I don't have a vested interest in Harris or OpenSky, but all of the poor coverage issues I've investigated have been protocol agnostic (doesn't matter if it was analog, P25, OpenSky, conventional, trunked, whatever) one of two factors: 1) poor site engineering (site placement or desensitization); and 2) decisions to cut necessary sites or spread them thinner than necessary for robust coverage. A good system costs money. One municipality built a single-site 800 MHz trunked system and "saved money" (and achieved more net loss) by using 300 ft. of 7/8" cable, which completely screwed their link budget (it's all a mathematical relationship between losses and gains and 300 ft. of even 7/8" cable loses about half of the power from one end to the other at 800 MHz). Another didn't want to light their towers and wanted to appease a political group by putting up 180 ft. towers (which fell below the 200 ft. FAA lighting/marking requirements). Surprise, surprise, all of their towers had other factors which required post-construction lighting, and the towers were woefully too short to cover the area and terrain needed.
Bottom line: whatever is going on is probably related to HOW things were put together more than what protocol the system uses. In just about all cases, the failure is a non-engineer/non-technical type who cut something critical to either meet sales margin (if it was the vendor/manufacturer) or made an indiscriminate cut to something critical to meet municipal budget.
The light activation is not unique to Harris. When Motorola first came out with their 100 W Spectra, the databus to the DEK units (push buttons that can be programmed to operate relays and work the Motorola siren package) would pick up stray RF inside the vehicle and mess up operation activating lights and siren when the mic was pressed. The mic cord was acting as an antenna. A snap on ferrite bead (one of those computer ones) "fixed" the problem.
