 1.24 or higher, your configuration will be
+          #          respected.
+          image: {{ .Values.scheduling.userScheduler.image.name }}:v1.23.14
+          {{- end }}
           {{- with .Values.scheduling.userScheduler.image.pullPolicy }}
           imagePullPolicy: {{ . }}
           {{- end }}

jupyterhub/templates/scheduling/user-scheduler/rbac.yaml

@@ -20,11 +20,8 @@ rules:
   #       - changed in 1.21: get/list/watch permission for namespace,
   #                          csidrivers, csistoragecapacities was added.
   #       - unchanged between 1.22 and 1.27
-  #       - changed in 1.28: permissions to get/update lock endpoint resource
-  #                          removed
-  #       - unchanged between 1.28 and 1.29
   #
-  # ref: https://github.com/kubernetes/kubernetes/blob/v1.29.0/plugin/pkg/auth/authorizer/rbac/bootstrappolicy/testdata/cluster-roles.yaml#L721-L862
+  # ref: https://github.com/kubernetes/kubernetes/blob/v1.27.0/plugin/pkg/auth/authorizer/rbac/bootstrappolicy/testdata/cluster-roles.yaml#L736-L892
   - apiGroups:
     - ""
     - events.k8s.io
@@ -49,6 +46,21 @@ rules:
     verbs:
     - get
     - update
+  - apiGroups:
+    - ""
+    resources:
+    - endpoints
+    verbs:
+    - create
+  - apiGroups:
+    - ""
+    resourceNames:
+    - {{ include "jupyterhub.user-scheduler-lock.fullname" . }}
+    resources:
+    - endpoints
+    verbs:
+    - get
+    - update
   - apiGroups:
     - ""
     resources:
@@ -171,9 +183,9 @@ rules:
   # Copied from the system:volume-scheduler ClusterRole of the k8s version
   # matching the kube-scheduler binary we use.
   #
-  # NOTE: These rules have not changed between 1.12 and 1.29.
+  # NOTE: These rules have not changed between 1.12 and 1.27.
   #
-  # ref: https://github.com/kubernetes/kubernetes/blob/v1.29.0/plugin/pkg/auth/authorizer/rbac/bootstrappolicy/testdata/cluster-roles.yaml#L1283-L1310
+  # ref: https://github.com/kubernetes/kubernetes/blob/v1.27.0/plugin/pkg/auth/authorizer/rbac/bootstrappolicy/testdata/cluster-roles.yaml#L1311-L1338
   - apiGroups:
     - ""
     resources:

jupyterhub/values.yaml

@@ -485,8 +485,8 @@ scheduling:
       allowPrivilegeEscalation: false
     image:
       # IMPORTANT: Bumping the minor version of this binary should go hand in
-      #            hand with an inspection of the user-scheduelr's RBAC
-      #            resources that we have forked in
+      #            hand with an inspection of the user-scheduelrs RBAC resources
+      #            that we have forked in
       #            templates/scheduling/user-scheduler/rbac.yaml.
       #
       #            Debugging advice:
@@ -519,7 +519,7 @@ scheduling:
       # here. We aim to stay around 1 minor version behind the latest k8s
       # version.
       #
-      tag: "v1.28.8" # ref: https://github.com/kubernetes/kubernetes/tree/master/CHANGELOG
+      tag: "v1.26.15" # ref: https://github.com/kubernetes/kubernetes/tree/master/CHANGELOG
       pullPolicy:
       pullSecrets: []
     nodeSelector: {}