Ok so I bit the bullet and moved from core to scale. My plan was to migrate to Dragonfish, the last supported in-place upgrade, and then move up from there.
After the upgrade, SMB is working well. NFS however, not so much.
The NFS service is refusing to start, and the the issue appears to be with rpcbind.
rpcbind is running as `/sbin/rpcbind -f -w -h 10.0.0.241`. At one time, long past that was the address of my Core system, but its been 192.168.20.241 for at least 5 years now.
Googling around I see that the problem appears to be a host=10.0.0.241 entry in /etc/nfsd.conf.d/local.conf.
I removed that and rebooted, but it made no difference. Also tried change it to 0.0.0.0 with no effect. After the restart the host= entry returns like a bad smell.
After more searching it seems that the address to bind to is now specified in /usr/lib/systemd/system/rpcbind.socket, but that file is as vanilla as could be and does not mention the 10.0.0.241 address anywhere.
Does anyone have a clue where this rpcbind parameter might be coming from?