My Morningstar Tristar PWM controller requires disconnecting the controller from the panels and the batteries after making DIP switch changes for them to take effect. Disconnect panels first, then battery. When hooking back up, hook to battery first and then to panels.
So while I would suspect that this could be necessary for the changes to take with the OP's controller, I don't know why it would indicate it is in float with an excessively high float voltage. OP has also not commented on if they have the temperature compensation sensor installed. While this sensor doesn't explain all of the behavior, it could at least point to a reason that the batteries are at a voltage that doesn't match a setpoint.
Idea: if the OP changed the DIP switches but didn't cycle power to the controller, and the controller requires this power cycle despite it not being mentioned in the manual, could the the controller be in a state where it just allows solar current/voltage to pass uncontrolled through the controller to the batteries? That could explain the high voltage reading at the batteries, it would be the uncontrolled panel voltage.