Is this requirement no more needed in newer ceph ?
And I might add that the "requirement" that the UID/GID be the same across the cluster is not really a hard requirement. It is, shall we say, advantageous because it enables swapping of OSD disks between nodes.
If two nodes in the cluster use the same UID/GID for ceph, and you swap an OSD disk between them, the OSD will "just work" in the new node.
Other than this, there is no requirement that they be the same.