in case someone else faces this problem…
looks like all servers restart are required…
I suspect that they change the replica set DNS and that’s cached somewhere…
Separately, Support is really bad… it say back after 2 days… you expect us to have production system down for 2 days?