I'm more familiar with Docker on Linux, where creating a bridge network makes it straightforward to give containers their own IP addresses and let them communicate with the host and with one another. I'm now working on Windows with Docker Desktop configured to use Hyper-V, and I need two containers to have separate addresses while maintaining connectivity to both the Windows host and each other. I found references to creating additional loopback adapters, but I'm not sure whether that's necessary or what the correct setup should be. What is the recommended way to configure this?
3 Answers
Try creating a user-defined bridge network instead of relying on Docker's default bridge. Containers on that network can communicate with each other through the Docker gateway, and traffic between containers generally stays within the network's subnet. This should provide the separate container addresses and host connectivity you need without manually creating loopback adapters.
If your goal is to learn and use normal Linux Docker networking, a small Debian virtual machine may be less frustrating than Docker Desktop. Install Docker inside that VM and create the bridge there; it behaves much more like the Linux setup you're used to. This is especially useful since using WSL is not an option in your environment.
Docker Desktop networking on Windows is more complicated than native Linux Docker because the containers run behind an additional virtualization layer. Hyper-V and Docker Desktop also have to account for VPNs, laptop connectivity, and preventing containers from being exposed directly to the local network. As a result, Linux bridge-network behavior does not map perfectly to Windows, and some host-networking scenarios may not be possible in the same way.
I can't use WSL in this setup, so I'll need either a Hyper-V-compatible Docker Desktop configuration or a separate Linux VM.