Docker环境下指定正确主机名后Celery仍无法连接Redis

问题概述 #

在Docker环境中，已指定Redis主机名为redis，Redis服务已正常启动，但Celery容器始终无法连接Redis，报错连接超时。

相关配置文件 #

Dockerfile #

FROM python:3.8.16-bullseye
RUN apt-get install ca-certificates
RUN update-ca-certificates
COPY requirements.txt requirements.txt
RUN pip install --upgrade pip && pip install -r requirements.txt
COPY . .
WORKDIR /task_manager
CMD celery -A worker.celery worker --loglevel info --uid nobody

worker.py代码 #

from celery import Celery
import os

BROKER_URI = os.environ["BROKER_URI"]
BACKEND_URI = os.environ["BACKEND_URI"]

celery = Celery('worker', backend=BACKEND_URI)

celery.conf.broker_url = BROKER_URI

docker-compose配置 #

version: "3.8"

services:
    worker:
        container_name: celery
        environment:
            - BROKER_URI=redis://redis
            - BACKEND_URI=redis://redis
        build:
            context: .
            dockerfile: Dockerfile
        depends_on:
            -   redis
    redis:
        image: "redis:alpine"
        ports:
            - 6379:6379

错误日志 #

celery    |  
celery    |  -------------- celery@f760551b7908 v5.2.7 (dawn-chorus)
celery    | --- ***** ----- 
celery    | -- ******* ---- Linux-5.19.0-41-generic-x86_64-with-glibc2.2.5 2023-05-05 15:05:25
celery    | - *** --- * --- 
celery    | - ** ---------- [config]
celery    | - ** ---------- .> app:         worker:0x7f223ace3a00
celery    | - ** ---------- .> transport:   redis://redis:6379/0
celery    | - ** ---------- .> results:     redis://redis:6379/0
celery    | - *** --- * --- .> concurrency: 8 (prefork)
celery    | -- ******* ---- .> task events: OFF (enable -E to monitor tasks in this worker)
celery    | --- ***** ----- 
celery    |  -------------- [queues]
celery    |                 .> celery           exchange=celery(direct) key=celery
celery    |                 
celery    | 
celery    | [tasks]
celery    | 
celery    | 
celery    | [2023-05-05 15:07:35,918: ERROR/MainProcess] consumer: Cannot connect to redis://redis:6379/0: Error 110 connecting to redis:6379. Connection timed out..
celery    | Trying again in 2.00 seconds... (1/100)

排查与解决方案 #

1. 解决depends_on的启动顺序局限性 #

depends_on仅保证容器启动顺序，不确保Redis服务完全初始化完成。Celery启动时Redis可能还未就绪，导致连接超时。

方案1：使用wait-for-it工具等待Redis端口
修改Dockerfile，添加工具安装和启动命令：

FROM python:3.8.16-bullseye
RUN apt-get update && apt-get install -y ca-certificates wait-for-it
RUN update-ca-certificates
COPY requirements.txt requirements.txt
RUN pip install --upgrade pip && pip install -r requirements.txt
COPY . .
WORKDIR /task_manager
CMD wait-for-it redis:6379 -- celery -A worker.celery worker --loglevel info --uid nobody

方案2：用Shell循环检测Redis端口（无需额外工具）
修改Dockerfile，安装netcat并添加等待逻辑：

FROM python:3.8.16-bullseye
RUN apt-get update && apt-get install -y ca-certificates netcat
RUN update-ca-certificates
COPY requirements.txt requirements.txt
RUN pip install --upgrade pip && pip install -r requirements.txt
COPY . .
WORKDIR /task_manager
CMD until nc -z redis 6379; do sleep 1; done && celery -A worker.celery worker --loglevel info --uid nobody

2. 检查Redis网络绑定配置 #

若Redis容器绑定了127.0.0.1，则仅允许本地访问，其他容器无法连接。修改docker-compose的Redis服务，显式指定绑定所有IP：

redis:
    image: "redis:alpine"
    ports:
        - 6379:6379
    command: redis-server --bind 0.0.0.0

3. 验证容器网络连通性 #

进入Celery容器，测试与Redis的连通性：

docker exec celery ping redis
# 或测试端口
docker exec celery nc -zv redis 6379

如果ping不通，说明网络配置异常，检查docker-compose是否使用默认网络，或是否自定义了网络但未将两个服务加入同一网络。

4. 显式指定Redis连接参数 #

将环境变量中的URI改为完整格式，避免默认参数歧义：

worker:
    environment:
        - BROKER_URI=redis://redis:6379/0
        - BACKEND_URI=redis://redis:6379/0

内容的提问来源于stack exchange，提问作者Bob