Like a fine wine, Red Hat Ceph Storage (RHCS) gets better with age. During Red Hat Summit 2015, we announced the availability of RHCS 1.3, a release that brings with it improvements and tuning designed to please many an admin. Let’s take a look at what you can expect but, before we do, remember that you can test drive RHCS by visiting this link. Do it today.
Robustness at scale
Data is growing at a mind-boggling rate every year, so it makes sense that RHCS should be able to handle multi-petabyte clusters where failure is a fact of life. To mitigate that failure and enhance performance, RHCS now has:
- Improved automatic rebalancing logic that prioritizes degraded, rather than misplaced, objects
- Rebalancing operations that can be temporarily disabled so those operations don’t impede performance
- Scrubbing that can be scheduled to avoid disruption at peak times
- Object buckets that can be sharded to avoid hot-spots
RHCS has been tweaked and tuned to improve speed and increase I/O consistency. This includes optimizations for flash storage devices; read-ahead caching, which accelerates virtual machine booting in OpenStack; allocation hinting, which reduces XFS fragmentation to minimize performance degradation over time; and cache hinting, which preserves the cache’s advantages and improves performance.
New features in RHCS 1.3 help admins manage storage more efficiently. For example, Calamari, the product’s management platform, now supports multiple users and clusters. Civetweb server greatly simplifies deployment of Ceph Object Gateway: it now takes only two commands to fully install RGW. CRUSH management via Calamari API allows programmatic adjustment of placement policies, and faster block device operations make resize, delete, and flatten operations quicker.
Your next step