So I'm pretty sure I know the answer. I'm building a very small cluster for now. Two nodes. One hosts the NAS and SAN and a few other 'infrastrcture-y' things. The other hosts most of the VMs that I want to boot *from* the SAN. I can manually migrate them from the service node to the infra node if I need to, say, to avoid interrupting a project during an update. I'm generally not worried about anything on the service node failing spontaneously if there's a system restart, and since it's clustered with network storage, I can bring it back up on the infra node if I need it. All that being the case, if I'm not doing anything overtly dumb like trying to join a third new machine to the cluster with a split brain, or running ceph, or the like, do I really care about quorum for the day-to-day operation? If one goes down, *can* I actually manually bring up a Non-HA VM on the other node without quorum? Reliability isn't a concern. I have a cold spare server, and could have a *dead* server replaced in a few hours in most cases, and a few days if I need spare parts (I'm also building up the spare part stockpile where applicable, slowly).
TL;DR: If I'm not using HA in a two-node cluster, how important is Quorum, and can a single node in the cluster continue to function semi-sanely without Quorum?
TL;DR: If I'm not using HA in a two-node cluster, how important is Quorum, and can a single node in the cluster continue to function semi-sanely without Quorum?
