Closed fabjan closed 3 years ago
Side notes:
I had to redo the apply a few times because DNS did not get setup properly on one node so it could not install anything with apt. I don't know if this was a knock-on-effect but the fireguard setup failed on one node as well so the etcd module never finished getting setup.
I fixed this by doing:
$ terraform destroy $(terraform state list | grep -e wireguard -e etcd | xargs printf -- "-target %s ")
and then re-applying the terraform (if the wireguard setup fails on one node the terraform setup can't recover them, so they all have to be torn down and redone).
(I also started out with 1G memory nodes at first but the preflight of kubeadm init did not like that and it was not obvious when debugging that that was the cause.)
Even though I didn't test this, code certainly seems to be clean. Thank you!
Thanks for making this guide! I'm setting up a hobby cluster for learning but I use UpCloud so I added a server setup using their provider and thought I'd share my changes.