Docker中PostgreSQL未监听指定端口5433的连接问题排查

我无法连接Docker中运行的PostgreSQL，容器并未监听指定的5433端口，而是使用了默认的5432端口（本地已有非Docker的PostgreSQL占用该端口）。通过Docker Python模块部署测试镜像/容器时表面无异常，但install.py连接数据库时出现以下错误：

psycopg2.OperationalError: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.

Docker日志显示容器启动后停止再重启，但始终监听5432端口。

日志内容：

waiting for server to shut down...2023-09-03 16:59:24.181 UTC [47] LOG:  received 
fast shutdown request
.2023-09-03 16:59:24.181 UTC [47] LOG:  aborting any active transactions
2023-09-03 16:59:24.182 UTC [47] LOG:  background worker "logical replication launcher" (PID 53) exited with exit code 1
2023-09-03 16:59:24.183 UTC [48] LOG:  shutting down
2023-09-03 16:59:24.184 UTC [48] LOG:  checkpoint starting: shutdown immediate
2023-09-03 16:59:24.269 UTC [48] LOG:  checkpoint complete: wrote 4473 buffers (27.3%); 0 WAL file(s) added, 0 removed, 2 recycled; write=0.023 s, sync=0.060 s, total=0.086 s; sync files=963, longest=0.001 s, average=0.001 s; distance=34921 kB, estimate=34921 kB
2023-09-03 16:59:24.285 UTC [47] LOG:  database system is shut down
 done
server stopped

PostgreSQL init process complete; ready for start up.

2023-09-03 16:59:24.415 UTC [1] LOG:  starting PostgreSQL 15.4 (Debian 15.4-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
2023-09-03 16:59:24.415 UTC [1] LOG:  listening on IPv4 address "0.0.0.0", port 5432
2023-09-03 16:59:24.415 UTC [1] LOG:  listening on IPv6 address "::", port 5432
2023-09-03 16:59:24.416 UTC [1] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
2023-09-03 16:59:24.419 UTC [71] LOG:  database system was shut down at 2023-09-03 16:59:24 UTC
2023-09-03 16:59:24.423 UTC [1] LOG:  database system is ready to accept connections

Dockerfile内容：

FROM postgres:12.15
FROM postgis/postgis
ENV POSTGRES_USER awdn
ENV POSTGRES_PASSWORD ****
ENV POSTGRES_DB awdn_data
ENV POSTGRES_PORT 5433
EXPOSE 5433

install.py内容：

import docker,psycopg2

client = docker.from_env()
client.containers.prune()

try:
    container = client.containers.get('dbase')
    container.stop()
    container.remove()
except docker.errors.NotFound as e:
    print("The container is not running.\n")

try:    
    awdnunittest = client.images.get('awdn-unit-test')
    client.images.remove(awdnunittest.short_id)
except docker.errors.NotFound as e:
    print("No image exists for the unittest.\nThe psql image is being installed...\n")

client.images.build(path='./',tag='awdn-unit-test')

client.containers.run('awdn-unit-test',name='dbase',ports={'5433':5433},detach=True,publish_all_ports=True)

conn = psycopg2.connect("dbname=awdn_data user=awdn port=5433 password=**** host=localhost")

1. 修复Dockerfile基础镜像与端口配置问题 #

• 你的Dockerfile使用了两次FROM指令，这会导致仅最后一个FROM postgis/postgis生效，之前的postgres:12.15被完全忽略，导致实际运行的PostgreSQL版本与预期不符。
• PostgreSQL官方镜像不识别POSTGRES_PORT环境变量，需使用PGPORT指定监听端口。

修改后的Dockerfile：

FROM postgis/postgis:12-3.3  # 选用对应PostgreSQL 12版本的PostGIS镜像，保持版本一致性
ENV POSTGRES_USER awdn
ENV POSTGRES_PASSWORD ****
ENV POSTGRES_DB awdn_data
ENV PGPORT 5433  # 替换POSTGRES_PORT为PGPORT
EXPOSE 5433

2. 修正容器端口映射逻辑 #

若容器内PostgreSQL已通过PGPORT改为监听5433，端口映射保持{'5433':5433}即可；若未修改容器内端口，需将容器内默认的5432端口映射到宿主机5433，即{'5432':5433}。

3. 添加容器启动等待逻辑 #

install.py中启动容器后立刻执行数据库连接，此时容器内PostgreSQL可能尚未完全初始化完成，需添加等待逻辑：

import time
# ... 启动容器的代码 ...
container = client.containers.run('awdn-unit-test',name='dbase',ports={'5433':5433},detach=True,publish_all_ports=True)

# 等待PostgreSQL服务就绪
while True:
    logs = container.logs().decode('utf-8')
    if "database system is ready to accept connections" in logs:
        break
    time.sleep(1)

# 执行连接操作
conn = psycopg2.connect("dbname=awdn_data user=awdn port=5433 password=**** host=localhost")

4. 验证容器端口监听状态 #

启动容器后，执行以下命令检查容器内端口监听情况，确认PostgreSQL是否监听在指定端口：

docker exec -it dbase netstat -tulpn

内容的提问来源于stack exchange，提问作者deadpickle