I spent a lot of time at a previous job supporting an NDB cluster and I can attest first hand that it is awful. Single node failures cause entire cluster shutd

I spent a lot of time at a previous job supporting an NDB cluster and I can attest first hand that it is awful. Single node failures cause entire cluster shutdowns, the cluster then takes forever to recover and must be done in a specific order. In fact just thinking about it makes me anxious.