I ended up solving this by manually updating the database and and then removing the offline node as the cluster came out of the locked state:
lxd sql global "UPDATE nodes SET schema=73, api_extensions=382 WHERE id=8"
lxc cluster remove --force glf-science-5