In this paper, we consider a parallel repairable system operating in a finite state Markovian random environment with L identical components subject to failures and with N repairmen. Accidental failures arrive to each component following a Markovian arrival process. The repair times follow phase-type distributions. The multi-dimensional Markov chains describing the behavior of the system is investigated. The algorithm for calculating the stationary state probabilities is elaborated. Useful performance measures in reliability ...