Ceph object storage?

I'm interested in how this might work, but since SuSE was divested the project has lost priority so we have no published experience with setup.
There are two options I think would work: either export the Ceph object store as a S3 bucket, or export as an NFS mount shared by all nodes. I am not sure which way will perform better but there are different setup and security tradeoffs that may be more important anyways. My guess is that NFS will be available on most if not all Ceph stores and would be the go-to choice. You might consider the mount options recommended by other object store vendors, e.g. mount on all nodes and set the mountpoint as a USER storage location: mount -t nfs -o intr,hard,tcp your.ceph.ip:/share /path/to/mountpoint