Beowulf cluster has created a lot of impact by delivering an excellent prize/performance and scalable computing systems to high performance computing community. Nevertheless, the widespread use of beowulf clusters has not happened as fast as it should be. One of the major problem is that the software tools on beowulf system is still in a very primitive state. Therefore, it is difficult for a small research group consisting mainly of non-computer scientists to build and maintain the cluster for their use. In this paper, many factors that contribute to the problem has been identified especially the unnescessary complexity that were built into the software. Some of the solutions to this problem have been proposed.