In this chapter, we looked at an overview of the implementation done in the previous Chapter 4, Using GlusterFS on the Cloud Infrastructure, so that we could have a fresh understanding of what was implemented in order to understand how we could test performance. Given the previous setup, the implementation should be capable of a theoretical 375 MB/s of throughput. We can validate this number with several tools that work at different levels.
For ZFS volumes, we can use the zpool iostat command, which provides data for each of the block devices that are part of the ZFS volume. iostat can be used to determine performance for all of the block devices present in the system. These commands can only be run on each of the nodes of the cluster. To be able to verify the actual performance of the implementation, we used the FIO tool, which can simulate specific workloads by changing...