Abstract
Ceph, like many other distributed systems, has traditionally stored data on local file systems like XFS or btrfs. However, the mismatch between Ceph’s requirements and the POSIX interface provided by kernel file systems has a huge performance cost and requires a lot of complexity to provide data safety. BlueStore, an entirely new OSD storage backend, utilizes block devices directly, doubling performance for most workloads.
This talk will cover the motivation a new backend, it’s overall design, the improved performance on HDDs, SSDs, and NVMe, and discuss some of the thornier issues we had to overcome when replacing tried and true kernel file systems with entirely new code running in userspace.