While initially targeted as a MySQL tool, the sysbench program is also valuable for testing low-level aspects of performance that impact databases, ones that apply equally well to PostgreSQL. It even supports running read-only tests against PostgreSQL databases too, a feature not demonstrated here. Its use in this chapter is just for its low-level hardware tests.
These specific sysbench tests should be considered secondary to the ones already shown. If you have confirmed that the sequential read, write, and seek speeds are good with Bonnie++, and know commit rates are good from doing INSERT tests with PostgreSQL, you really don't need this data too. It's mainly useful as an alternate way to do more thorough and specific seek and ...
