Missing `/etc/systemd/system/kubelet.service.d` directory on new installation

[aja114]

I am trying to run the role on a new virtual machine running ubuntu22.04. I am using most of the defaults from the role including the kubernetest_version: 1.25.0. When running the role I obtain the error from this deprecated task that setups the KUBELET_EXTRA_ARGS. I imagine it will be removed soon, but currently the role would not run.

The new installed kubeadm package seem to create the equivalent directory at /lib/systemd/system/kubelet.service.d, maybe that has changed in the recent releases. The solution is to create a symlink from /lib/systemd/system/kubelet.service.d to /etc/systemd/system/kubelet.service.d before running the role.

[geerlingguy]

Thanks! It'll be good to figure out when this change occurred, and if we can just drop support for the old path or have to have some sort of shim for newer versions :/

[seagercortez]

@aja114 Thank you for making this post about the symlink. I haven't used Kubernetes before but we're reconstructing our backend services for production and I was finally looking into Kubernetes, following some old guides for setting up a cluster of about 15 servers. This saved me a lot of time searching for a solution haha! 😂