gcp-cloudrun简单使用

发表于 分类于
演示 GCP Cloud Run 的简单用法：部署普通服务、部署 H100 GPU 任务，以及通过 YAML 挂载目录。
  1. 简单跑任务
1
2
3
4
5
6
7
8
gcloud run deploy my-nginx-service \
  --image nginx:latest \
  --platform managed \
  --region us-east4 \
  --allow-unauthenticated \
  --port 80 \
  --cpu 1 \
  --memory 256Mi
  1. 跑任务到 GPU 上
1
2
3
4
5
6
7
8
9
10
11
12
13
14
gcloud alpha run deploy comfyui \
--image us-east1-docker.pkg.dev/eng-spot-466907-i6/mytest-public/comfyui:v1 \
--concurrency 4 \
--cpu 20 \
--set-env-vars OLLAMA_NUM_PARALLEL=4 \
--gpu 1 \
--gpu-type nvidia-h100-80gb \
--max-instances 1 \
--memory 80Gi \
--no-allow-unauthenticated \
--no-cpu-throttling \
--no-gpu-zonal-redundancy \
--timeout=600 \
--startup-probe tcpSocket.port=8080,initialDelaySeconds=0,failureThreshold=10,timeoutSeconds=30,periodSeconds=60
  1. 挂载目录
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: comfyui
  labels:
    run.googleapis.com/satisfiesPzs: 'true'
    cloud.googleapis.com/location: us-east4
  annotations:
    run.googleapis.com/launch-stage: ALPHA
    run.googleapis.com/ingress: all
    run.googleapis.com/ingress-status: all
spec:
  template:
    metadata:
      labels:
        run.googleapis.com/startupProbeType: Default
      annotations:
        autoscaling.knative.dev/maxScale: '1'
        run.googleapis.com/client-name: gcloud
        run.googleapis.com/gpu-zonal-redundancy-disabled: 'true'
        run.googleapis.com/cpu-throttling: 'false'
        run.googleapis.com/startup-cpu-boost: 'true'
    spec:
      containerConcurrency: 4
      timeoutSeconds: 600
      containers:
      - image: >-
          us-east4-docker.pkg.dev/eng-spot-466907-i6/mytest-public/comfyui:v1
        command:
        - bash
        - '-c'
        - >-
          chmod +x /home/ns-comfyui-api/script/ComfyUI/start.sh; chmod +x
          /data/ns-comfyui-api/script/ComfyUI/start.sh; bash
          /data/ns-comfyui-api/script/ComfyUI/start.sh;
        ports:
        - name: http1
          containerPort: 12023
        env:
        - name: OLLAMA_NUM_PARALLEL
          value: '4'
        resources:
          limits:
            cpu: '20'
            nvidia.com/gpu: '1'
            memory: 80Gi
        volumeMounts:
        - name: gcs-1
          mountPath: /data
        startupProbe:
          timeoutSeconds: 240
          periodSeconds: 240
          failureThreshold: 1
          tcpSocket:
            port: 12023
      volumes:
      - name: gcs-1
        csi:
          driver: gcsfuse.run.googleapis.com
          volumeAttributes:
            bucketName: comfyui-data
            mountOptions: only-dir=Comfyui_Docker_Build
      nodeSelector:
        run.googleapis.com/accelerator: nvidia-h100-80gb
  traffic:
  - percent: 100
    latestRevision: true