Persistent etcd

This release features a new mechanism for etcd in BTrDB. In prior releases we used etcd-operator but they don't really handle persistence well (servers restarting) and no longer offer a good backup solution. This release runs etcd as daemonsets on nodes that are configured in the siteconfig.yaml. This will make the cluster tolerant to restarts and power cycling.

If upgrading from a previous version of BTrDB, you will need to backup your etcd data and restore it. Please take a look at the backuptool for a guide on how this is done.