Q: Docker部署后ipv4不通的现象 || Q: IPv4 is blocked after Docker deployment

[MWTJC]

Describe the bug | 问题描述

我偶然在一个仅有ipv4的环境下，想访问具有公网ip（v4与v6均可达）的经Caddy反向代理（代理配置见文末）的Waline服务端后台，发现不可达，但相同配置（均通过docker-compose部署）下的lanraragi可达，artalk不可达，遂迷惑具体原因究竟是Docker层面还是哪个层面，故来此求助，
防火墙是设定完全放行caddy.exe，
以下是摸索出的可能的局域网内可复现步骤：（应该与直接在关闭ipv6后，访问下面的问题网站是相同的现象）

• windows11
• docker desktop （我是微软商店版）
• docker-compose.yaml

services:
  waline:
    container_name: waline
    image: lizheming/waline:latest
    restart: always
    ports:
      - 8360:8360
    volumes:
      - ${PWD}:/app/data
    environment:
      TZ: 'Asia/Shanghai' 
      SQLITE_PATH: '/app/data'  
      JWT_TOKEN: '********'  
      IPQPS: 10        

• 放置waline.sqlite
• pwsl

docker compose up -d --force-recreate

• 本机网页浏览器访问 localhost:8360/ui -> 正常
• 本机网页浏览器访问 192.168.5.2:8360/ui -> ERR_CONNECTION_TIMED_OUT
  （192.168.5.2为部署机器的局域网ip，此时dockerdesktop -> Resources -> network相关配置如下）
  • Default networking mode: Dual IPv4/IPv6
  • Enable host networking: True

CaddyFile(caddy.exe)

mwtjc.online:18360 {
	tls {
		dns dnspod "*****************"
	}
	reverse_proxy :8360 {
	    header_up X-Forwarded-For {header.X-Forwarded-For}
	}
}

Website URL | 问题网站

https://mwtjc.online:14001/blog-demo/

Where your waline deploy? | 服务部署在哪里？

Docker

Where your comment data store? | 数据存储在哪里？

SQLite

---

Describe the bug | Problem description

I accidentally wanted to access the Waline server backend via Caddy reverse proxy (see the end of the article for proxy configuration) with a public IP address (both v4 and v6 are reachable) in an environment with only IPv4. I found that it was unreachable, but the lanraragi under the same configuration (both deployed through docker-compose) was reachable, and artark was unreachable. I was confused about whether the specific cause was the Docker layer or which layer, so I came here for help.
The firewall is set to fully allow caddy.exe.
The following are possible steps that can be reproduced within the local area network: (It should be the same phenomenon as directly accessing the following problem website after turning off ipv6)

• windows11
• docker desktop (I am the Microsoft store version)
• docker-compose.yaml

services:
  waline:
    container_name:waline
    image: lizheming/waline:latest
    restart: always
    ports:
      - 8360:8360
    volumes:
      - ${PWD}:/app/data
    environment:
      TZ: 'Asia/Shanghai'
      SQLITE_PATH: '/app/data'
      JWT_TOKEN: '********'
      IPQPS: 10

• place waline.sqlite
  -pwsl

docker compose up -d --force-recreate

• Native web browser access localhost:8360/ui -> normal
• Native web browser access 192.168.5.2:8360/ui -> ERR_CONNECTION_TIMED_OUT
  (192.168.5.2 is the LAN IP of the deployment machine. At this time, the relevant configuration of dockerdesktop -> Resources -> network is as follows)
  • Default networking mode: Dual IPv4/IPv6
  • Enable host networking: True

CaddyFile(caddy.exe)

mwtjc.online:18360 {
	tls {
		dns dnspod "*******************"
	}
	reverse_proxy :8360 {
	    header_up X-Forwarded-For {header.X-Forwarded-For}
	}
}

Website URL | Problem website

https://mwtjc.online:14001/blog-demo/

Where your waline deploy? | Where is the service deployed?

Docker

Where your comment data store? | Where is the data stored?

SQLite