Kasmweb problem with docker

Software & Web Applications
1

Anyone running Kasmweb 1.16 and good with docker willing to give me a hand?

I upgraded from a different type of install with 1.15 or maybe 1.14, and everything broke. Tried a fresh install on a fresh Debian 12 VM, and still broke. Deleted the VM and built a fresh VM on my Lab, still broke.

I finally dug down into the problems yesterday and figured out the cause, but not the fix.

Cause is in the docker bridge docker0, it is assigning this bridge to 172.17.0.0/16… My production network is on 172.17.0.0/16 so all traffic sent to that server, or coming back from that server to my production network gets caught in that bridge. If I change the IP range in Advanced Networking GUI, everything is good. But if I reboot then a new docker0 is created and back to the 172.17.0.0 network.

I looked through the yaml files to see if I could find where these networks are built, but I’m not finding that info. It doesn’t help that I’m not up to speed on Docker, not even close. About the only thing I know is that Docker Compose builds containers based on a yaml file full of instructions.

I’m probably going to go through my notes for the old install and see if I can still replicate it, that would at least give me web browser and reminna so I can get into my lab from home. Trying to learn VSphere with the $200 personal license (1 year term). There are other things I want to do for my coworkers, but right now my focus is on my lab. Maybe I can just “disconnect” my lab from production and forward the port from the firewall to the Kasm server on the Lab network. The latter might be the plan for the weekend.

2

I fixed part of the issues, still have two more bridges automatically assigned where I don’t want them, but I’ll get that done later.

Needed to create /etc/docker/daemon.json file with the following in the file:

{
    "bip": "192.168.50.1/24"
}

Did a reboot to make sure it held, and so far so good. You could also stop docker and start it after making the file, but reboot is the acid test.

Need to figure out why the Guac containers can’t find the server with DNS, the host can find itself after an entry into DNS, and I made entries in the local hosts file too. Another thing to run down to make it all work as desired.

But at least things can talk to each other now.

3

to avoid collisions of bridge addresses with your production network addresses you can put something along these lines in /etc/docker/daemon.json:

{
“default-address-pools”: [
{“base”: “172.30.0.0/16”, “size”: 24},
{“base”: “10.30.0.0/16”, “size”: 24}
]
}

4

Thanks, getting a better idea of what I want to change. Your info lead me to this: