Which version of real flow do you have? There is a sheeting algorithm in the newer versions that makes all the difference in the look of things especially where it clings to smooth things.
As for length, I think it might take quite a while (think days) to run the mesh in RF for that length of time. As the frames increase there is more history to consider when calculating the particles so it tends to get slower as it goes. Normally you can engage a loop function but that tends to be for a fountain type arrangement where there is a constant stream of new particles being generated and old ones exiting the build zone, but in your case I think you want to generate a block of particles and then turn the emitter off.
I think you are going to have drag the attractor (or a few of them) around over time to keep the flow going. The most important factor will probably be the friction setting between liquid and sphere, too high and the liquid would rather grind to a halt. Keeping the liquid moving is also good for the illusion: as the particles settle they can wiggle as they try and find their lowest energy – you won’t see this if you keep it dynamic. There used to be “use geometry” instead of center of mass for gravity or attraction. I haven’t used that feature in a while so I’m not sure if that is still part of the package, but if it is you could use hollow arcs bigger and beyond your sphere to attract the liquid.
Another approach would be increase the friction and surface tension and then rotate the sphere in short quick bursts – the liquid will “stick” and get flung around and then flow each time the sphere stops.