PSA: Docker nukes your firewall rules and replaces them with its own.
PSA: Docker nukes your firewall rules and replaces them with its own.
I use nftables to set my firewall rules. I typically manually configure the rules myself. Recently, I just happened to dump the ruleset, and, much to my surprise, my config was gone, and it was replaced with an enourmous amount of extremely cryptic firewall rules. After a quick examination of the rules, I found that it was Docker that had modified them. And after some brief research, I found a number of open issues, just like this one, of people complaining about this behaviour. I think it's an enourmous security risk to have Docker silently do this by default.
I have heard that Podman doesn't suffer from this issue, as it is daemonless. If that is true, I will certainly be switching from Docker to Podman.
You're viewing a single thread.
If you use firewalld, both docker and podman apply rules in a special zone separate from your main one.
That being said, podman is great. Podman in rootful mode, along with podman-docker and docker-compose, is basically a drop-in replacement for Docker.
Is it? Last time I tried none of my docker compose files would start correctly in podman compose.
I just set it up last week, it works exceptionally well.
Did you also install podman-docker, make sure that the podman socket was running, and verify that the socket directory referenced in the config files was correct?
Those are the 3 things that I got a bit stuck on. In the end, I RTFM and all was well.
podman-compose is different from docker-compose. It runs your containers in rootless mode. This may break certain containers if configured incorrectly. This is why I suggested podman-docker, which allows podman to emulate docker, and the native docker-compose tool. Then you use
sudo docker-composeto run your compose files in rootful mode.How is Podman rootful better than Docker? I was mostly attracted by the rootless path, but the breakage deterred me. Would you be so kind to tell me ?
It isn't that much better. I use it as drop-in docker replacement. It's better integrated with things like cockpit though and the idea is that it's easier to eventually migrate to rootless if you're already in the podman ecosystem.
Ok that sounds intetesting, I've found Cockpit easier to use than Proxmox, I'm new to virtualization and I don't want do nesting... I fear it will complicate things when I'll need to do GPU passthrough.
How is Podman integrated into Cockpit?
Also, I had so much trouble trying to bridge my Home Assistant VM to my LAN. Are there any tutorials on how to do this from Cockpit?
Your containers show up in Cockpit under the "Podman containers" section and you can view logs, type commands into their consoles, etc. You can even start up containers, manage images, etc.
Are there any tutorials on how to do this from Cockpit?
I have not done this personally, but I would assume you need to create a bridge device in Network Manager or via Cockpit and then tell your VM to use that. Keep in mind, bridge devices only work over Ethernet.
bridge devices only work over Ethernet
Yes, I want to reach my HA VM from my LAN connected devices.
Cockpit definitely has the ability to create bridge devices. I haven't found a tutorial specifically for cockpit, but you can follow something like this and apply the same principles to the "Add Bridge" dialog in Cockpit's network settings.
I'm a podman user, but what's the point of using podman if you are going to use a daemon and run it as root? I like podman so I can specifically avoid those things.
I am using it as a migration tool tbh. I am trying to get to rootless, but some of the stuff I host just don't work well in rootless yet, so I use rootful for those containers. Meanwhile, I am using rootless for dev purposes or when testing out new services that I am unsure about.
Podman also has good integration into Cockpit, which is nice for monitoring purposes.