This message was deleted.

Give it a static IP

Yeah, that is what I was afraid of (vagrant sadly can't do it when provisioner is Hyper-V). Is there a way I could configure the cluster to use domain instead? (We have a logic with vagrant-hostmanager that automatically assigns a local domain for the VM)

no, etcd uses the node’s IP address as part of the member identifier. If the IP changes, then it doesn’t recognize that it’s part of the cluster. You can start with --cluster-reset to update the membership, but that’s not particularly easy to do. You could instead configure a dummy interface with a static IP and use that as the --node-ip ?

Apologies, not sure what you meant with dummy interface, not very savy with linux networks. However kinda gave up with Hyper-V, only one user in our company used it. A simpler question though However can I somehow configure which of the IP's is used as node IP, if it machine has multiple IPs? I noticed with Virtubalbox and VMWare providers, even when we have static IP configured, it's not the IP kubernets chooses as node-ip as it's not the first interface for the VM. I imagine I have to use the

node-ip

config value?

it uses the primary IP of the interface with the default route. If you want to override it, use

node-ip

Thanks, will try to play around with that.