Docker behind VPN

Question

When I start the company VPN om my Ubuntu 16.04 laptop, I get a bunch of static routes. Among these are 172.26.0.0/16 (I checked this in the kern.log, and it also shows up when i run route). This worked just fine for a week or so, but suddenly it stopped working.

For some reason I can nslookup the IP address, but ping (or traceroute or dig) returns a "host not found" error.

If I enter the route as a static route in VPN IPv4 settings via the nm-applet, and restart the VPN, then I can ping the server indeed AND ping it successfully.

However this seems to kill all networking powers of the docker images I need to run to do my job, which worked fine before.

I have tried to reinstall the openconnect packages I use to connect to the Cicso AnyConnect server, but still no play.

I've been digging around the net for a while and the closest match I found was this posting Suddenly not able to access resources on VPN network in Ubuntu 18.04? which, needless to say, offer next to no comfort. I tried the solution offered by @crip659, but it didn't work for me.

I really hope someone can shine some light on this.

EDIT

Could this be caused the fact that my docker0 bridge has an IP-adress that coincides with the IP-range of one of the static routes?

Answer 1

So I fixed this.

Apparently the problem was, indeed, an IP-conflict between the docker0 bridge and an IP-range in one of the static routes.

I had already tried to assign a new IP-address to the docker0 bridge, by inserting a "bip"-entry by entering the command sudo nano /etc/docker/daemon.json and editing the file to make it look somewhat similar to this.

{
    "bip":"171.2.2.0/8"
}

but all the other addresses I tried resulted in sudo systemctl restart docker failing spectacularly.

I had (still dont have) no idea what IP-address to choose, so this 171.2.2.0/8 was a lucky guess I suppose. Maybe it will cause problems later, but that's another story.