I’ve got a three node cluster with one node just running etcd and control plane (basically just to be the quorum, i don’t want workloads on there). However, when I deploy longhorn it seems like it wants to use that host, but that host isn’t going to have the storage available to participate. Is there a way to get longhorn to just ignore a host?
It’s not supported currently. But we’re working on multiple disk scheduling (https://github.com/rancher/longhorn/issues/47) which will include this feature in the next release.
1 Like