Ceph Datastore for several LXD nodes

Hello, I’ve been using LXD 2.0 so far with and I need to use ceph via latest lxd. I’ve read some doc about it and there are things I’m not quite certain:

1- I have already set up a production Ceph Cluster used by other hypervisors (LXD 2, with some workarounds, and KVM). Can I reutilize that existing cluster with a new LXD host with ceph support?
2- I’ve read https://github.com/lxc/lxd/issues/3716#issuecomment-324956166 https://github.com/lxc/lxd/issues/3747 https://github.com/lxc/lxd/issues/3958 and seems there are some issues related to use a common ceph storage pool with several LXD nodes. Can I use a ceph cluster across several LXD nodes? Do they need to be clustered over the network (lxc remote add lxd-node fashion way) or neither way will this work?
3- May I point the root disk inside the container configuration to a specific block ? ie.

    type: disk
    source: /dev/rbd1
    path: /
    pool: default

Another thing, when will be released another lts version for LXD? By the way great work with the development.

Also, will I need a specific Ceph version and/or kernel version?

I think i figured out the process you can find the procedures and some troubleshooting here https://github.com/lxc/lxd/issues/3993, the only thing left out was the ceph client configuration, you need to configure the ceph client on the LXD node. I tried it with lxd 2.18 and ceph jewel.