apps/base/jellyfin/deployment.yaml

@@ -0,0 +1,97 @@
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: jellyfin
+  namespace: jellyfin
+spec:
+  # Single-replica for now. The jellyfin-ha fork adds the Redis transcode store
+  # and RWX transcode scratch that make scaling to true HA a follow-up.
+  replicas: 1
+  strategy:
+    # Config PVC is RWO; Recreate avoids two pods contending for it.
+    type: Recreate
+  selector:
+    matchLabels:
+      app: jellyfin
+  template:
+    metadata:
+      labels:
+        app: jellyfin
+    spec:
+      securityContext:
+        fsGroup: 1000
+      nodeSelector:
+        feature.node.kubernetes.io/pci-0300_8086.present: "true"
+      containers:
+        - name: jellyfin
+          image: git.unkin.net/unkin/jellyfin-ha:v0.1.0
+          imagePullPolicy: IfNotPresent
+          ports:
+            - name: http
+              containerPort: 8096
+              protocol: TCP
+          env:
+            - name: TZ
+              value: Australia/Sydney
+            - name: PUID
+              value: "1000"
+            - name: PGID
+              value: "1000"
+            - name: JELLYFIN_PublishedServerUrl
+              value: https://jellyfin.k8s.syd1.au.unkin.net
+            # Distributed transcode session store (jellyfin-ha additions).
+            - name: Jellyfin__TranscodeStore__RedisConnectionString
+              value: "jellyfin-redis:6379,abortConnect=false"
+            - name: Jellyfin__TranscodeStore__LeaseDurationSeconds
+              value: "30"
+          resources:
+            requests:
+              cpu: 100m
+              memory: 1Gi
+            limits:
+              cpu: "4"
+              memory: 8Gi
+              gpu.intel.com/i915: "1"
+          livenessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 5
+            failureThreshold: 3
+          readinessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 10
+            periodSeconds: 10
+            timeoutSeconds: 5
+            failureThreshold: 3
+          volumeMounts:
+            - name: config
+              mountPath: /config
+            - name: cache
+              mountPath: /cache
+            - name: transcode
+              mountPath: /transcode
+            - name: media-library
+              mountPath: /mnt/movies
+              subPath: movies
+            - name: media-library
+              mountPath: /mnt/tvseries
+              subPath: tvseries
+      volumes:
+        - name: config
+          persistentVolumeClaim:
+            claimName: jellyfin-config
+        - name: cache
+          persistentVolumeClaim:
+            claimName: jellyfin-cache
+        - name: transcode
+          persistentVolumeClaim:
+            claimName: jellyfin-transcode
+        - name: media-library
+          persistentVolumeClaim:
+            claimName: media-library

apps/base/jellyfin/gateway.yaml

@@ -0,0 +1,37 @@
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: Gateway
+metadata:
+  labels:
+    traefik.io/instance: internal
+  annotations:
+    cert-manager.io/cluster-issuer: vault-issuer
+    cert-manager.io/common-name: jellyfin.k8s.syd1.au.unkin.net
+    cert-manager.io/private-key-size: "4096"
+    external-dns.alpha.kubernetes.io/hostname: jellyfin.k8s.syd1.au.unkin.net
+    external-dns.alpha.kubernetes.io/target: 198.18.200.4
+  name: jellyfin
+  namespace: jellyfin
+spec:
+  gatewayClassName: traefik-internal
+  listeners:
+    - allowedRoutes:
+        namespaces:
+          from: Same
+      hostname: jellyfin.k8s.syd1.au.unkin.net
+      name: http
+      port: 80
+      protocol: HTTP
+    - allowedRoutes:
+        namespaces:
+          from: Same
+      hostname: jellyfin.k8s.syd1.au.unkin.net
+      name: https
+      port: 443
+      protocol: HTTPS
+      tls:
+        certificateRefs:
+          - group: ""
+            kind: Secret
+            name: jellyfin-tls
+        mode: Terminate

apps/base/jellyfin/httproute.yaml

@@ -0,0 +1,49 @@
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: http-redirect
+  namespace: jellyfin
+spec:
+  hostnames:
+    - jellyfin.k8s.syd1.au.unkin.net
+  parentRefs:
+    - group: gateway.networking.k8s.io
+      kind: Gateway
+      name: jellyfin
+      sectionName: http
+  rules:
+    - filters:
+        - type: RequestRedirect
+          requestRedirect:
+            scheme: https
+            statusCode: 301
+      matches:
+        - path:
+            type: PathPrefix
+            value: /
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: jellyfin
+  namespace: jellyfin
+spec:
+  hostnames:
+    - jellyfin.k8s.syd1.au.unkin.net
+  parentRefs:
+    - group: gateway.networking.k8s.io
+      kind: Gateway
+      name: jellyfin
+      sectionName: https
+  rules:
+    - backendRefs:
+        - group: ""
+          kind: Service
+          name: jellyfin
+          port: 8096
+          weight: 1
+      matches:
+        - path:
+            type: PathPrefix
+            value: /

apps/base/jellyfin/kustomization.yaml

@@ -0,0 +1,18 @@
+---
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+
+resources:
+  - namespace.yaml
+  - pv_media-library.yaml
+  - pvc_media-library.yaml
+  - pvc_config.yaml
+  - pvc_cache.yaml
+  - pvc_transcode.yaml
+  - deployment.yaml
+  - service.yaml
+  - redis-deployment.yaml
+  - redis-service.yaml
+  - redis-pvc.yaml
+  - gateway.yaml
+  - httproute.yaml

apps/base/jellyfin/namespace.yaml

@@ -0,0 +1,5 @@
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: jellyfin

apps/base/jellyfin/pv_media-library.yaml

@@ -0,0 +1,41 @@
+---
+# Static CephFS PersistentVolume bound to the pre-existing, ACTIVELY-USED
+# puppet media library (ceph filesystem "mediafs", mounted by the VM/incus
+# instances at /shared/media). ceph-csi only mounts this volume; staticVolume
+# tells it the storage pre-exists and it must never provision or delete it.
+#
+# reclaimPolicy MUST stay Retain: deleting this PV or its PVC must NEVER be able
+# to reclaim or destroy the underlying CephFS data that the VM instances use.
+apiVersion: v1
+kind: PersistentVolume
+metadata:
+  name: jellyfin-media-library
+spec:
+  accessModes:
+    - ReadWriteMany
+  capacity:
+    storage: 10Ti
+  # Load-bearing safety control. Do not change to Delete.
+  persistentVolumeReclaimPolicy: Retain
+  storageClassName: ""
+  volumeMode: Filesystem
+  # Pre-bind to the media-library claim so nothing else can grab this PV.
+  claimRef:
+    apiVersion: v1
+    kind: PersistentVolumeClaim
+    name: media-library
+    namespace: jellyfin
+  csi:
+    driver: cephfs.csi.ceph.com
+    volumeHandle: jellyfin-media-library-static
+    nodeStageSecretRef:
+      name: csi-cephfs-secret
+      namespace: csi-cephfs
+    volumeAttributes:
+      # clusterID maps (in the csi-cephfs ceph-csi-config) to the mon set that
+      # also serves mediafs; for a static volume only the mon lookup is used.
+      clusterID: cephfs_csi_ssd_ec_4_1
+      fsName: mediafs
+      staticVolume: "true"
+      # Filesystem-internal root of the library (mediafs root == /shared/media).
+      rootPath: /

apps/base/jellyfin/pvc_cache.yaml

@@ -0,0 +1,15 @@
+---
+# Local transcode/image cache. Scratch, delete reclaim policy.
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: jellyfin-cache
+  namespace: jellyfin
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 200Gi
+  storageClassName: cephrbd-fast-delete
+  volumeMode: Filesystem

apps/base/jellyfin/pvc_config.yaml

@@ -0,0 +1,15 @@
+---
+# Jellyfin config + SQLite library database. Single-writer, block storage.
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: jellyfin-config
+  namespace: jellyfin
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 10Gi
+  storageClassName: cephrbd-fast-retain
+  volumeMode: Filesystem

apps/base/jellyfin/pvc_media-library.yaml

@@ -0,0 +1,19 @@
+---
+# Claim bound to the static mediafs PV. RWX so every app in the stack shares the
+# one library. storageClassName "" + volumeName pin it to the static PV (no
+# dynamic provisioning). Deleting this claim cannot reclaim the data (PV is
+# Retain).
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: media-library
+  namespace: jellyfin
+spec:
+  accessModes:
+    - ReadWriteMany
+  storageClassName: ""
+  volumeName: jellyfin-media-library
+  resources:
+    requests:
+      storage: 10Ti
+  volumeMode: Filesystem

apps/base/jellyfin/pvc_transcode.yaml

@@ -0,0 +1,17 @@
+---
+# Shared transcode scratch. ReadWriteMany is the hard requirement for the HA
+# fork: a taking-over pod must read the in-flight HLS segments written by the
+# pod it replaces. Scratch data, so delete reclaim policy.
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: jellyfin-transcode
+  namespace: jellyfin
+spec:
+  accessModes:
+    - ReadWriteMany
+  resources:
+    requests:
+      storage: 100Gi
+  storageClassName: cephfs-raid5-delete
+  volumeMode: Filesystem

apps/base/jellyfin/redis-deployment.yaml

@@ -0,0 +1,64 @@
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: jellyfin-redis
+  namespace: jellyfin
+spec:
+  replicas: 1
+  strategy:
+    type: Recreate
+  selector:
+    matchLabels:
+      app: jellyfin-redis
+  template:
+    metadata:
+      labels:
+        app: jellyfin-redis
+    spec:
+      restartPolicy: Always
+      containers:
+        - name: redis
+          image: redis:7-alpine
+          imagePullPolicy: IfNotPresent
+          command:
+            - redis-server
+            - --save
+            - "20"
+            - "1"
+          ports:
+            - name: redis
+              containerPort: 6379
+              protocol: TCP
+          livenessProbe:
+            exec:
+              command:
+                - redis-cli
+                - ping
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 5
+            failureThreshold: 3
+          readinessProbe:
+            exec:
+              command:
+                - redis-cli
+                - ping
+            initialDelaySeconds: 5
+            periodSeconds: 10
+            timeoutSeconds: 5
+            failureThreshold: 3
+          resources:
+            requests:
+              cpu: 50m
+              memory: 128Mi
+            limits:
+              cpu: 500m
+              memory: 512Mi
+          volumeMounts:
+            - name: data
+              mountPath: /data
+      volumes:
+        - name: data
+          persistentVolumeClaim:
+            claimName: jellyfin-redis-data

apps/base/jellyfin/redis-pvc.yaml

@@ -0,0 +1,14 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: jellyfin-redis-data
+  namespace: jellyfin
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 5Gi
+  storageClassName: cephrbd-fast-delete
+  volumeMode: Filesystem

apps/base/jellyfin/redis-service.yaml

@@ -0,0 +1,17 @@
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: jellyfin-redis
+  namespace: jellyfin
+spec:
+  type: ClusterIP
+  internalTrafficPolicy: Cluster
+  sessionAffinity: None
+  selector:
+    app: jellyfin-redis
+  ports:
+    - name: redis
+      port: 6379
+      targetPort: redis
+      protocol: TCP

apps/base/jellyfin/service.yaml

@@ -0,0 +1,17 @@
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: jellyfin
+  namespace: jellyfin
+spec:
+  type: ClusterIP
+  internalTrafficPolicy: Cluster
+  sessionAffinity: None
+  selector:
+    app: jellyfin
+  ports:
+    - name: http
+      port: 8096
+      targetPort: http
+      protocol: TCP

apps/overlays/au-syd1/jellyfin/kustomization.yaml

@@ -0,0 +1,6 @@
+---
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+
+resources:
+  - ../../../base/jellyfin

argocd/applicationsets/kustomization.yaml

@@ -5,6 +5,7 @@ kind: Kustomization
 resources:
   - aitooling.yaml
   - logging.yaml
+  - media.yaml
   - observability.yaml
   - platform.yaml
   - storage.yaml

argocd/applicationsets/media.yaml

@@ -0,0 +1,32 @@
+---
+apiVersion: argoproj.io/v1alpha1
+kind: ApplicationSet
+metadata:
+  name: media-apps
+  namespace: argocd
+spec:
+  generators:
+    - git:
+        repoURL: https://git.unkin.net/unkin/argocd-apps
+        revision: HEAD
+        directories:
+          # jellyfin only for now; downloads/managers apps join in later PRs.
+          - path: apps/overlays/*/jellyfin
+  template:
+    metadata:
+      name: 'media-{{path[3]}}'
+    spec:
+      project: media
+      source:
+        repoURL: https://git.unkin.net/unkin/argocd-apps
+        targetRevision: HEAD
+        path: '{{path}}'
+      destination:
+        server: https://kubernetes.default.svc
+        namespace: jellyfin
+      syncPolicy:
+        automated:
+          prune: true
+          selfHeal: true
+        syncOptions:
+          - ServerSideApply=true

argocd/projects/kustomization.yaml

@@ -5,6 +5,7 @@ kind: Kustomization
 resources:
   - aitooling.yaml
   - logging.yaml
+  - media.yaml
   - observability.yaml
   - platform.yaml
   - storage.yaml

argocd/projects/media.yaml

@@ -0,0 +1,21 @@
+---
+apiVersion: argoproj.io/v1alpha1
+kind: AppProject
+metadata:
+  name: media
+  namespace: argocd
+spec:
+  description: Media services (jellyfin; downloads/managers namespaces to follow)
+  sourceRepos:
+    - https://git.unkin.net/unkin/argocd-apps
+  destinations:
+    - namespace: jellyfin
+      server: https://kubernetes.default.svc
+  clusterResourceWhitelist:
+    - group: ''
+      kind: Namespace
+    - group: ''
+      kind: PersistentVolume
+  namespaceResourceWhitelist:
+    - group: '*'
+      kind: '*'