Is it possible to use a already installed bare-metal machine in a cluster?

[silenium-dev]

I'd like to use CoreOS for my nodes, which isn't supported to be installed with installimage.
So I went ahead and wrote an ansible playbook to deploy and set up my bare-metal machine with CoreOS and all required settings/tools.
I didn't find any documentation on using these set up machines in a cluster whatsoever, and the HetznerBareMetalMachine requires a config for installimage.

Is there any way to use CoreOS (either already set up or via installimage) with the CAPH?

[batistein]

You can specify any custom image in installimage, you just need to create a tarball of it and save it to an FTP server, for example. See more about how to create an image and what types of storage are supported: https://docs.hetzner.com/robot/dedicated-server/operating-systems/installing-custom-images/

[silenium-dev]

In the documentation it says, the custom image "must be one of Debian, Ubuntu, OpenSUSE or CentOS to get the network settings properly configured". CoreOS a special variant of Fedora with an ostree based root filesystem, and uses systemd-networkd instead of NetworkManager. I will try it though.

[batistein]

Sorry, as i have never used coreos with installimage i don't know. But I would ask here: https://github.com/hetzneronline/installimage or directly the Hetzner support

[silenium-dev]

I've opened an issue here: hetzneronline/installimage#72