Not sure if it is specific to 2072A, you can test whatever model you have.
If a channel was on 50 Ohm input impedance setting, on power up the setting has no effect, i.e. the input port is still on high impedance though the menu shows the port is on 50 Ohm.
It seems the scope should be powered off for a while for the bug to be reproduced. Just power cycling does not reproduce the bug. My scope was turned off for 3 days or so. Today I used it, connected a signal gen and noticed the wrong amplitude swing, twice as it should be. The port was on 50 Ohm on power up. Toggling the port impedance in the menu fixed the issue. I am going to leave the scope powered off for the night and then will retest.