Software-defined object storage project Ceph has issued a new release its makers say is a “major release” that “will be the foundation for the next long-term stable release.”
Ceph aims to provide storage that has no single point of failure, by spreading data across a cluster. The open-source project is of course free and aspires to scale to exabytes of storage.
The headline items in this new release are the first stable, feature-complete release of the Ceph Filesytem. That code offers snapshots, volume repair and disaster recovery tools, and experimental support for multiple ceph file systems across a single cluster.
Considerable effort has gone into greater compatibility with OpenStack's Swift storage API. Another small-but-notable change means the Ceph process is managed by systemd while Ceph's daemon now runs as a user named 'ceph' rather than as root. SHEC erasure coding has been upgraded from experimental status.
The project's trying to tidy up its development process to standardise on C++11 and is therefore dropping support for older operating systems. So CentOS 6 is gone, but CentOS 7.x is supported. So are Debian Jessie 8.x and Debian Wheezy 7.x. Ubuntu Precise 12.04 is outta here, but Ubuntu Xenial 16.04 and Trusty 14.04 are welcome, as are Fedora 22 or later and aarch64/arm64.
Ceph's post detailing all the changes in the new release is a little curious, as the project's release schedule says Jewel has long-term stable (LTS) status, but also says this release is a foundation for the next LTS release. We've asked for clarification on that overlap. ®