> a satellite orbiting around the Earth is really going in a straight line but space is curved.
This is not quite how it works, I think. The gravity of the Earth isn't high enough to cause space to curve around in a circle, and the path of the satellite depends on its orbital velocity. If it fired a rocket forwards to reduce speed, it would fall towards the Earth, but that's not due to a change in space curvature.
The paths taken by light will follow the curvature, however. That would give a very small deflection for light passing near the Earth.
The gravity of Terra is exactly enough to cause the curved orbit. You are just missing how shallowly curved it really is, by squashing out the time part of spacetime. In 4 dimensional spacetime, the circle is (to simplify a little) really a helix, because it does not return to its starting point, and it's very stretched out along the time axis. (Or very narrow along the space axes, depending from how one choses one's units.) The curvature only looks greater because you are discounting how much greater length of time it extends across in order to do all of that curving that gets it all of the way around Terra.
The curvature is quite small but is responsible for orbital motion in the steady state condition. I don't think the parent comment was suggesting that space would "curve into a circle" unless you are implying something else specifically.
All ideal orbits, with no other forces involved, are equivalent to geodesics. However that only applies in true free fall. Actual orbits decay due to tiny non-ideal characteristics which have a cumulative effect or stochastic "butterfly effect"-like characteristic over long periods of time. For example in the ideal case, as soon as a satellite fires maneuvering thrusters, it is not following a geodesic for the time that the rockets are active. After the maneuver is completed, a new steady state condition can be calculated piecewise and then it's following a geodesic again. Another example is the ideal state of a satellite being in a low orbit that decays due to friction with the atmosphere, for which there are good estimates but not exact predictions. Due to the friction, there is a tiny acceleration (which happens to become larger over time) and this makes it follow something that starts close to a geodesic but deviates from it.
This is not quite how it works, I think. The gravity of the Earth isn't high enough to cause space to curve around in a circle, and the path of the satellite depends on its orbital velocity. If it fired a rocket forwards to reduce speed, it would fall towards the Earth, but that's not due to a change in space curvature.
The paths taken by light will follow the curvature, however. That would give a very small deflection for light passing near the Earth.