fix qwen3.8 deployment

llamacpp/README.md

@@ -13,13 +13,13 @@ Ollama endpoint.
 |---|---|---|---|
 | `qwen3.8-27b` | Qwen3.8-27B with MTP | `Q4_K_M` primary, `Q4_0` draft, 196k context, q8_0 K/V cache | `llamacpp-qwen38-27b.llamacpp:80` |
 
-The active Deployment uses llama.cpp's Hugging Face downloader for both model
-repositories:
+An initContainer downloads both model files atomically before llama-server
+starts:
 
 - Primary: `ggml-org/Qwen3.8-27B-GGUF:Q4_K_M`
 - Draft: `ggml-org/Qwen3.8-27B-GGUF:Q4_0`
 
-The model cache is stored on the shared hostPath PVC at
+The model files are stored on the shared hostPath PVC at
 `/data/llamacpp/models` on the NUCBox. The server is configured with
 `--spec-default --spec-type draft-mtp`, `--reasoning-preserve`, `--fit off`,
 and `--agent`.

llamacpp/deployment-qwen38-27b.yaml

@@ -1,6 +1,6 @@
 # Qwen3.8-27B with MTP speculative decoding served by llama.cpp.
-# The primary and draft GGUFs are downloaded by llama-server into the shared
-# persistent llama.cpp cache on the NUCBox Radeon 8060S.
+# The primary and draft GGUFs are downloaded by an initContainer into the
+# shared model PVC before llama-server starts.
 ---
 apiVersion: apps/v1
 kind: Deployment
@@ -27,15 +27,41 @@ spec:
       nodeSelector:
         kubernetes.io/arch: amd64
         hardware: high-memory
+      initContainers:
+      - name: fetch-models
+        image: alpine:3.20
+        command: ["/bin/sh", "-c"]
+        args:
+        - |
+          set -eu
+          apk add --no-cache curl
+          for entry in \
+            "https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF/resolve/main/Qwen3.8-27B-Q4_K_M.gguf|Qwen3.8-27B-Q4_K_M.gguf" \
+            "https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF/resolve/main/mtp-Qwen3.8-27B-Q4_0.gguf|mtp-Qwen3.8-27B-Q4_0.gguf"; do
+            url=${entry%%|*}
+            file=${entry##*|}
+            if [ -s "/models/$file" ]; then
+              echo "$file already present - skipping download."
+              continue
+            fi
+            echo "Downloading $file ..."
+            curl -fL --retry 5 --retry-delay 5 -C - \
+              -o "/models/$file.partial" "$url"
+            mv "/models/$file.partial" "/models/$file"
+            echo "Download complete: $(ls -lh "/models/$file")"
+          done
+        volumeMounts:
+        - name: models
+          mountPath: /models
       containers:
       - name: llama-server
         image: ghcr.io/ggml-org/llama.cpp:server-vulkan
         imagePullPolicy: IfNotPresent
         args:
-        - -hf
-        - ggml-org/Qwen3.8-27B-GGUF:Q4_K_M
-        - -hfd
-        - ggml-org/Qwen3.8-27B-GGUF:Q4_0
+        - -m
+        - /models/Qwen3.8-27B-Q4_K_M.gguf
+        - --model-draft
+        - /models/mtp-Qwen3.8-27B-Q4_0.gguf
         - --spec-default
         - --spec-type
         - draft-mtp
@@ -79,14 +105,15 @@ spec:
           httpGet:
             path: /health
             port: 8080
-          initialDelaySeconds: 180
+          initialDelaySeconds: 600
           periodSeconds: 30
           failureThreshold: 5
         securityContext:
           privileged: true
         volumeMounts:
         - name: models
-          mountPath: /root/.cache/llama.cpp
+          mountPath: /models
+          readOnly: true
         - name: dri
           mountPath: /dev/dri
       volumes:

llamacpp/pv.yaml

@@ -10,7 +10,7 @@
 #
 # IMPORTANT: capacity is only metadata for a hostPath volume — k8s does NOT
 # enforce it and bumping it does NOT add physical disk space. The active Qwen
-# primary and draft GGUFs are downloaded into this cache, so the hostPath
+# primary and draft GGUFs are downloaded into this directory, so the hostPath
 # filesystem must have enough free space for both models.
 apiVersion: v1
 kind: PersistentVolume