Docker Compose 서비스 의존성 문제 해결

Docker Compose에서 여러 서비스가 있을 때, 의존성이 있는 서비스가 아직 준비되지 않은 상태에서 다른 서비스가 시작되는 문제가 빈번합니다.

Environment

$ docker-compose --version
Docker Compose version v2.23.0

$ docker --version
Docker version 24.0.7, build afdd53b

Problem: 의존성이 있는 서비스가 준비되기 전에 시작

# docker-compose.yml
version: '3.8'

services:
  web:
    build: .
    ports:
      - "8000:8000"
    depends_on:
      - db
      - redis
  
  db:
    image: postgres:15
    environment:
      POSTGRES_DB: mydb
      POSTGRES_PASSWORD: secret
  
  redis:
    image: redis:7-alpine

$ docker-compose up -d
$ docker-compose logs web
# Error: Connection refused to database
# Error: Connection refused to redis

Analysis: 서비스 시작 순서 이해

# Check service status
$ docker-compose ps
NAME      IMAGE         COMMAND                  STATUS
db        postgres:15   "docker-entrypoint.s…"   Up 5 seconds
redis     redis:7       "docker-entrypoint.s…"   Up 5 seconds
web       myapp         "python app.py"          Restarting (1)

# Check logs
$ docker-compose logs --tail 20 web

Solution: 의존성 관리 방법

방법 1: depends_on + condition

# docker-compose.yml
version: '3.8'

services:
  web:
    build: .
    depends_on:
      db:
        condition: service_healthy
      redis:
        condition: service_healthy
  
  db:
    image: postgres:15
    environment:
      POSTGRES_DB: mydb
      POSTGRES_PASSWORD: secret
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U postgres"]
      interval: 5s
      timeout: 5s
      retries: 5
  
  redis:
    image: redis:7-alpine
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 5s
      retries: 5

방법 2: wait-for-it 스크립트

#!/bin/bash
# wait-for-it.sh
host="$1"
port="$2"
timeout="$3"

for i in $(seq 1 $timeout); do
    nc -z $host $port && echo "Service $host:$port is ready" && exit 0
    sleep 1
done

echo "Service $host:$port is not ready after $timeout seconds"
exit 1

version: '3.8'

services:
  web:
    build: .
    command: >
      sh -c "./wait-for-it.sh db 5432 30 &&
             ./wait-for-it.sh redis 6379 30 &&
             python app.py"
    depends_on:
      - db
      - redis

방법 3: Entrypoint 스크립트

# Dockerfile
FROM python:3.11-slim

WORKDIR /app

# Install wait-for-it
RUN apt-get update && apt-get install -y netcat && rm -rf /var/lib/apt/lists/*

COPY wait-for-it.sh /app/wait-for-it.sh
RUN chmod +x /app/wait-for-it.sh

COPY . .

CMD ["./wait-for-it.sh", "db", "5432", "--", "python", "app.py"]

방법 4: Dockerize 사용

version: '3.8'

services:
  web:
    build: .
    command: dockerize -wait tcp://db:5432 -wait tcp://redis:6379 -timeout 30s python app.py
    depends_on:
      - db
      - redis

방법 5: Python에서 재시도 로직

# app.py
import time
import sys
from sqlalchemy import create_engine
from sqlalchemy.exc import OperationalError

def wait_for_database(max_retries=30, delay=1):
    """Wait for database to be ready"""
    engine = create_engine("postgresql://postgres:secret@db:5432/mydb")
    
    for i in range(max_retries):
        try:
            engine.connect()
            print("Database is ready!")
            return True
        except OperationalError:
            print(f"Waiting for database... ({i+1}/{max_retries})")
            time.sleep(delay)
    
    print("Database connection failed after max retries")
    sys.exit(1)

if __name__ == "__main__":
    wait_for_database()
    # Start application
    start_app()

Lessons Learned

1. depends_on만으로는 부족: depends_on은 컨테이너 시작만 보장할 뿐, 서비스 준비 완료를 보장하지 않습니다.

2. Healthcheck 필수: condition: service_healthy를 사용하려면 서비스에 healthcheck를 정의해야 합니다.

3. Application-level 재시도: Infrastructure-level 대기보다 application-level 재시도가 더 견고합니다.

4. 적절한 타임아웃: 모든 대기 로직에는 합리적인 타임아웃을 설정하여 무한 대기를 방지하세요.

5. ** 로그 모니터링**: docker-compose logs -f로 서비스 시작 순서와 에러를 모니터링하세요.

This blog does not accept any external sponsorships, affiliate marketing, or ad revenue.