@Piotr et al
Align issues are usually related to bad scale and position estimates (inputs to the solver). Often unchecking "Use Scale" and "Use Position" in the Align settings can fix the problem (though possibly slow down the solving).
It is also possible that the star detection parameters are bad, as a reasonable number of stars need to be detected to solve.
Also, check the fits header of a file that doesn't solve and see if the relevant values look approximately ok (e.g. the reported ra/dec, focal length, sensor size).
StellarSolver comes with a StellarSolverTester tool that Rob provides to help debug solving. You should try it out and see if that can help--e.g. load a problematic file, see if stars are detected well, disable use position, use scale, ... You might then copy over the star detection parameters.
It can also be interesting to have the image solved by the online astrometry.net (
nova.astrometry.net/upload).
With regards to logging...around a year ago, guessing that's 3.5.8 or so, there were many reports of KStars crashing. I narrowed many of the crashes down to memory issues in the StellarSolver. (Internally, stellarSolver uses older professional Astronomy C libraries that are good at detecting stars but probably not written/tested for these kind of parallel computations.) Despite spending a ton of time debugging, I was unable to find the culprit code. However, I disabled some of the parallelism in star detection and disabled the logging and the crashes went away. Honestly the logs weren't that informative in helping resolve normal user issues, and the system has been much more stable since. it was well worth the tradeoff (applied to 3.5.9 and on I believe, or perhaps one release later).
If you are unable to figure out why you can't solve, please post a link to a fits file that doesn't solve, along with your system parameters (e.g. focal length, pixel size, where you were pointing) and I can take a look.
Hy