Ceph: object storage, block storage, file system, replication, massive scalability, and then some!
|Wiki Page:||Ceph: object storage, block storage, file system, replication, massive scalability, and then some!|
Ceph is one of the most exciting new technologies to recently emerge in the Linux storage space. Based on the RADOS object store, the Ceph stack boasts massive scalability and high availability using nothing but commercial, off-the shelf hardware and free and open source software.
Ceph includes a massively distributed filesystem (Ceph FS), a striped, replicated, highly available block device (RADOS block device, RBD), S3 and Swift object storage capability through the RESTful RADOS Gateway, and a simple, well-documented native API with language bindings for C, C++ and Python. The Ceph filesystem and RBD have been part of the mainline kernel since the 2.6.3x releases, and the server-side stack has recently undergone an extensive cleanup and stabilization phase.
The Ceph stack is also well integrated into OpenStack, making it a potential "one-stop shop" for OpenStack object, image and volume storage.
In this hands-on tutorial, Florian and Tim will walk you through the initial setup of a Ceph cluster, explore its capabilities, highlight its most important features and identify current shortcomings, discuss performance considerations, and identify common Ceph failure modes and troubleshooting steps.
Attendees should have a good understanding of Linux systems administration. Prior experience with distributed storage (like Lustre, GlusterFS, DRBD, Swift) is a plus but not required. Prior knowledge of the Ceph stack is not necessary. Attendees will have the opportunity to follow the tutorial in a virtual Ceph cluster of their own; pre-installed Libvirt/KVM virtual images will be available for that purpose.
Co-presenter: Tim Serong, SUSE
Florian is a Linux high availability and storage specialist, experienced technical consultant, seasoned training instructor, and technical writer. He frequently consults and conducts training on both OpenStack and the Ceph stack.
Tim is currently employed by SUSE as Senior Clustering Engineer, working on the SUSE Linux Enterprise High Availability Extension and the SUSE Cloud product, which is based on OpenStack. Through his work on HA clustering, he discovered that you can never ever have too many log files. His recent work with OpenStack has caused him to doubt this assessment. Interests include motorcycling, beer, and making technology simpler and more reliable (although not necessarily in that order, and usually not all at the same time).