Im unable to view the original thread, so apologies if the content of my message is already repeated,
judging by the title of the email, i would say sandro is right, if one of the node in your gluster enabled cluster has a failed disk, you could replace the disk with a new one(thus losing all your old data thereby making your host new) and run "same node fqdn replace host procedure" . follow this instruction to setup your replace host  playbook instructions.

Kind Regards,
Prajith Kesava Prasad.