fix(longhorn): isolate GPU worker from general storage scheduling

- Remove expand-replicas-job (blindly forced all volumes to 3 replicas,
  ignoring StorageClass settings)
- Add diskSelector: 'storage' to longhorn and longhorn-cnpg StorageClasses
  so replicas only land on CP nodes (cp-1, cp-2, cp-3)
- Tag all CP node disks with 'storage' via PostSync job
  (disk names are runtime-discovered, can't hardcode in Node CRs)
- Disable scheduling on worker-1 Node CR — only longhorn-llm-local
  (diskSelector: 'llm') can use it
- worker-1 is GPU-only: llm-models and comfyui use dedicated SCs

k8s/infra/longhorn/longhorn-storageclass.yaml

@@ -6,13 +6,15 @@ metadata:
   name: longhorn
   annotations:
     storageclass.kubernetes.io/is-default-class: "true"
-    description: "Longhorn distributed storage - 3 replicas, WaitForFirstConsumer"
+    description: "Longhorn distributed storage - 3 replicas, CP nodes only"
+    argocd.argoproj.io/sync-options: Replace=true,Force=true
 provisioner: driver.longhorn.io
 allowVolumeExpansion: true
 reclaimPolicy: Delete
 volumeBindingMode: Immediate  # Immediate binding for StatefulSets (matches deployed config)
 parameters:
-  numberOfReplicas: "3"           # HA across all 3 nodes
+  numberOfReplicas: "3"           # HA across all 3 CP nodes
+  diskSelector: "storage"           # Only schedule on CP node disks, not GPU worker
   staleReplicaTimeout: "30"
   fromBackup: ""
   dataLocality: "disabled"        # Match deployed config (not best-effort)