So here's what I think went wrong in parts of the western half of North Carolina with the precipitation amounts. The first map from Thurs at 10AM shows how the core zone of the 850mb warm advection (low level overrunning) stayed to the south from north Bama to SE North Carolina. This is the region that received the solid precip shield. NC missed out on that forcing for precipitation. The second map from Thurs 4PM shows how a large part of NC has flipped over to cold air advection at 850mb. Warm advection at 850 produces lift (for precip production), cold advection does not. Along with it being cold advection, there is also a downsloping component there with it coming across the mtns. The models certainly showed this all along and adjusted downward with QPF late in the game, but were limited all along on the NW side. With this being a weak wave and not a very dynamic system aloft, it was a must for the western 1/2 of NC to get in on this overrunning precip. The eastern 1/2 of NC was able to benefit from other processes leading to precip production higher up in the atmosphere - that is, increasing mid-level frontogenesis and upper level divergence as the storm matured and the final trailing upper wave strengthened a little as it moved thru. The SPC archive maps are really good to view in post analysis - https://www.spc.noaa.gov/exper/ma_archive/
In summary, the text above further supplements the idea that it is much preferred to have cold air in place prior to the storm's arrival. Not only is it preferred for precip type and pre-storm cold for efficient accumulation when precip arrives, it also helps with lift in the atmosphere as you want warmer air from the south running into colder air to the north to generate overrunning precip - the kind that produces the nice, consistent shield of precip on radar.