chore: import upstream snapshot with attribution

2026-07-13 12:24:33 +08:00
commit f213ec8976
2101 changed files with 494002 additions and 0 deletions
@@ -0,0 +1,52 @@
+---
+apiVersion: apps/v1
+kind: DaemonSet
+metadata:
+  name: lmcache-server
+  namespace: multi-process
+spec:
+  selector:
+    matchLabels:
+      app: lmcache-server
+  template:
+    metadata:
+      labels:
+        app: lmcache-server
+    spec:
+      hostNetwork: true
+      volumes:
+        - name: dev-shm
+          hostPath:
+            path: /dev/shm
+            type: Directory
+      containers:
+        - name: lmcache-server
+          image: lmcache/standalone:nightly
+          command:
+            - /opt/venv/bin/python3
+            - -m
+            - lmcache.v1.multiprocess.server
+            - --host
+            - "0.0.0.0"
+            - --port
+            - "6555"
+            - --l1-size-gb
+            - "60"
+            - --eviction-policy
+            - LRU
+            - --max-workers
+            - "4"
+          volumeMounts:
+            - name: dev-shm
+              mountPath: /dev/shm
+          resources:
+            requests:
+              # 65Gi = 60GB l1-size-gb + ~5GB overhead for Python runtime and server operations
+              memory: "65Gi"
+              cpu: "4"
+            limits:
+              memory: "120Gi"
+              cpu: "8"
+          env:
+            - name: LMCACHE_LOG_LEVEL
+              value: "DEBUG"
@@ -0,0 +1,63 @@
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: vllm-deployment
+  namespace: multi-process
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: vllm-deployment
+  template:
+    metadata:
+      labels:
+        app: vllm-deployment
+    spec:
+      volumes:
+        - name: dshm
+          hostPath:
+            path: /dev/shm
+            type: Directory
+      containers:
+        - name: vllm
+          volumeMounts:
+            - name: dshm
+              mountPath: /dev/shm
+          image: lmcache/vllm-openai:latest-nightly
+          command:
+            - /bin/bash
+            - -c
+            - |
+              vllm serve Qwen/Qwen3-14B \
+                --host 0.0.0.0 \
+                --port 8000 \
+                --no-enable-prefix-caching \
+                --max-model-len 32768 \
+                --gpu-memory-utilization 0.85 \
+                --tensor-parallel-size 4 \
+                --kv-transfer-config "{\"kv_connector\":\"LMCacheMPConnector\", \"kv_role\":\"kv_both\", \"kv_connector_extra_config\": {\"lmcache.mp.host\": \"tcp://${HOST_IP}\", \"lmcache.mp.port\": 6555}}"
+          env:
+            - name: HOST_IP
+              valueFrom:
+                fieldRef:
+                  fieldPath: status.hostIP
+            - name: PYTHONHASHSEED
+              value: "0"
+            - name: PROMETHEUS_MULTIPROC_DIR
+              value: "/tmp"
+          resources:
+            limits:
+              nvidia.com/gpu: "4"
+            requests:
+              nvidia.com/gpu: "4"
+          startupProbe:
+            failureThreshold: 60
+            httpGet:
+              path: /health
+              port: 8000
+              scheme: HTTP
+            initialDelaySeconds: 30
+            periodSeconds: 10
+            successThreshold: 1
+            timeoutSeconds: 1