The use of industrial and domestic equipment is increasingly dependent on computerized control systems. This evolution awakens in the users the feeling of reliability of the equipment, which is not always achieved. However, system designers implement fault-tolerance methodologies and attributes to eliminate faults or any error in the system.Industrially, the increase in system reliability is achieved by the redundancy of control systems based on the replication of conventional and centralized programmable logic controllers. In distributed systems, reliability is achieved by replicating and distributing the most critical elements, leaving a single copy of the remaining components. On the other hand, given the nature of the distributed systems, it will also be necessary to ensure that the data set received by each of the replicas has the same order. Thus, any change in the order and data set received will result in different results, in each of the replicas, which may manifest in erroneous behavior.In this paper, the interactions and the erroneous behavior of the replicas are explained, depending on the data set received, in a fault tolerant distributed system. Its tendency, behavior and possible influences on reliability are presented, considering the failure rate and availability based on the mean time to failure.