ceph/doc/rados/operations/data-placement.rst

   1 =========================
   2  Data Placement Overview
   3 =========================
   4
   5 Ceph stores, replicates, and rebalances data objects across a RADOS cluster
   6 dynamically. Because different users store objects in different pools for
   7 different purposes on many OSDs, Ceph operations require a certain amount of
   8 data- placement planning. The main data-placement planning concepts in Ceph
   9 include:
  10
  11 - **Pools:** Ceph stores data within pools, which are logical groups used for
  12   storing objects. Pools manage the number of placement groups, the number of
  13   replicas, and the CRUSH rule for the pool. To store data in a pool, it is
  14   necessary to be an authenticated user with permissions for the pool. Ceph is
  15   able to make snapshots of pools. For additional details, see `Pools`_.
  16
  17 - **Placement Groups:** Ceph maps objects to placement groups. Placement
  18   groups (PGs) are shards or fragments of a logical object pool that place
  19   objects as a group into OSDs. Placement groups reduce the amount of
  20   per-object metadata that is necessary for Ceph to store the data in OSDs. A
  21   greater number of placement groups (for example, 100 PGs per OSD as compared
  22   with 50 PGs per OSD) leads to better balancing. For additional details, see
  23   :ref:`placement groups`.
  24
  25 - **CRUSH Maps:**  CRUSH plays a major role in allowing Ceph to scale while
  26   avoiding certain pitfalls, such as performance bottlenecks, limitations to
  27   scalability, and single points of failure. CRUSH maps provide the physical
  28   topology of the cluster to the CRUSH algorithm, so that it can determine both
  29   (1) where the data for an object and its replicas should be stored and (2)
  30   how to store that data across failure domains so as to improve data safety.
  31   For additional details, see `CRUSH Maps`_.
  32
  33 - **Balancer:** The balancer is a feature that automatically optimizes the
  34   distribution of placement groups across devices in order to achieve a
  35   balanced data distribution, in order to maximize the amount of data that can
  36   be stored in the cluster, and in order to evenly distribute the workload
  37   across OSDs.
  38
  39 It is possible to use the default values for each of the above components.
  40 Default values are recommended for a test cluster's initial setup. However,
  41 when planning a large Ceph cluster, values should be customized for
  42 data-placement operations with reference to the different roles played by
  43 pools, placement groups, and CRUSH.
  44
  45 .. _Pools: ../pools
  46 .. _CRUSH Maps: ../crush-map
  47 .. _Balancer: ../balancer