Joining a cluster with a deviating zpool name

I have an existing cluster where all nodes have a separate partition backing a zfs pool, all named “local”. I have now installed a new debian machine with zfs on root where only one pool exists named “zroot”.

Have I understood it correctly that it’s now impossible to join this machine into the cluster simply because of the zpool name? It seems strange to me that it’s not enough to call the incus storage pool “local” like in the rest of the cluster, but that the backing zfs pool also needs this name. If I had entered a device or partition here instead, it doesn’t need to be the same (presumably because incus then creates the pool itself and then “owns” the name)

It seems to me like the incus pool name is already an abstraction and I cannot understand why the pool name itself also has to be identical instead of something that can be fetched from the attributes of the incus pool.

Is there something I’m missing or do I really have to ensure that my literal zpool name is correct already when installing Debian to be able to join an existing cluster, when it’s the only disk I have?

Well, it seems I was just confused. The right way seems to be to specify this when joining the cluster:

Choose “source” property for storage pool “local”: zroot/incus
Choose “zfs.pool_name” property for storage pool “local”: zroot/incus

While previously what I’ve done when it’s a separate partition has been something like:

Choose "source" property for storage pool "local": /dev/nvme0n1p3
Choose "zfs.pool_name" property for storage pool "local": local