I just had a little incident after a reboot where rpc.statd grabbed 993/tcp and effectively stopped mailboxd from running. I caught the problem and worked around it, but it made me wonder if anyone else has a workaround:

We use NFS (as a client) on our Zimbra server to store /opt/zimbra/backup, as well as for other services. It's useful and I'd rather not have to find another way (SMB, iSCSI, etc). I'm not completely sure that I need all the RPC services to allow NFS clients to function and, rather than test it and try to find out, I was wondering:
  • If anyone know if I can do without the RPC services started by portmap, idmap, nfslock and such and still have a functioning NFS client?
  • If anyone has a good way of ensuring ports that Zimbra needs (I'm thinking 993 and 995, typically) aren't stolen by RPC services? I can't find a way to do this without moving to RHEL6 as per this bug.