An overflow pipe would be nice, but I caution its incorrect use. It would be nice to have should you decide to place a ball valve between barrel one and barrel two. This way you could shut off the flow to barrel two and bypass it for when you need to service it (this should not be often at all, however).
It's continual use however might deflect more water than you care too from being cycled into the second barrel. I can't say exactly how much because of the many variables at play here. I see that you have changed your design from a 1" pipe to a 2" pipe. This will help but only address's one of the variables I brought up earlier in your bottom fed second barrel.
Let me try a different approach to explain what is happening (and did happen in your original design).
Have you ever noticed when buying a pump that flow rates are listed at different head heights? Did you notice that as the head height increases, flow rate decreases? Why is this?
As water fills the pipe and is directed uphill, the wieght (mass) of the water in the pipe increases. The higher up the water has to travel, the more wieght the water has ( due to an increase in volume). Additionally, not only the vertical distance the water has to travel effects the weight, but the size of the pipe as well. For instance, if you take a small pond pump of say 2000 GPH and feed it through a 2" pipe a distance of 8 feet up a 45 degree slope into a waterfall basin, you will probably get a pretty decent flow ( I am guessing around 1400 GPH depending on the pump). Now if you take that same pump and feed it into a 20 inch pipe up the same distance, you probably won't get much flow at all as the pump cannot overcome the weight of the water and its resulting backpressure.
Now, why is this all relevant to you filter design? Well, when you bottom feed a barrel and force the water up, your effectively pushig the water through a 20" pipe (the barrel). Now all this is fine and good on your first barrel, but your doing it twice. So your first barrell fills and then begins to fill the second barrel. As the weight of the water accumulates in the second barrel causing backpressure, your going to reach a point where it is easier for the water to simply overfill and overflow in the first barrel rather than to continue to fill the second (in theory). Now in all likelyhood, the second barrel will fill and work, but your going to have x amount of overflow in the first barrel as well.
Seeing how we would like to eliminate this, why not seal the top of the first barrel, or better yet, top feed the second barrel? Why are you so set on bottom feeding the second barrel????