I agree with all the feedback you have been provided and had some other points for your consideration:
1. Remove two connections in your drawing…
You suggested in one response (~9:22 pm) that all your connections in your drawing are correct, except you needed to remove Vcc.
“minus the fact that I drew the line to the VCC instead of the JD-VCC”
Careful here…your drawing shows two ties between the pi and Relay Module…VCC and the common relay GND…which can be a path for the aforementioned EMF feedback.
Remove both the Vcc and GND connections between the Relay Module and the Rpi.
Summary…do not have any Vcc or GND connections between the pi and Relay Module.
2. Don’t trust the Relay Module Schematic…
After reviewing the schematic for the Relay Module, I would be concerned with the accuracy. It does not seem to make sense that the Relay Module header “J?” would have a pin tied to one opto isolator, rather one would expect that header “J?” would be the relay's power connection for JD-Vcc and GND?
Further thoughts, while your suggestion that relay power can be obtained from GPIO header, given only one relay will be under power at any given time, please do not consider this as a valid option and consider the following thoughts:
1. Your program locks up or is buggy which might enable more than one or all relays to energized at the same time, do you really want to service your pi?
2. What is the in-rush current (peak) that any relay might require?
3. Other risks:
a. Increase current, one of the relays develops a short or the clamping diode leaks?
b. Relay energized by board component failure, a transistor shorts and energizes one relay and maintains this state and then at some time your program enables another relay, net effect more than one relay energized simultaneously.
Consider all eventualities (the what if) for greater dependency and perhaps peace of mind…
