I worked for a while for a municipal water department. We had a lot of deep wells (thousands of feet) many of the older ones used a shaft driven pump and large motor on the surface. (hundreds of HP) The shafts were very thin considering the power...like 1-1/2" maybe. I worked on the controls end, not the mechanics of it, so I only saw them at a distance. (Due to sanitation concerns my curiosity was not encouraged.)
But they were supported by many spider bearings all the way down. That took care of the radial vibration and deflection issues, which have been discussed in this thread. Also care was taken to keep them irregularly spaced, to separate the resonance of each segment.
What hasn't been mentioned is torsional deflection. With the well pumps, we had a several minute delay between shut-down and restart. This was to allow the shafts to unwind, which they would go through several cycles clockwise and anti-clockwise before finally coming to rest. If you tried starting a motor while the shaft was wound-up the opposite direction, you could end up twisting it in two.
While the subject shaft isn't going to be twisting several turns under load like those half-mile long (more in many cases) pump shafts did, it will still be a torsion spring, and the VAWT and the generator are going to have significant moments of inertia, so there is a possibility that you will have torsional resonance issues...especially if the load changes abruptly, like if you short the output to lock the mill out in high winds, or even switching in a load dump arraignment. Just something to be aware of, and might be good to keep in mind if you have troubles down the road.