The high adoption of Internet services and cloud computing leads to increaseddemands for reliable high performance storage in data centers. DistributedParallel Fault Tolerant File Systems (DPFTFS) is the evolution of traditionalDistributed File Systems (DFS) trying to address these increased demands.GlusterFS is a relatively new open source DPFTFS gaining lots of momentumlately and this thesis is dedicated to explore its overall performance and reliabil-ity. The experiments conducted by scaling up the number of concurrent clientsaccessing the le system, while generating high workloads at the same time.Di erent metrics are measured in both servers and clients including networkbandwidth usage and system load, giving an overview of performance underdi erent setups using the Elastic Volume Management feature of GlusterFS.The data integrity on these setups under di erent failing scenarios was alsotested providing insight of the reliability. Results show the sustained perfor-mance of GlusterFS with the increased number of connected clients, and betterwrite instead of read performance for the distributed only and striped only vol-ume types. Data loss and corruption has also been observed in these types ofvolume correspondingly. In a setup combining distributed and replication vol-ume types, no data loss or corruption and better read instead of write resultshave been obtained. The experiment exposed di erent bottlenecks related tothe network capacity, topology and hardware heterogeneity, giving an indica-tion of factors need to be taken into consideration before the deployment of aGlusterFS storage solution.