I have two NFS servers and two CIFS servers, all four with a lustre backend. Users are running into problems that changes to files on one protocol are not visible on the other protocol.

I have CTDB on the NFS servers and on the CIFS servers separately but that’s apparently not enough. Do I link all four servers together in CTDB or do I turn them into four NFS servers and run CIFS over NFS on two of those?

What’s the right way to proceed here?