Recently, LoRaWAN has attracted much attention for the realization of many Internet of Things applications because it offers low-power, long-distance, and low-cost wireless communication. Recent works have shown that the LoRaWAN specification for class A devices comes with scalability limitations due to the ALOHA-like nature of the MAC layer. In this paper, we propose a synchronization and scheduling mechanism for LoRaWAN networks consisting of class A devices. The mechanism runs on top of the LoRaWAN MAC layer. A central network synchronization and scheduling entity will schedule uplink and downlink transmissions. In order to reduce the synchronization packet length, all time slots that are being assigned to an end node are encoded in a probabilistic space-efficient data structure. An end node will check if a time slot is part of the received data structure in order to determine when to transmit. Time slots are assigned based on the traffic needs of the end nodes. We show that in case of a nonsaturated multichannel LoRaWAN network with synchronization being done in a separate channel, the packet delivery ratio (PDR) is easily 7% (for SF7) to 30% (for SF12) higher than in an unsynchronized LoRaWAN network. For saturated networks, the differences in PDR become more profound as nodes are only scheduled as long as they can be accommodated given the remaining capacity of the network. The synchronization process will use less than 3-mAh extra battery capacity per end node during a one year period, for synchronization periods longer than three days. This is less than the battery capacity used to transmit packets that are going to be lost in an unsynchronized network due to collisions.