ceph/doc/cephfs/administration.rst

   1
   2 CephFS Administrative commands
   3 ==============================
   4
   5 Filesystems
   6 -----------
   7
   8 These commands operate on the CephFS filesystems in your Ceph cluster.
   9 Note that by default only one filesystem is permitted: to enable
  10 creation of multiple filesystems use ``ceph fs flag set enable_multiple true``.
  11
  12 ::
  13
  14     fs new <filesystem name> <metadata pool name> <data pool name>
  15
  16 ::
  17
  18     fs ls
  19
  20 ::
  21
  22     fs rm <filesystem name> [--yes-i-really-mean-it]
  23
  24 ::
  25
  26     fs reset <filesystem name>
  27
  28 ::
  29
  30     fs get <filesystem name>
  31
  32 ::
  33
  34     fs set <filesystem name> <var> <val>
  35
  36 ::
  37
  38     fs add_data_pool <filesystem name> <pool name/id>
  39
  40 ::
  41
  42     fs rm_data_pool <filesystem name> <pool name/id>
  43
  44
  45 Settings
  46 --------
  47
  48 ::
  49
  50     fs set <fs name> max_file_size <size in bytes>
  51
  52 CephFS has a configurable maximum file size, and it's 1TB by default.
  53 You may wish to set this limit higher if you expect to store large files
  54 in CephFS. It is a 64-bit field.
  55
  56 Setting ``max_file_size`` to 0 does not disable the limit. It would
  57 simply limit clients to only creating empty files.
  58
  59
  60 Maximum file sizes and performance
  61 ----------------------------------
  62
  63 CephFS enforces the maximum file size limit at the point of appending to
  64 files or setting their size. It does not affect how anything is stored.
  65
  66 When users create a file of an enormous size (without necessarily
  67 writing any data to it), some operations (such as deletes) cause the MDS
  68 to have to do a large number of operations to check if any of the RADOS
  69 objects within the range that could exist (according to the file size)
  70 really existed.
  71
  72 The ``max_file_size`` setting prevents users from creating files that
  73 appear to be eg. exabytes in size, causing load on the MDS as it tries
  74 to enumerate the objects during operations like stats or deletes.
  75
  76
  77 Taking the cluster down
  78 -----------------------
  79
  80 Taking a CephFS cluster down is done by reducing the number of ranks to 1,
  81 setting the cluster_down flag, and then failing the last rank. For example:
  82
  83 ::
  84     ceph fs set <fs_name> max_mds 1
  85     ceph mds deactivate <fs_name>:1 # rank 2 of 2
  86     ceph status # wait for rank 1 to finish stopping
  87     ceph fs set <fs_name> cluster_down true
  88     ceph mds fail <fs_name>:0
  89
  90 Setting the ``cluster_down`` flag prevents standbys from taking over the failed
  91 rank.
  92
  93 Daemons
  94 -------
  95
  96 These commands act on specific mds daemons or ranks.
  97
  98 ::
  99
 100     mds fail <gid/name/role>
 101
 102 Mark an MDS daemon as failed.  This is equivalent to what the cluster
 103 would do if an MDS daemon had failed to send a message to the mon
 104 for ``mds_beacon_grace`` second.  If the daemon was active and a suitable
 105 standby is available, using ``mds fail`` will force a failover to the standby.
 106
 107 If the MDS daemon was in reality still running, then using ``mds fail``
 108 will cause the daemon to restart.  If it was active and a standby was
 109 available, then the "failed" daemon will return as a standby.
 110
 111 ::
 112
 113     mds deactivate <role>
 114
 115 Deactivate an MDS, causing it to flush its entire journal to
 116 backing RADOS objects and close all open client sessions. Deactivating an MDS
 117 is primarily intended for bringing down a rank after reducing the number of
 118 active MDS (max_mds). Once the rank is deactivated, the MDS daemon will rejoin the
 119 cluster as a standby.
 120 ``<role>`` can take one of three forms:
 121
 122 ::
 123
 124     <fs_name>:<rank>
 125     <fs_id>:<rank>
 126     <rank>
 127
 128 Use ``mds deactivate`` in conjunction with adjustments to ``max_mds`` to
 129 shrink an MDS cluster.  See :doc:`/cephfs/multimds`
 130
 131 ::
 132
 133     tell mds.<daemon name>
 134
 135 ::
 136
 137     mds metadata <gid/name/role>
 138
 139 ::
 140
 141     mds repaired <role>
 142
 143
 144 Global settings
 145 ---------------
 146
 147 ::
 148
 149     fs dump
 150
 151 ::
 152
 153     fs flag set <flag name> <flag val> [<confirmation string>]
 154
 155 "flag name" must be one of ['enable_multiple']
 156
 157 Some flags require you to confirm your intentions with "--yes-i-really-mean-it"
 158 or a similar string they will prompt you with. Consider these actions carefully
 159 before proceeding; they are placed on especially dangerous activities.
 160
 161
 162 Advanced
 163 --------
 164
 165 These commands are not required in normal operation, and exist
 166 for use in exceptional circumstances.  Incorrect use of these
 167 commands may cause serious problems, such as an inaccessible
 168 filesystem.
 169
 170 ::
 171
 172     mds compat rm_compat
 173
 174 ::
 175
 176     mds compat rm_incompat
 177
 178 ::
 179
 180     mds compat show
 181
 182 ::
 183
 184     mds getmap
 185
 186 ::
 187
 188     mds set_state
 189
 190 ::
 191
 192     mds rmfailed
 193
 194 Legacy
 195 ------
 196
 197 The ``ceph mds set`` command is the deprecated version of ``ceph fs set``,
 198 from before there was more than one filesystem per cluster. It operates
 199 on whichever filesystem is marked as the default (see ``ceph fs
 200 set-default``.)
 201
 202 ::
 203
 204     mds stat
 205     mds dump  # replaced by "fs get"
 206     mds stop  # replaced by "mds deactivate"
 207     mds set_max_mds # replaced by "fs set max_mds"
 208     mds set # replaced by "fs set"
 209     mds cluster_down  # replaced by "fs set cluster_down"
 210     mds cluster_up  # replaced by "fs set cluster_up"
 211     mds newfs # replaced by "fs new"
 212     mds add_data_pool # replaced by "fs add_data_pool"
 213     mds remove_data_pool #replaced by "fs remove_data_pool"
 214