File System Performance Testing


One Line Summary

Discussion surrounding performance regression testing


This talk is aimed at fostering discussion about performance testing. There are several frameworks in place for automating tests, but no one seems to add tests, or run the test suites on a regular basis. Let’s discuss what the barriers are, and how we move forward to get more complete testing to find performance regressions sooner.


testing, performance, file systems


  • Jeff Moyer

    Red Hat, Inc.


    Jeff Moyer is a principal software engineer at Red Hat who has been hacking on Linux since 1995. While at Red Hat, he has worked on improving the Linux automounter, the netpoll API used by netconsole and netdump, the Linux in-kernel asynchronous and direct I/O code paths, and the block layer. He currently spends the bulk of his time tracking down bugs and performance problems in the block I/O path. Before coming to Red Hat, Jeff was involved in the design and implementation of a couple of different high availability clustering infrastructures.

  • Eric Sandeen

    Red Hat


    Eric has been developing Linux filesystems for about a decade, starting with XFS at SGI and now working on ext4, ext3, xfs, and other filesystem/storage related bits at Red Hat.

Leave a private comment to organizers about this proposal