Yeah, that didn’t work…
Eighty four images representing four and a half hours of exposure on the target. In this case NGC2244, the Rosette Nebula. It takes a few minutes to setup the software to align all of the images, then it takes about half an hour for the software to digest everything and produce a result. At the end of this you get to see if the effort worked. Sometimes it does not.
Complicating the issue is that a number of the sub-frames are taken at different exposures. Finding the correct stars in the these different exposures gives the software fits.

Daunting…and I was under the impression that this was second nature for you. I may have to be content with single raw frames in Aperture for a long time…
I continually find new ways to screw it up. In this case the only things lost were time and effort. I ended up processing each of the different exposure sets separately, then combining. That worked.