ceph/doc/start/intro.rst

   1 ===============
   2  Intro to Ceph
   3 ===============
   4
   5 Whether you want to provide :term:`Ceph Object Storage` and/or
   6 :term:`Ceph Block Device` services to :term:`Cloud Platforms`, deploy
   7 a :term:`Ceph Filesystem` or use Ceph for another purpose, all
   8 :term:`Ceph Storage Cluster` deployments begin with setting up each
   9 :term:`Ceph Node`, your network, and the Ceph Storage Cluster. A Ceph
  10 Storage Cluster requires at least one Ceph Monitor, Ceph Manager, and
  11 Ceph OSD (Object Storage Daemon). The Ceph Metadata Server is also
  12 required when running Ceph Filesystem clients.
  13
  14 .. ditaa::  +---------------+ +------------+ +------------+ +---------------+
  15             |      OSDs     | | Monitors   | |  Managers  | |      MDSs     |
  16             +---------------+ +------------+ +------------+ +---------------+
  17
  18 - **Monitors**: A :term:`Ceph Monitor` (``ceph-mon``) maintains maps
  19   of the cluster state, including the monitor map, manager map, the
  20   OSD map, and the CRUSH map.  These maps are critical cluster state
  21   required for Ceph daemons to coordinate with each other.  Monitors
  22   are also responsible for managing authentication between daemons and
  23   clients.  At least three monitors are normally required for
  24   redundancy and high availability.
  25
  26 - **Managers**: A :term:`Ceph Manager` daemon (``ceph-mgr``) is
  27   responsible for keeping track of runtime metrics and the current
  28   state of the Ceph cluster, including storage utilization, current
  29   performance metrics, and system load.  The Ceph Manager daemons also
  30   host python-based modules to manage and expose Ceph cluster
  31   information, including a web-based :ref:`mgr-dashboard` and
  32   `REST API`_.  At least two managers are normally required for high
  33   availability.
  34
  35 - **Ceph OSDs**: A :term:`Ceph OSD` (object storage daemon,
  36   ``ceph-osd``) stores data, handles data replication, recovery,
  37   rebalancing, and provides some monitoring information to Ceph
  38   Monitors and Managers by checking other Ceph OSD Daemons for a
  39   heartbeat. At least 3 Ceph OSDs are normally required for redundancy
  40   and high availability.
  41
  42 - **MDSs**: A :term:`Ceph Metadata Server` (MDS, ``ceph-mds``) stores
  43   metadata on behalf of the :term:`Ceph Filesystem` (i.e., Ceph Block
  44   Devices and Ceph Object Storage do not use MDS). Ceph Metadata
  45   Servers allow POSIX file system users to execute basic commands (like
  46   ``ls``, ``find``, etc.) without placing an enormous burden on the
  47   Ceph Storage Cluster.
  48
  49 Ceph stores data as objects within logical storage pools. Using the
  50 :term:`CRUSH` algorithm, Ceph calculates which placement group should
  51 contain the object, and further calculates which Ceph OSD Daemon
  52 should store the placement group.  The CRUSH algorithm enables the
  53 Ceph Storage Cluster to scale, rebalance, and recover dynamically.
  54
  55 .. _REST API: ../../mgr/restful
  56
  57 .. raw:: html
  58
  59         <style type="text/css">div.body h3{margin:5px 0px 0px 0px;}</style>
  60         <table cellpadding="10"><colgroup><col width="50%"><col width="50%"></colgroup><tbody valign="top"><tr><td><h3>Recommendations</h3>
  61
  62 To begin using Ceph in production, you should review our hardware
  63 recommendations and operating system recommendations.
  64
  65 .. toctree::
  66    :maxdepth: 2
  67
  68    Hardware Recommendations <hardware-recommendations>
  69    OS Recommendations <os-recommendations>
  70
  71
  72 .. raw:: html
  73
  74         </td><td><h3>Get Involved</h3>
  75
  76    You can avail yourself of help or contribute documentation, source
  77    code or bugs by getting involved in the Ceph community.
  78
  79 .. toctree::
  80    :maxdepth: 2
  81
  82    get-involved
  83    documenting-ceph
  84
  85 .. raw:: html
  86
  87         </td></tr></tbody></table>