cap/docker-compose.yml.example at main · mobr-ai/cap · GitHub

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
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
services:
  api:
    build: .
    platform: linux/amd64
    container_name: cap_server
    volumes:
      - /home/cap/data/share-images:/var/lib/cap/share-images
    extra_hosts:
      - "host.docker.internal:host-gateway"
    ports:
      - "8000:8000"
    env_file:
      - .env
    command: >
      sh -c "sleep 15 && uvicorn src.cap.main:app --host 0.0.0.0 --port 8000"
    depends_on:
      jaeger:
        condition: service_started
      postgres:
        condition: service_healthy
      vllm:
        condition: service_healthy
    networks:
      - app-network

  postgres:
    image: postgres:16
    container_name: cap_postgres
    ports:
      - "5433:5432"   # host access (avoid clashing with any local 5432)
    env_file:
      - .env
    volumes:
      - /home/cap/data/postgres:/var/lib/postgresql/data
      - ./ops/sql/init-cap.sql:/docker-entrypoint-initdb.d/10-init-cap.sql:ro
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER} -d ${POSTGRES_DB:-postgres} -h localhost"]
      interval: 5s
      timeout: 3s
      retries: 20
    networks:
      - app-network

  redis:
    image: redis:7-alpine
    container_name: cap_redis
    ports:
      - "6379:6379"
    volumes:
      - /home/cap/data/redis:/data
    command: redis-server --appendonly yes
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 20
    networks:
      - app-network

  qlever:
    image: adfreiburg/qlever:latest
    platform: linux/amd64
    container_name: qlever
    ports:
      - "7001:7001"   # QLever UI
      - "7000:7000"   # SPARQL endpoint
    env_file:
      - .env
    user: "1000:1000"
    volumes:
      - /home/cap/data/qlever:/data
    working_dir: /data
    entrypoint: ["/bin/bash"]
    command: ["-c", "ServerMain -i cardano -p 7000 -a ${VIRTUOSO_PASSWORD} --persist-updates --memory-max-size 16g"]
    healthcheck:
      test: ["CMD-SHELL", "curl -fsS http://localhost:7000/ >/dev/null || exit 1"]
      interval: 10s
      timeout: 5s
      retries: 30
      start_period: 30s
    networks:
      - app-network

  jaeger:
    image: jaegertracing/all-in-one:latest
    platform: linux/amd64
    container_name: jaeger
    ports:
      - "16688:16686" # UI
      - "4319:4317"   # OTLP gRPC
      - "4322:4318"   # OTLP HTTP
    env_file:
      - .env
    networks:
      - app-network

  vllm:
    image: vllm/vllm-openai:cu130-nightly-5bff999d12dd061c102381b0c9c5d364c5953ea2
    container_name: vllm

    ipc: host
    gpus: all

    ports:
      - "8001:8000"

    volumes:
      - /home/cap/data/huggingface:/root/.cache/huggingface

    environment:
      - NVIDIA_VISIBLE_DEVICES=all
      - NVIDIA_DRIVER_CAPABILITIES=compute,utility
      - CUDA_VISIBLE_DEVICES=${CUDA_VISIBLE_DEVICES:-0}
      # Mint/Ubuntu typical driver lib locations (inside container when mounted by nvidia runtime)
      - LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu/nvidia/current

    entrypoint:
      - /bin/bash
      - -lc
      - >
          set -euxo pipefail;
          echo "=== nvidia-smi (inside container) ===";
          nvidia-smi || true;
          echo "=== libcuda candidates ===";
          ldconfig -p | grep -E 'libcuda\.so|libnvidia-ml\.so' || true;
          echo "=== show compat config ===";
          ls -la /etc/ld.so.conf.d/ || true;
          echo "=== remove cuda compat conf if present ===";
          rm -f /etc/ld.so.conf.d/00-cuda-compat.conf || true;
          ldconfig;
          echo "=== libcuda after ldconfig ===";
          ldconfig -p | grep -E 'libcuda\.so|libnvidia-ml\.so' || true;
          echo "=== launching vllm serve ===";
          exec vllm serve
          --host 0.0.0.0
          --port 8000
          --model Qwen/Qwen3-8B-AWQ
          --dtype auto

    healthcheck:
      test: ["CMD-SHELL", "curl -fsS http://localhost:8000/health >/dev/null || exit 1"]
      interval: 10s
      timeout: 5s
      retries: 60
      start_period: 120s

    restart: unless-stopped
    networks:
      - app-network

networks:
  app-network:
    driver: bridge