It would be helpful it you could post a link to your RAW file.
1. Manual settings... that's fine.
2. Focal length 24mm... fine.
3. 500 Rule ... maybe. That's for full-frame cameras. If you use an APS-C camera then divide 500 by the crop factor (1.6) or multiply the lens focal length by 1.6 before doing the math. For an APS-C camera it works out to 13 seconds.
4. ISO 2000 ... the best ISO depends on the camera model. You can always stretch in post-processing (which is basically what ISO gain does in the camera except it applies the stretch before writing the file.)
5. Aperture 2.8 ... fine (btw, was this "wide open" for that lens?
6. Focused on star... how did you do this? Can you provide some details on your method.
7. Steady tripod ... required... yep.
The image should not have been "black" and it made me wonder... did you have a lens cap on? Did you have a focus mask on the camera?
I usually tell people to crank up the ISO all the way just to get a shot (albeit with loads of noise) that you can use to confirm your compositional framing and focus are good... then dial the ISO back to a more sane value.
You should certainly have seen stars at ISO 2000 & f/2.8 (with any lens).
For example, here's a test shot using my 14mm f/2.8 at ISO 800 & f/4 and you can see lots of stars. (this is straight-out-of-the-camera other than the RAW to JPEG conversion and size reduction to be web-friendly).
That was 2.25 stops down in ISO and 1 stop down in f-stops but about 1/2 stop longer on exposure time. So this should be darker than what you got ... but you can see it's not "black". Your image *should* be brighter than mine. (this was a night trying capture Perseid meteors ... and shot from somewhat light-polluted skies). But the main point is that you can see the stars at this exposure.
And here's another (this is shot through an f/5.4 telescope) completely un-processed out of the camera shot. This is the Trifid Nebula. It's ISO 800, f/5.4 and it's a 4 minute shot. That's 2.25 stops slower on ISO, about 2 stops slower on f-stop (so that's 4.25 stops) but it's much longer. Still... if we multiple your 21 sec exposure time by 4.25 stops... it would be "as if" you could use my exposure settings but shoot for 420 seconds (and I only shot for 240 secs).
BTW, that was shot with my 60Da... which is a special edition of the 60D designed for astrophotography and is particularly sensitive to reds (hence the very nice sensitiivty to the red part of the nebula and the strong color cast on the stars). A processed version of this image would have a white balance correction... but this is straight out of the camera.
You should be seeing stars in your image. Clearly something is wrong.
Did you switch off the auto-focus on the lens? If the lens shifts focus... blurred stars will appear to more or less just vanish. This is, in part, why it would be good to post a link to the RAW file so we can inspect all the camera settings active at the time of the shot.
