I work mostly as a side man. I play whatever volume the guy that hires me tells me to play. Everybody wants something different. If you want to work you have to give them want they ask for. That way you get asked back.

It sounds to me like the real problem is understanding how to deal with monitors. The singer absolutely has to hear himself as not being able to hear well leads to over singing and over singing leads to serious and eventually permanent vocal damage. The reality is, using a powered mixer is going to give you very limited control over gain before feedback. Make sure he is using a microphone with a very limited pattern. Then, begin to get an understanding of how EQ really works. At the very least you need a 1/3 octave EQ patched to the monitor output. Possibly a better and more controlled approach would be to utilize a parametric EQ that will allow you to really tailor what frequencies you are cutting and the width of the Q on that.

My suggestion is to work out the feedback issue as that should not be a problem if you know what to do. Your singer will never be successful until that is sorted out. Then use the suggestions presented here to dial in the overall mix.

For rehearsal, maybe the singer could try headphones. A little extreme, but it will fix the problem for him. Or...and this is really extreme, but since this is a vintage forum, you could try a REALLY retro audio approach: tape two microphones together, wire one out of phase and tell the singer to sing into only one of them. You have effectively created a noise canceling mic. You'll find lots of old photos of the Grateful Dead, The Who, etc. using that technique in the late 60's and early 70's. It's goofy but it works. I would not play live that way though because, like I said, it's goofy but it worked for them when the technology did not exist to play that loud or when you placed the entire PA behind you like Dead's famed Wall of Sound.