用于 windows 的 WSL2 和 Docker 桌面的网络问题

Question

TL;DR

It seems I'm constantly running into networking problems when using WSL2 and Docker Desktop for Windows.

Normally, I can access a port started from a WSL2 instance ("distribution" in WSL2-speak) from inside Docker containers, by finding the IP address of the WSL2 instance, and referencing to that IP address from inside the containers. But today that doesn't work for some reason. I'm getting used to rebooting when I run into WSL2+DockerDesktop problems, but now I'd like to finally understand what is actually wrong.

I'm pretty sure that when I reboot in a moment, all will be fine and dandy.

Do you have an idea on how to debug something this?

I also have a corporate Forticlient VPN that could be messing this up. But I have no choice in that matter... ;-(

Details

I find the IP with:

WSL2:» ip -br a
lo               UNKNOWN        127.0.0.1/8 ::1/128 
bond0            DOWN           
dummy0           DOWN           
sit0@NONE        DOWN           
eth0             UP             172.26.183.209/20 fe80::215:5dff:fef9:cc5c/64 

So the WSL2 machine's IP address is 172.26.183.209 .

So now that I have a web server listening on port 8080 in a process my WSL2, it can be accessed with http://172.26.183.209:8080 , as confirmed by nmap and curl from WSL2:

WSL2@~» nmap -Pn -p 8080 172.26.183.209            
Starting Nmap 7.80 ( https://nmap.org ) at 2020-12-23 16:15 CET
Nmap scan report for ubuntu-wsl2 (172.26.183.209)
Host is up (0.000045s latency).

PORT     STATE SERVICE
8080/tcp open  http-proxy

Nmap done: 1 IP address (1 host up) scanned in 0.02 seconds

WSL2@~» curl http://172.26.183.209:8080 > /dev/null
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0

From inside a container, however:

[root@0467b7ef17e7 /]# nmap -Pn -p 8080 172.26.183.209
Starting Nmap 7.70 ( https://nmap.org ) at 2020-12-23 15:12 UTC
Nmap done: 1 IP address (0 hosts up) scanned in 0.43 seconds
[root@0467b7ef17e7 /]# curl http://172.26.183.209:8080
curl: (7) Failed to connect to 172.26.183.209 port 8080: No route to host

It also works from inside the docker-desktop WSL2 container:

WSL2@~» wsl.exe -d docker-desktop wget http://172.26.183.209:8080
Connecting to 172.26.183.209:8080 (172.26.183.209:8080)     
Connecting to host.docker.internal:8180 (192.168.0.202:8180)
index.html           100% |********************************|  7308  0:00:00 ETA

Just not from inside the containers. Until I reboot. AAAARRRRRHHHHHHHHHHHHHHHH:-)

One hint - something that doesn't make sense to me - is this from inside a container:

[root@0467b7ef17e7 /]# ping 172.26.183.209
PING 172.26.183.209 (172.26.183.209) 56(84) bytes of data.
From 172.26.0.2 icmp_seq=1 Destination Host Unreachable
From 172.26.0.2 icmp_seq=2 Destination Host Unreachable
From 172.26.0.2 icmp_seq=3 Destination Host Unreachable
^C
--- 172.26.183.209 ping statistics ---
6 packets transmitted, 0 received, +3 errors, 100% packet loss, time 202ms
pipe 4

What is 172.26.0.2 ? Hmm...

Networking setup:

(I've modified Ethernet 3's IP address since it's in the corporate network)

WSL2@~» ip -br a
lo               UNKNOWN        127.0.0.1/8 ::1/128 
bond0            DOWN           
dummy0           DOWN           
sit0@NONE        DOWN           
eth0             UP             172.26.183.209/20 fe80::215:5dff:fef9:cc5c/64 
WSL2@~» ipconfig.exe

Windows IP Configuration


Ethernet adapter Ethernet 3:

   Connection-specific DNS Suffix  . :
   Link-local IPv6 Address . . . . . : fe80::405f:b7d2:70b4:b405%19
   IPv4 Address. . . . . . . . . . . : 10.14.11.17
   Subnet Mask . . . . . . . . . . . : 255.255.255.255
   Default Gateway . . . . . . . . . : 10.14.11.18

Ethernet adapter Ethernet 4:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Ethernet adapter Ethernet:

   Connection-specific DNS Suffix  . :
   Link-local IPv6 Address . . . . . : fe80::7cd1:5641:ac65:7004%15
   IPv4 Address. . . . . . . . . . . : 192.168.0.202
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . : 192.168.0.1

Ethernet adapter VirtualBox Host-Only Network:

   Connection-specific DNS Suffix  . :
   Link-local IPv6 Address . . . . . : fe80::f06d:9785:cff5:2ad0%6
   IPv4 Address. . . . . . . . . . . : 192.168.225.1
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . :

Wireless LAN adapter Wi-Fi:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Wireless LAN adapter Local Area Connection* 7:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Wireless LAN adapter Local Area Connection* 10:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Ethernet adapter Ethernet 2:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Ethernet adapter Bluetooth Network Connection:

   Media State . . . . . . . . . . . : Media disconnected
   Connection-specific DNS Suffix  . :

Ethernet adapter vEthernet (WSL):

   Connection-specific DNS Suffix  . :
   Link-local IPv6 Address . . . . . : fe80::c507:3c2b:62d:7270%39
   IPv4 Address. . . . . . . . . . . : 172.26.176.1
   Subnet Mask . . . . . . . . . . . : 255.255.240.0
   Default Gateway . . . . . . . . . :

WSL2@~» route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         172.26.176.1    0.0.0.0         UG    0      0        0 eth0
172.26.176.0    0.0.0.0         255.255.240.0   U     0      0        0 eth0

Answer 1

Have you tried just restarting the container and not the whole machine? It sounds like it's the container losing track of its network configuration. A lot of things with containers are written at startup and never updated or not consistently updated.

Answer 2

I'm still looking for a solution, but I found a quicker workaround: just restart WSL2. From a cmd window, enter:

wsl --shutdown

Docker will complain about having stopped unexpectedly, and then will give you another message when you restart it, but that's not harmful.

Answer 3

I had a similar issue where the IP address for WSL2 conflicted with the 172.17.0.0/16 address space which Docker uses internally. In my case, in WSL2:

$ hostname -I
172.17.112.35

The fix was to change the address space Docker uses to a different one which also doesn't conflict with my local network. In the settings in Docker Desktop, set "bip": "192.168.200.1/24" on the Docker Engine configuration page.

Credit: https://www.beyondjava.net/docker-wsl-network

Restarting your PC or WSL "fixes" this issue temporarily because the IP assigned to WSL happens to not conflict with addresses used by Docker.