Fog computing is considered a promising technology to reduce latency and network congestion. Meanwhile, Millimeter-wave (mmWave) communication owing to its potential for multi-gigabit of wireless channel capacity could be employed to further improve the performance of fog computing networks. In this context, we study the feasibility of using 28 GHz and 38 GHz mmWaves in fog radio access networks (F-RANs). The multi-slope path loss model is used to calculate the interference impacts because it provides a more accurate approximation of the wireless links. Simulations are carried out for uplink scenario considering the following fog node (FN) deployment models: Poisson point process (PPP), Ginibre point process (GPP), square grid, and ultra-dense network (UDN). The results depict that at low FN densities the massive accumulation of interference components severely impacts the performance. However, the performance can be improved by increasing the FN density and choosing a deployment strategy with high degree of regularity. Based on the results, we verify that it is feasible to use 28 GHz and 38 GHz mmWaves in F-RANs when the density of the interfering users is less than 150 user/km2 where capacities higher than 1 Gbps are achieved for all considered scenarios.