ceph/doc/cephadm/adoption.rst

   1 .. _cephadm-adoption:
   2
   3 Converting an existing cluster to cephadm
   4 =========================================
   5
   6 Cephadm allows you to convert an existing Ceph cluster that
   7 has been deployed with ceph-deploy, ceph-ansible, DeepSea, or similar tools.
   8
   9 Limitations
  10 -----------
  11
  12 * Cephadm only works with BlueStore OSDs.  If there are FileStore OSDs
  13   in your cluster you cannot manage them.
  14
  15 Preparation
  16 -----------
  17
  18 #. Get the ``cephadm`` command line tool on each host in the existing
  19    cluster.  See :ref:`get-cephadm`.
  20
  21 #. Prepare each host for use by ``cephadm``::
  22
  23      # cephadm prepare-host
  24
  25 #. Determine which Ceph version you will use.  You can use any Octopus (15.2.z)
  26    release or later.  For example, ``docker.io/ceph/ceph:v15.2.0``.  The default
  27    will be the latest stable release, but if you are upgrading from an earlier
  28    release at the same time be sure to refer to the upgrade notes for any
  29    special steps to take while upgrading.
  30
  31    The image is passed to cephadm with::
  32
  33      # cephadm --image $IMAGE <rest of command goes here>
  34
  35 #. Cephadm can provide a list of all Ceph daemons on the current host::
  36
  37      # cephadm ls
  38
  39    Before starting, you should see that all existing daemons have a
  40    style of ``legacy`` in the resulting output.  As the adoption
  41    process progresses, adopted daemons will appear as style
  42    ``cephadm:v1``.
  43
  44
  45 Adoption process
  46 ----------------
  47
  48 #. Ensure the ceph configuration is migrated to use the cluster config database.
  49    If the ``/etc/ceph/ceph.conf`` is identical on each host, then on one host::
  50
  51      # ceph config assimilate-conf -i /etc/ceph/ceph.conf
  52
  53    If there are config variations on each host, you may need to repeat
  54    this command on each host.  You can view the cluster's
  55    configuration to confirm that it is complete with::
  56
  57      # ceph config dump
  58
  59 #. Adopt each monitor::
  60
  61      # cephadm adopt --style legacy --name mon.<hostname>
  62
  63    Each legacy monitor should stop, quickly restart as a cephadm
  64    container, and rejoin the quorum.
  65
  66 #. Adopt each manager::
  67
  68      # cephadm adopt --style legacy --name mgr.<hostname>
  69
  70 #. Enable cephadm::
  71
  72      # ceph mgr module enable cephadm
  73      # ceph orch set backend cephadm
  74
  75 #. Generate an SSH key::
  76
  77      # ceph cephadm generate-key
  78      # ceph cephadm get-pub-key > ceph.pub
  79
  80 #. Install the cluster SSH key on each host in the cluster::
  81
  82      # ssh-copy-id -f -i ceph.pub root@<host>
  83
  84 #. Tell cephadm which hosts to manage::
  85
  86      # ceph orch host add <hostname> [ip-address]
  87
  88    This will perform a ``cephadm check-host`` on each host before
  89    adding it to ensure it is working.  The IP address argument is only
  90    required if DNS does not allow you to connect to each host by its
  91    short name.
  92
  93 #. Verify that the adopted monitor and manager daemons are visible::
  94
  95      # ceph orch ps
  96
  97 #. Adopt all OSDs in the cluster::
  98
  99      # cephadm adopt --style legacy --name <name>
 100
 101    For example::
 102
 103      # cephadm adopt --style legacy --name osd.1
 104      # cephadm adopt --style legacy --name osd.2
 105
 106 #. Redeploy MDS daemons by telling cephadm how many daemons to run for
 107    each file system.  You can list file systems by name with ``ceph fs
 108    ls``.  For each file system::
 109
 110      # ceph orch apply mds <fs-name> <num-daemons>
 111
 112    For example, in a cluster with a single file system called `foo`::
 113
 114      # ceph fs ls
 115      name: foo, metadata pool: foo_metadata, data pools: [foo_data ]
 116      # ceph orch apply mds foo 2
 117
 118    Wait for the new MDS daemons to start with::
 119
 120      # ceph orch ps --daemon-type mds
 121
 122    Finally, stop and remove the legacy MDS daemons::
 123
 124      # systemctl stop ceph-mds.target
 125      # rm -rf /var/lib/ceph/mds/ceph-*
 126
 127 #. Redeploy RGW daemons.  Cephadm manages RGW daemons by zone.  For each
 128    zone, deploy new RGW daemons with cephadm::
 129
 130      # ceph orch apply rgw <realm> <zone> <placement> [--port <port>] [--ssl]
 131
 132    where *<placement>* can be a simple daemon count, or a list of
 133    specific hosts (see :ref:`orchestrator-cli-placement-spec`).
 134
 135    Once the daemons have started and you have confirmed they are functioning,
 136    stop and remove the old legacy daemons::
 137
 138      # systemctl stop ceph-rgw.target
 139      # rm -rf /var/lib/ceph/radosgw/ceph-*
 140
 141 #. Check the ``ceph health detail`` output for cephadm warnings about
 142    stray cluster daemons or hosts that are not yet managed.