My 3 metre solid dish is almost large enough to limit diffraction losses at 21cm, and gave reasonable results when I tried some quick tests last year, but steering that beast needs a lot of serious engineering, and it really needs to be bigger! One of my friends has a 4.5m solid dish and another has a 5.6m mesh dish. Those would be excellent RA tools, but they are used for moonbounce at 23cm. Booo!
HOWEVER.... Any antenna is better than not trying. Mesh dish kits are available from RF Hamdesign in the Netherlands, and 1.2m offset parabolic satellite dishes are readily available. Steering gets more challenging from an engineering perspective with size at about the cube of the diameter!
Another option to consider is an array of small dipole/reflector antennas feeding individual SDR receivers using a common local oscillator clock, then processing the resulting data streams to create a synthetic steerable phased array.
I've tried with a 90cm dish and a simple round "coffee can" feedhorn but the feed is huge and unwieldy on that little dish mount.
A set of four phased Yagis or loop yagis works OK for moonbounce at 23cm, and is fairly easy to build in a home workshop. Loop yagis are particularly simple to build.
Anyway, the point is to try SOMETHING, and see what results you can get, then when (not if) you get hooked and start measuring thermal noise from the moon and doing detailed 21cm mapping and all that good stuff, you can blame me for your garden becoming Jodrell.
One if my chums is selling a 2.4 metre mesh dish and some feeds. If anyone is interested, I can put you in touch. Location is Harrogate, North Yorkshire.
Neil