Release Notes

Dell PowerEdge R730xd Performance and Sizing Guide for Red Hat Ceph Storage - A Dell Red Hat Technical White Paper 15
Replicated storage pools: Replication makes full copies of stored objects, and is ideal for quick recovery. In
a replicated storage pool, Ceph configuration defaults to a replication factor of three, involving a primary
OSD and two secondary OSDs. If two of the three OSDs in a placement group become unavailable, data
may be read, but write operations are suspended until at least two OSDs are operational.
Erasure-coded storage pools: Erasure coding provides a single copy of data plus parity, and it is useful for
archive storage and cost-effective durability and availability. With erasure coding, storage pool objects are
divided into chunks by using the n=k+m notation, where
k is the number of data chunks that are created,
m is the number of coding chunks that will be created to provide data protection, and n is the total
number of chunks placed by CRUSH after the erasure coding process. So for instance,
n disks are needed
to store
k disks worth of data with data protection and fault tolerance of m disks.
Ceph block storage is typically configured with 3x replicated pools and is currently not supported directly
on erasure-coded pools. Ceph object storage is supported on either replicated or erasure-coded pools.
Depending on the performance needs and read/write mix of an object storage workload, an erasure-
coded pool can provide an extremely cost effective solution while meeting performance requirements.
For more information on Ceph architecture, see the Ceph documentation at
http://docs.ceph.com/docs/master/architecture/.