Yes it is perfecly fine to have antenna as low as 2m, but what has to be noted is that you are getting substantial amount of attenuation by ground proximity.
If you have the antenna 1/8 wavelength high, so on 40m band it is 5m, you are getting NVIS patern but not much loss. If you go higher, the elevation agle gets lower and you will get closer and closer to freespace pattern and gain. If you go lower you are getting more and more ground attenuation.
That is the theory, you would say, because you are making contacts - yes you do, but maybe you work only 50% chasers you could have worked if you had the antenna at 5m. Of course this might not apply in countries with low chaser density, so you maybe work them all no matter what your setup is, but in Europe where you can easily make 20-40 chasers in half an hour, it makes a difference.
To support this, I give an example: When I activate SOTA with kit transceiver running of low sensitivity NA612 mixers and having no preamp, I work about 10-15 chasers. With my KX3 I am getting 20-30 chasers on average. But if there were only 10 chasers out there, I would be likely to get all 10 on both setups no matter how I improved my station. So during your field test consider if your sample is statistically sufficient.
But dont get my wrong I have nothing against low antennas, it is only about what are your expectations. If you want rapid-deployment (RADAR) and do not care about how many contacts you make as long as you activate the summit, this is the way to go. But if you are like me and enjoy little pileup, at expense of setup taking longer, higher antenna is way to go. We should say all the facts and let everyone decide what are their expectations.
Lastly, a bit of reading on the NVIS height topic with a graph: