UserUtilities.xml

   1 <?xml version="1.0" encoding="utf-8"?>
   2 <chapter xmlns="http://docbook.org/ns/docbook"
   3 xmlns:xl="http://www.w3.org/1999/xlink" version="5.0" xml:lang="en-US"
   4 xml:id="userutilities"
   5 xmlns:xi="http://www.w3.org/2001/XInclude">
   6   <title xml:id="userutilities.title">User Utilities</title>
   7   <para>This chapter describes user utilities.</para>
   8   <section xml:id="dbdoclet.50438206_94597">
   9     <title>
  10       <indexterm>
  11         <primary>lfs</primary>
  12       </indexterm>
  13       <literal>lfs</literal>
  14     </title>
  15     <para>The
  16     <literal>lfs</literal> utility can be used for user configuration routines
  17     and monitoring.</para>
  18     <section remap="h5">
  19       <title>Synopsis</title>
  20       <screen>
  21 lfs
  22 lfs changelog [--follow] <replaceable>mdt_name</replaceable> [startrec [endrec]]
  23 lfs changelog_clear <replaceable>mdt_name id endrec</replaceable>
  24 lfs check <replaceable>mds|osts|servers</replaceable>
  25 lfs data_version [-nrw] <replaceable>filename</replaceable>
  26 lfs df [-i] [-h] [--pool]-p <replaceable>fsname</replaceable>[.<replaceable>pool</replaceable>] [<replaceable>path</replaceable>] [--lazy]
  27 lfs find [[!] --atime|-A [-+]N] [[!] --mtime|-M [-+]N]
  28          [[!] --ctime|-C [-+]N] [--maxdepth|-D N] [--name|-n <replaceable>pattern</replaceable>]
  29          [--print|-p] [--print0|-P] [[!] --obd|-O <replaceable>ost_name</replaceable>[,<replaceable>ost_name...</replaceable>]]
  30          [[!] --size|-S [+-]N[kMGTPE]] --type |-t {bcdflpsD}]
  31          [[!] --gid|-g|--group|-G <replaceable>gname|gid</replaceable>]
  32          [[!] --uid|-u|--user|-U <replaceable>uname|uid</replaceable>]
  33          <replaceable>dirname|filename</replaceable>
  34 lfs getname [-h]|[path...]
  35 lfs getstripe [--obd|-O <replaceable>ost_name</replaceable>] [--quiet|-q] [--verbose|-v]
  36               [--stripe-count|-c] [--stripe-index|-i]
  37               [--stripe-size|-s] [--pool|-p] [--directory|-d]
  38               [--mdt-index|-M] [--recursive|-r] [--raw|-R]
  39               [--layout|-L]
  40               <replaceable>dirname|filename</replaceable> ...
  41 lfs setstripe [--size|-s stripe_size] [--stripe-count|-c <replaceable>stripe_count</replaceable>]
  42               [--overstripe-count|-C <replaceable>stripe_count</replaceable>]
  43               [--stripe-index|-i <replaceable>start_ost_index</replaceable>]
  44               [--ost-list|-o <replaceable>ost_indicies</replaceable>]
  45               [--pool|-p <replaceable>pool</replaceable>]
  46               <replaceable>dirname|filename</replaceable>
  47 lfs setstripe -d <replaceable>dir</replaceable>
  48 lfs osts [path]
  49 lfs pool_list <replaceable>filesystem</replaceable>[.<replaceable>pool</replaceable>]| <replaceable>pathname</replaceable>
  50 lfs quota [-q] [-v] [-h] [-o <replaceable>obd_uuid</replaceable>|-I <replaceable>ost_idx</replaceable>|-i <replaceable>mdt_idx</replaceable>]
  51           [-u <replaceable>username|uid|-g</replaceable> <replaceable>group|gid</replaceable>|-p <replaceable>projid</replaceable>] <replaceable>/mount_point</replaceable>
  52 lfs quota -t -u|-g|-p <replaceable>/mount_point</replaceable>
  53 lfs quotacheck [-ug] <replaceable>/mount_point</replaceable>
  54 lfs quotachown [-i] <replaceable>/mount_point</replaceable>
  55 lfs quotainv [-ug] [-f] <replaceable>/mount_point</replaceable>
  56 lfs quotaon [-ugf] <replaceable>/mount_point</replaceable>
  57 lfs quotaoff [-ug] <replaceable>/mount_point</replaceable>
  58 lfs setquota {-u|--user|-g|--group|-p|--project} <replaceable>uname|uid|gname|gid|projid</replaceable>
  59              [--block-softlimit <replaceable>block_softlimit</replaceable>]
  60              [--block-hardlimit <replaceable>block_hardlimit</replaceable>]
  61              [--inode-softlimit <replaceable>inode_softlimit</replaceable>]
  62              [--inode-hardlimit <replaceable>inode_hardlimit</replaceable>]
  63              <replaceable>/mount_point</replaceable>
  64 lfs setquota -u|--user|-g|--group|-p|--project <replaceable>uname|uid|gname|gid|projid</replaceable>
  65              [-b <replaceable>block_softlimit</replaceable>] [-B <replaceable>block_hardlimit</replaceable>]
  66              [-i <replaceable>inode-softlimit</replaceable>] [-I <replaceable>inode_hardlimit</replaceable>]
  67              <replaceable>/mount_point</replaceable>
  68 lfs setquota -t -u|-g|-p [--block-grace <replaceable>block_grace</replaceable>]
  69              [--inode-grace <replaceable>inode_grace</replaceable>]
  70              <replaceable>/mount_point</replaceable>
  71 lfs setquota -t -u|-g|-p [-b <replaceable>block_grace</replaceable>] [-i <replaceable>inode_grace</replaceable>]
  72              <replaceable>/mount_point</replaceable>
  73 lfs help
  74 </screen>
  75       <note>
  76         <para>In the above example, the
  77         <literal>
  78           <replaceable>/mount_point</replaceable>
  79         </literal> parameter refers to the mount point of the Lustre file
  80         system.</para>
  81       </note>
  82       <note>
  83         <para>The old lfs quota output was very detailed and contained
  84         cluster-wide quota statistics (including cluster-wide limits for a
  85         user/group and cluster-wide usage for a user/group), as well as
  86         statistics for each MDS/OST. Now,
  87         <literal>lfs quota</literal> has been updated to provide only
  88         cluster-wide statistics, by default. To obtain the full report of
  89         cluster-wide limits, usage and statistics, use the
  90         <literal>-v</literal> option with
  91         <literal>lfs quota</literal>.</para>
  92       </note>
  93     </section>
  94     <section remap="h5">
  95       <title>Description</title>
  96       <para>The
  97       <literal>lfs</literal> utility is used to create a new file with a
  98       specific striping pattern, determine the default striping pattern, gather
  99       the extended attributes (object numbers and location) for a specific
 100       file, find files with specific attributes, list OST information or set
 101       quota limits. It can be invoked interactively without any arguments or in
 102       a non-interactive mode with one of the supported arguments.</para>
 103     </section>
 104     <section remap="h5">
 105       <title>Options</title>
 106       <para>The various
 107       <literal>lfs</literal> options are listed and described below. For a
 108       complete list of available options, type help at the
 109       <literal>lfs</literal> prompt.</para>
 110       <informaltable frame="all">
 111         <tgroup cols="3">
 112           <colspec colname="c1" colwidth="20*" />
 113           <colspec colname="c2" colwidth="30*" />
 114           <colspec colname="c3" colwidth="50*" />
 115           <thead>
 116             <row>
 117               <entry nameend="c2" namest="c1">
 118                 <para>
 119                   <emphasis role="bold">Option</emphasis>
 120                 </para>
 121               </entry>
 122               <entry>
 123                 <para>
 124                   <emphasis role="bold">Description</emphasis>
 125                 </para>
 126               </entry>
 127             </row>
 128           </thead>
 129           <tbody>
 130             <row>
 131               <entry nameend="c2" namest="c1">
 132                 <para>
 133                   <literal>changelog</literal>
 134                 </para>
 135               </entry>
 136               <entry>
 137                 <para>Shows the metadata changes on an MDT. Start and end
 138                 points are optional. The
 139                 <literal>--follow</literal> option blocks on new changes; this
 140                 option is only valid when run directly on the MDT node.</para>
 141               </entry>
 142             </row>
 143             <row>
 144               <entry nameend="c2" namest="c1">
 145                 <para>
 146                   <literal>changelog_clear</literal>
 147                 </para>
 148               </entry>
 149               <entry>
 150                 <para>Indicates that changelog records previous to
 151                 <literal>
 152                   <replaceable>endrec</replaceable>
 153                 </literal> are no longer of interest to a particular consumer
 154                 <literal>
 155                   <replaceable>id</replaceable>
 156                 </literal>, potentially allowing the MDT to free up disk space.
 157                 An
 158                 <literal>
 159                   <replaceable>endrec</replaceable>
 160                 </literal> of 0 indicates the current last record. Changelog
 161                 consumers must be registered on the MDT node using
 162                 <literal>lctl</literal>.</para>
 163               </entry>
 164             </row>
 165             <row>
 166               <entry nameend="c2" namest="c1">
 167                 <literal>check</literal>
 168               </entry>
 169               <entry>
 170                 <para>Displays the status of MDS or OSTs (as specified in the
 171                 command) or all servers (MDS and OSTs).</para>
 172               </entry>
 173             </row>
 174             <row>
 175               <entry nameend="c2" namest="c1">
 176                 <para>
 177                   <literal>data_version [-nrw]
 178                   <replaceable>filename</replaceable></literal>
 179                 </para>
 180               </entry>
 181               <entry>
 182                 <para>Displays the current version of file data. If
 183                 <literal>-n</literal> is specified, the data version is read
 184                 without taking a lock. As a consequence, the data version could
 185                 be outdated if there are dirty caches on filesystem clients, but
 186                 this option will not force data flushes and has less of an
 187                 impact on the filesystem. If <literal>-r</literal> is specified,
 188                 the data version is read after dirty pages on clients are
 189                 flushed. If <literal>-w</literal> is specified, the data version
 190                 is read after all caching pages on clients are flushed.
 191                 </para>
 192                 <para>
 193                 Even with <literal>-r</literal> or <literal>-w</literal>, race
 194                 conditions are possible and the data version should be checked
 195                 before and after an operation to be confident the data did not
 196                 change during it.
 197                 </para>
 198                 <para>
 199                 The  data  version is the sum of the last committed transaction
 200                 numbers of all data objects of a file. It is used by HSM policy
 201                 engines for verifying that file data has not been changed during
 202                 an archive operation or before a release operation, and by OST
 203                 migration, primarily for verifying that file data has not been
 204                 changed during a data copy, when done in non-blocking mode.
 205                 </para>
 206               </entry>
 207             </row>
 208             <row>
 209               <entry nameend="c2" namest="c1">
 210                 <literal>df [-i] [-h] [--pool|-p
 211                 <replaceable>fsname</replaceable>[.
 212                 <replaceable>pool</replaceable>] [
 213                 <replaceable>path</replaceable>] [--lazy]</literal>
 214               </entry>
 215               <entry>
 216                 <para>Use
 217                 <literal>-i</literal> to report file system disk space usage or
 218                 inode usage of each MDT or OST or, if a pool is specified with
 219                 the
 220                 <literal>-p</literal> option, a subset of OSTs.</para>
 221                 <para>By default, the usage of all mounted Lustre file systems
 222                 is reported. If the
 223                 <literal>path</literal> option is included, only the usage for
 224                 the specified file system is reported. If the
 225                 <literal>-h</literal> option is included, the output is printed
 226                 in human-readable format, using SI base-2 suffixes for
 227                 <emphasis role="bold">M</emphasis>ega-,
 228                 <emphasis role="bold">G</emphasis>iga-,
 229                 <emphasis role="bold">T</emphasis>era-,
 230                 <emphasis role="bold">P</emphasis>eta-, or
 231                 <emphasis role="bold">E</emphasis>xabytes.</para>
 232                 <para>If the
 233                 <literal>--lazy</literal> option is specified, any OST that is
 234                 currently disconnected from the client will be skipped. Using
 235                 the
 236                 <literal>--lazy</literal> option prevents the
 237                 <literal>df</literal> output from being blocked when an OST is
 238                 offline. Only the space on the OSTs that can currently be
 239                 accessed are returned. The
 240                 <literal>llite.*.lazystatfs</literal> tunable can be enabled to
 241                 make this the default behaviour for all
 242                 <literal>statfs()</literal> operations.</para>
 243               </entry>
 244             </row>
 245             <row>
 246               <entry nameend="c2" namest="c1">
 247                 <para>
 248                   <literal>find</literal>
 249                 </para>
 250               </entry>
 251               <entry>
 252                 <para>Searches the directory tree rooted at the given
 253                 directory/filename for files that match the given
 254                 parameters.</para>
 255                 <para>Using
 256                 <literal>!</literal> before an option negates its meaning (files
 257                 NOT matching the parameter). Using
 258                 <literal>+</literal> before a numeric value means files with the
 259                 parameter OR MORE. Using
 260                 <literal>-</literal> before a numeric value means files with the
 261                 parameter OR LESS.</para>
 262               </entry>
 263             </row>
 264             <row>
 265               <entry />
 266               <entry>
 267                 <literal>--atime</literal>
 268               </entry>
 269               <entry>
 270                 <para>File was last accessed N*24 hours ago. (There is no
 271                 guarantee that
 272                 <literal>atime</literal> is kept coherent across the
 273                 cluster.)</para>
 274                 <para>OSTs store a transient
 275                 <literal>atime</literal> that is updated when clients do read
 276                 requests. Permanent
 277                 <literal>atime</literal> is written to the MDS when the file is
 278                 closed. However, on-disk atime is only updated if it is more
 279                 than 60 seconds old (
 280                 <literal>/proc/fs/lustre/mds/*/max_atime_diff</literal>). The
 281                 Lustre software considers the latest
 282                 <literal>atime</literal> from all OSTs. If a
 283                 <literal>setattr</literal> is set by user, then it is updated on
 284                 both the MDS and OST, allowing the
 285                 <literal>atime</literal> to go backward.</para>
 286               </entry>
 287             </row>
 288             <row>
 289               <entry>
 290                 <para>&#160;</para>
 291               </entry>
 292               <entry>
 293                 <para>
 294                   <literal>--ctime</literal>
 295                 </para>
 296               </entry>
 297               <entry>
 298                 <para>File status was last changed N*24 hours ago.</para>
 299               </entry>
 300             </row>
 301             <row>
 302               <entry>
 303                 <para>&#160;</para>
 304               </entry>
 305               <entry>
 306                 <para>
 307                   <literal>--mtime</literal>
 308                 </para>
 309               </entry>
 310               <entry>
 311                 <para>File data was last modified N*24 hours ago.</para>
 312               </entry>
 313             </row>
 314             <row>
 315               <entry>
 316                 <para>&#160;</para>
 317               </entry>
 318               <entry>
 319                 <para>
 320                   <literal>--obd</literal>
 321                 </para>
 322               </entry>
 323               <entry>
 324                 <para>File has an object on a specific OST(s).</para>
 325               </entry>
 326             </row>
 327             <row>
 328               <entry>
 329                 <para>&#160;</para>
 330               </entry>
 331               <entry>
 332                 <para>
 333                   <literal>--size</literal>
 334                 </para>
 335               </entry>
 336               <entry>
 337                 <para>File has a size in bytes, or kilo-, Mega-, Giga-, Tera-,
 338                 Peta- or Exabytes if a suffix is given.</para>
 339               </entry>
 340             </row>
 341             <row>
 342               <entry>
 343                 <para>&#160;</para>
 344               </entry>
 345               <entry>
 346                 <para>
 347                   <literal>--type</literal>
 348                 </para>
 349               </entry>
 350               <entry>
 351                 <para>File has the type - block, character, directory, pipe,
 352                 file, symlink, socket or door (used in Solaris operating
 353                 system).</para>
 354               </entry>
 355             </row>
 356             <row>
 357               <entry>
 358                 <para>&#160;</para>
 359               </entry>
 360               <entry>
 361                 <para>
 362                   <literal>--uid</literal>
 363                 </para>
 364               </entry>
 365               <entry>
 366                 <para>File has a specific numeric user ID.</para>
 367               </entry>
 368             </row>
 369             <row>
 370               <entry>
 371                 <para>&#160;</para>
 372               </entry>
 373               <entry>
 374                 <para>
 375                   <literal>--user</literal>
 376                 </para>
 377               </entry>
 378               <entry>
 379                 <para>File owned by a specific user (numeric user ID
 380                 allowed).</para>
 381               </entry>
 382             </row>
 383             <row>
 384               <entry>
 385                 <para>&#160;</para>
 386               </entry>
 387               <entry>
 388                 <para>
 389                   <literal>--gid</literal>
 390                 </para>
 391               </entry>
 392               <entry>
 393                 <para>File has a specific group ID.</para>
 394               </entry>
 395             </row>
 396             <row>
 397               <entry>
 398                 <para>&#160;</para>
 399               </entry>
 400               <entry>
 401                 <para>
 402                   <literal>--group</literal>
 403                 </para>
 404               </entry>
 405               <entry>
 406                 <para>File belongs to a specific group (numeric group ID
 407                 allowed).</para>
 408               </entry>
 409             </row>
 410             <row>
 411               <entry>
 412                 <para>&#160;</para>
 413               </entry>
 414               <entry>
 415                 <para>-
 416                 <literal>-maxdepth</literal></para>
 417               </entry>
 418               <entry>
 419                 <para>Limits find to descend at most N levels of the directory
 420                 tree.</para>
 421               </entry>
 422             </row>
 423             <row>
 424               <entry>
 425                 <para>&#160;</para>
 426               </entry>
 427               <entry>
 428                 <para>
 429                 <literal>--print</literal>/
 430                 <literal>--print0</literal></para>
 431               </entry>
 432               <entry>
 433                 <para>Prints the full filename, followed by a new line or NULL
 434                 character correspondingly.</para>
 435               </entry>
 436             </row>
 437             <row>
 438               <entry nameend="c2" namest="c1">
 439                 <para>
 440                   <literal>osts [path]</literal>
 441                 </para>
 442               </entry>
 443               <entry>
 444                 <para>Lists all OSTs for the file system. If a path located on
 445                 a mounted Lustre file system is specified, then only OSTs
 446                 belonging to this file system are displayed.</para>
 447               </entry>
 448             </row>
 449             <row>
 450               <entry nameend="c2" namest="c1">
 451                 <para>
 452                   <literal>getname [path...]</literal>
 453                 </para>
 454               </entry>
 455               <entry>
 456                 <para>List each Lustre file system instance associated with
 457                 each Lustre mount point. If no path is specified, all Lustre
 458                 mount points are interrogated. If a list of paths is provided,
 459                 the instance of each path is provided. If the path is not a
 460                 Lustre instance 'No such device' is returned.</para>
 461               </entry>
 462             </row>
 463             <row>
 464               <entry nameend="c2" namest="c1">
 465                 <para>
 466                   <literal>getstripe</literal>
 467                 </para>
 468               </entry>
 469               <entry>
 470                 <para>Lists striping information for a given filename or
 471                 directory. By default, the stripe count, stripe size and offset
 472                 are returned.</para>
 473                 <para>If you only want specific striping information, then the
 474                 options of
 475                 <literal>--stripe-count</literal>,
 476                 <literal>--stripe-size</literal>,
 477                 <literal>--stripe-index</literal>,
 478                 <literal>--layout</literal>, or
 479                 <literal>--pool</literal> plus various combinations of these
 480                 options can be used to retrieve specific information.</para>
 481                 <para>If the
 482                 <literal>--raw</literal> option is specified, the stripe
 483                 information is printed without substituting the file system
 484                 default values for unspecified fields. If the striping EA is
 485                 not set, 0, 0, and -1 will be printed for the stripe count,
 486                 size, and offset respectively.</para>
 487                 <para condition="l24">The
 488                 <literal>--mdt-index</literal> prints the index of the MDT for a given
 489                 directory. See
 490                 <xref linkend="lustremaint.rmremotedir" />.</para>
 491               </entry>
 492             </row>
 493             <row>
 494               <entry>
 495                 <para>&#160;</para>
 496               </entry>
 497               <entry>
 498                 <para>
 499                   <literal>--obd
 500                   <replaceable>ost_name</replaceable></literal>
 501                 </para>
 502               </entry>
 503               <entry>
 504                 <para>Lists files that have an object on a specific OST.</para>
 505               </entry>
 506             </row>
 507             <row>
 508               <entry>
 509                 <para>&#160;</para>
 510               </entry>
 511               <entry>
 512                 <para>
 513                   <literal>--quiet</literal>
 514                 </para>
 515               </entry>
 516               <entry>
 517                 <para>Lists details about the file's object ID
 518                 information.</para>
 519               </entry>
 520             </row>
 521             <row>
 522               <entry>
 523                 <para>&#160;</para>
 524               </entry>
 525               <entry>
 526                 <para>
 527                   <literal>--verbose</literal>
 528                 </para>
 529               </entry>
 530               <entry>
 531                 <para>Prints additional striping information.</para>
 532               </entry>
 533             </row>
 534             <row>
 535               <entry>
 536                 <para>&#160;</para>
 537               </entry>
 538               <entry>
 539                 <para>
 540                   <literal>--stripe-count</literal>
 541                 </para>
 542               </entry>
 543               <entry>
 544                 <para>Lists the stripe count (how many OSTs to use).</para>
 545               </entry>
 546             </row>
 547             <row>
 548               <entry>
 549                 <para>&#160;</para>
 550               </entry>
 551               <entry>
 552                 <para>
 553                   <literal>--index</literal>
 554                 </para>
 555               </entry>
 556               <entry>
 557                 <para>Lists the index for each OST in the file system.</para>
 558               </entry>
 559             </row>
 560             <row>
 561               <entry>
 562                 <para>&#160;</para>
 563               </entry>
 564               <entry>
 565                 <para>
 566                   <literal>--offset</literal>
 567                 </para>
 568               </entry>
 569               <entry>
 570                 <para>Lists the OST index on which file striping starts.</para>
 571               </entry>
 572             </row>
 573             <row>
 574               <entry>
 575                 <para>&#160;</para>
 576               </entry>
 577               <entry>
 578                 <para>
 579                   <literal>--pool</literal>
 580                 </para>
 581               </entry>
 582               <entry>
 583                 <para>Lists the pools to which a file belongs.</para>
 584               </entry>
 585             </row>
 586             <row>
 587               <entry>
 588                 <para>&#160;</para>
 589               </entry>
 590               <entry>
 591                 <para>
 592                   <literal>--size</literal>
 593                 </para>
 594               </entry>
 595               <entry>
 596                 <para>Lists the stripe size (how much data to write to one OST
 597                 before moving to the next OST).</para>
 598               </entry>
 599             </row>
 600             <row>
 601               <entry>
 602                 <para>&#160;</para>
 603               </entry>
 604               <entry>
 605                 <para>
 606                   <literal>--directory</literal>
 607                 </para>
 608               </entry>
 609               <entry>
 610                 <para>Lists entries about a specified directory instead of its
 611                 contents (in the same manner as
 612                 <literal>ls -d</literal>).</para>
 613               </entry>
 614             </row>
 615             <row>
 616               <entry>
 617                 <para>&#160;</para>
 618               </entry>
 619               <entry>
 620                 <para>
 621                   <literal>--recursive</literal>
 622                 </para>
 623               </entry>
 624               <entry>
 625                 <para>Recurses into all sub-directories.</para>
 626               </entry>
 627             </row>
 628             <row>
 629               <entry nameend="c2" namest="c1">
 630                 <para>
 631                   <literal>setstripe</literal>
 632                 </para>
 633               </entry>
 634               <entry>
 635                 <para>Create new files with a specific file layout (stripe
 636                 pattern) configuration.
 637                 <footnote>
 638                   <para>The file cannot exist prior to using
 639                   <literal>setstripe</literal>. A directory must exist prior to
 640                   using
 641                   <literal>setstripe</literal>.</para>
 642                 </footnote></para>
 643               </entry>
 644             </row>
 645             <row>
 646               <entry>
 647                 <para>&#160;</para>
 648               </entry>
 649               <entry>
 650                 <para>
 651                   <literal>--stripe-count stripe_cnt</literal>
 652                 </para>
 653               </entry>
 654               <entry>
 655                 <para>Number of OSTs over which to stripe a file. A
 656                 <literal>stripe_cnt</literal> of 0 uses the file system-wide
 657                 default stripe count (default is 1). A
 658                 <literal>stripe_cnt</literal> of -1 stripes over all available
 659                 OSTs.</para>
 660               </entry>
 661             </row>
 662             <row>
 663               <entry>
 664                 <para>&#160;</para>
 665               </entry>
 666               <entry>
 667                 <para>
 668                   <literal>--overstripe-count stripe_cnt</literal>
 669                 </para>
 670               </entry>
 671               <entry>
 672                 <para>The same as --stripe-count, but allows overstriping,
 673                 which will place more than one stripe per OST if
 674                 <literal>stripe_cnt</literal> is greater than the number of
 675                 OSTs. Overstriping is useful for matching the number of stripes
 676                 to the number of processes, or with very fast OSTs, where one
 677                 stripe per OST is not enough to get full performance.</para>
 678               </entry>
 679             </row>
 680             <row>
 681               <entry>
 682                 <para>&#160;</para>
 683               </entry>
 684               <entry>
 685                 <para>
 686                 <literal>--size stripe_size</literal>
 687                 <footnote>
 688                   <para>The default stripe-size is 0. The default start-ost is
 689                   -1. Do NOT confuse them! If you set start-ost to 0, all new
 690                   file creations occur on OST 0 (seldom a good idea).</para>
 691                 </footnote>&#160;</para>
 692               </entry>
 693               <entry>
 694                 <para>Number of bytes to store on an OST before moving to the
 695                 next OST. A stripe_size of 0 uses the file system's default
 696                 stripe size, (default is 1 MB). Can be specified with
 697                 <emphasis role="bold">k</emphasis>(KB),
 698                 <emphasis role="bold">m</emphasis>(MB), or
 699                 <emphasis role="bold">g</emphasis>(GB), respectively.</para>
 700               </entry>
 701             </row>
 702             <row>
 703               <entry>
 704                 <para>&#160;</para>
 705               </entry>
 706               <entry>
 707                 <para>
 708                   <literal>--stripe-index start_ost_index</literal>
 709                 </para>
 710               </entry>
 711               <entry>
 712                 <para>The OST index (base 10, starting at 0) on which to start
 713                 striping for this file. A start_ost_index value of -1 allows
 714                 the MDS to choose the starting index. This is the default
 715                 value, and it means that the MDS selects the starting OST as it
 716                 wants. We strongly recommend selecting this default, as it
 717                 allows space and load balancing to be done by the MDS as
 718                 needed. The
 719                 <literal>start_ost_index</literal> value has no relevance on
 720                 whether the MDS will use round-robin or QoS weighted allocation
 721                 for the remaining stripes in the file.</para>
 722               </entry>
 723             </row>
 724             <row>
 725               <entry>
 726                 <para>&#160;</para>
 727               </entry>
 728               <entry>
 729                 <para>
 730                   <literal>--ost-index ost_indices</literal>
 731                 </para>
 732               </entry>
 733               <entry>
 734                                 <para>This option is used to specify the exact stripe
 735                 layout on the the file system. <literal>ost_indices</literal>
 736                 is a list of OSTs referenced by their indices and index ranges
 737                 separated by commas, e.g. <literal>1,2-4,7</literal>.</para>
 738               </entry>
 739             </row>
 740             <row>
 741               <entry>
 742                 <para>&#160;</para>
 743               </entry>
 744               <entry>
 745                 <para>
 746                   <literal>--pool
 747                   <replaceable>pool</replaceable></literal>
 748                 </para>
 749               </entry>
 750               <entry>
 751                 <para>Name of the pre-defined pool of OSTs (see
 752                 <xref linkend="dbdoclet.50438219_38274" />) that will be used
 753                 for striping. The
 754                 <literal>stripe_cnt</literal>,
 755                 <literal>stripe_size</literal> and
 756                 <literal>start_ost</literal> values are used as well. The
 757                 start-ost value must be part of the pool or an error is
 758                 returned.</para>
 759               </entry>
 760             </row>
 761             <row>
 762               <entry nameend="c2" namest="c1">
 763                 <para>
 764                   <literal>setstripe -d</literal>
 765                 </para>
 766               </entry>
 767               <entry>
 768                 <para>Deletes default striping on the specified
 769                 directory.</para>
 770               </entry>
 771             </row>
 772             <row>
 773               <entry nameend="c2" namest="c1">
 774                 <para>
 775                   <literal>pool_list {filesystem}[.poolname]|{pathname}</literal>
 776                 </para>
 777               </entry>
 778               <entry>
 779                 <para>Lists pools in the file system or pathname, or OSTs in
 780                 the file system's pool.</para>
 781               </entry>
 782             </row>
 783             <row>
 784               <entry nameend="c2" namest="c1">
 785                 <para>
 786                   <literal>quota [-q] [-v] [-o
 787                   <replaceable>obd_uuid</replaceable>|-i
 788                   <replaceable>mdt_idx</replaceable>|-I
 789                   <replaceable>ost_idx</replaceable>] [-u|-g|-p
 790                   <replaceable>uname|uid|gname|gid|projid]</replaceable>
 791                   <replaceable>/mount_point</replaceable></literal>
 792                 </para>
 793                 <para>&#160;</para>
 794               </entry>
 795               <entry>
 796                 <para>Displays disk usage and limits, either for the full file
 797                 system or for objects on a specific OBD. A user or group name
 798                 or an usr, group and project ID can be specified. If all user,
 799                 group project ID are omitted, quotas for the current UID/GID
 800                 are shown. The <literal>-q</literal> option disables printing
 801                 of additional descriptions (including column titles). It fills
 802                 in blank spaces in the
 803                 <literal>grace</literal> column with zeros (when there is no
 804                 grace period set), to ensure that the number of columns is
 805                 consistent. The
 806                 <literal>-v</literal> option provides more verbose (per-OBD
 807                 statistics) output.</para>
 808               </entry>
 809             </row>
 810             <row>
 811               <entry nameend="c2" namest="c1">
 812                 <para>
 813                   <literal>quota -t
 814                   <replaceable>-u|-g|-p</replaceable>
 815                   <replaceable>/mount_point</replaceable></literal>
 816                 </para>
 817               </entry>
 818               <entry>
 819                 <para>Displays block and inode grace times for user (
 820                 <literal>-u</literal>) or group (
 821                 <literal>-g</literal>) or project (
 822                 <literal>-p</literal>) quotas.</para>
 823               </entry>
 824             </row>
 825             <row>
 826               <entry nameend="c2" namest="c1">
 827                 <para>
 828                   <literal>quotachown</literal>
 829                 </para>
 830               </entry>
 831               <entry>
 832                 <para>Changes the file's owner and group on OSTs of the
 833                 specified file system.</para>
 834               </entry>
 835             </row>
 836             <row>
 837               <entry nameend="c2" namest="c1">
 838                 <para>
 839                   <literal>quotacheck [-ugf]
 840                   <replaceable>/mount_point</replaceable></literal>
 841                 </para>
 842               </entry>
 843               <entry>
 844                 <para>Scans the specified file system for disk usage, and
 845                 creates or updates quota files. Options specify quota for users
 846                 (
 847                 <literal>-u</literal>), groups (
 848                 <literal>-g</literal>), and force (
 849                 <literal>-f</literal>).</para>
 850               </entry>
 851             </row>
 852             <row>
 853               <entry nameend="c2" namest="c1">
 854                 <para>
 855                   <literal>quotaon [-ugf]
 856                   <replaceable>/mount_point</replaceable></literal>
 857                 </para>
 858               </entry>
 859               <entry>
 860                 <para>Turns on file system quotas. Options specify quota for
 861                 users (
 862                 <literal>-u</literal>), groups (
 863                 <literal>-g</literal>), and force (
 864                 <literal>-f</literal>).</para>
 865               </entry>
 866             </row>
 867             <row>
 868               <entry nameend="c2" namest="c1">
 869                 <para>
 870                   <literal>quotaoff [-ugf]
 871                   <replaceable>/mount_point</replaceable></literal>
 872                 </para>
 873               </entry>
 874               <entry>
 875                 <para>Turns off file system quotas. Options specify quota for
 876                 users (
 877                 <literal>-u</literal>), groups (
 878                 <literal>-g</literal>), and force (
 879                 <literal>-f</literal>).</para>
 880               </entry>
 881             </row>
 882             <row>
 883               <entry nameend="c2" namest="c1">
 884                 <para>
 885                   <literal>quotainv [-ug] [-f]
 886                   <replaceable>/mount_point</replaceable></literal>
 887                 </para>
 888               </entry>
 889               <entry>
 890                 <para>Clears quota files (administrative quota files if used
 891                 without
 892                 <literal>-f</literal>, operational quota files otherwise), all
 893                 of their quota entries for users (
 894                 <literal>-u</literal>) or groups (
 895                 <literal>-g</literal>). After running
 896                 <literal>quotainv</literal>, you must run
 897                 <literal>quotacheck</literal> before using quotas.</para>
 898                 <caution>
 899                   <para>Use extreme caution when using this command; its
 900                   results cannot be undone.</para>
 901                 </caution>
 902               </entry>
 903             </row>
 904             <row>
 905               <entry nameend="c2" namest="c1">
 906                 <para>
 907                   <literal>setquota {-u|-g|-p
 908                   <replaceable>uname|uid|gname|gid|projid}</replaceable>
 909                   [--block-softlimit
 910                   <replaceable>block_softlimit</replaceable>]
 911                   [--block-hardlimit
 912                   <replaceable>block_hardlimit</replaceable>]
 913                   [--inode-softlimit
 914                   <replaceable>inode_softlimit</replaceable>]
 915                   [--inode-hardlimit
 916                   <replaceable>inode_hardlimit</replaceable>]
 917                   <replaceable>/mount_point</replaceable></literal>
 918                 </para>
 919               </entry>
 920               <entry>
 921                 <para>Sets file system quotas for users, groups or one project.
 922                 Limits can be specified with
 923                 <literal>--{block|inode}-{softlimit|hardlimit}</literal> or
 924                 their short equivalents
 925                 <literal>-b</literal>,
 926                 <literal>-B</literal>,
 927                 <literal>-i</literal>,
 928                 <literal>-I</literal>. Users can set 1, 2, 3 or 4 limits.
 929                 <footnote>
 930                   <para>The old
 931                   <literal>setquota</literal> interface is supported, but it may
 932                   be removed in a future Lustre software release.</para>
 933                 </footnote>Also, limits can be specified with special suffixes,
 934                 -b, -k, -m, -g, -t, and -p to indicate units of 1, 2^10, 2^20,
 935                 2^30, 2^40 and 2^50, respectively. By default, the block limits
 936                 unit is 1 kilobyte (1,024), and block limits are always
 937                 kilobyte-grained (even if specified in bytes). See
 938                 <xref linkend="dbdoclet.50438206_11903" />.</para>
 939               </entry>
 940             </row>
 941             <row>
 942               <entry nameend="c2" namest="c1">
 943                 <para>
 944                   <literal>setquota -t -u|-g|-p [--block-grace
 945                   <replaceable>block_grace</replaceable>] [--inode-grace
 946                   <replaceable>inode_grace</replaceable>]
 947                   <replaceable>/mount_point</replaceable></literal>
 948                 </para>
 949               </entry>
 950               <entry>
 951                 <para>Sets the file system quota grace times for users or
 952                 groups. Grace time is specified in '
 953                 <literal>XXwXXdXXhXXmXXs</literal>' format or as an integer
 954                 seconds value. See
 955                 <xref linkend="dbdoclet.50438206_11903" />.</para>
 956               </entry>
 957             </row>
 958             <row>
 959               <entry nameend="c2" namest="c1">
 960                 <para>
 961                   <literal>help</literal>
 962                 </para>
 963               </entry>
 964               <entry>
 965                 <para>Provides brief help on various
 966                 <literal>lfs</literal> arguments.</para>
 967               </entry>
 968             </row>
 969             <row>
 970               <entry nameend="c2" namest="c1">
 971                 <para>
 972                   <literal>exit/quit</literal>
 973                 </para>
 974               </entry>
 975               <entry>
 976                 <para>Quits the interactive
 977                 <literal>lfs</literal> session.</para>
 978               </entry>
 979             </row>
 980           </tbody>
 981         </tgroup>
 982       </informaltable>
 983     </section>
 984     <section remap="h5" xml:id="dbdoclet.50438206_11903">
 985       <title>Examples</title>
 986       <para>Creates a file striped on two OSTs with 128 KB on each
 987       stripe.</para>
 988       <screen>
 989 $ lfs setstripe -s 128k -c 2 /mnt/lustre/file1
 990 </screen>
 991       <para>Deletes a default stripe pattern on a given directory. New files
 992       use the default striping pattern.</para>
 993       <screen>
 994 $ lfs setstripe -d /mnt/lustre/dir
 995 </screen>
 996       <para>Lists the detailed object allocation of a given file.</para>
 997       <screen>
 998 $ lfs getstripe -v /mnt/lustre/file1
 999 </screen>
1000       <para>List all the mounted Lustre file systems and corresponding Lustre
1001       instances.</para>
1002       <screen>
1003 $ lfs getname
1004 </screen>
1005       <para>Efficiently lists all files in a given directory and its
1006       subdirectories.</para>
1007       <screen>
1008 $ lfs find /mnt/lustre
1009 </screen>
1010       <para>Recursively lists all regular files in a given directory more than
1011       30 days old.</para>
1012       <screen>
1013 $ lfs find /mnt/lustre -mtime +30 -type f -print
1014 </screen>
1015       <para>Recursively lists all files in a given directory that have objects
1016       on OST2-UUID. The lfs check servers command checks the status of all
1017       servers (MDT and OSTs).</para>
1018       <screen>
1019 $ lfs find --obd OST2-UUID /mnt/lustre/
1020 </screen>
1021       <para>Lists all OSTs in the file system.</para>
1022       <screen>
1023 $ lfs osts
1024 </screen>
1025       <para>Lists space usage per OST and MDT in human-readable format.</para>
1026       <screen>
1027 $ lfs df -h
1028 </screen>
1029       <para>Lists inode usage per OST and MDT.</para>
1030       <screen>
1031 $ lfs df -i
1032 </screen>
1033       <para>List space or inode usage for a specific OST pool.</para>
1034       <screen>
1035 $ lfs df --pool
1036 <replaceable>filesystem</replaceable>[.
1037 <replaceable>pool</replaceable>] |
1038 <replaceable>pathname</replaceable>
1039 </screen>
1040       <para>List quotas of user 'bob'.</para>
1041       <screen>
1042 $ lfs quota -u bob /mnt/lustre
1043 </screen>
1044       <para>List quotas of project ID '1'.</para>
1045       <screen>
1046 $ lfs quota -p 1 /mnt/lustre
1047 </screen>
1048       <para>Show grace times for user quotas on
1049       <literal>/mnt/lustre</literal>.</para>
1050       <screen>
1051 $ lfs quota -t -u /mnt/lustre
1052 </screen>
1053       <para>Changes file owner and group.</para>
1054       <screen>
1055 $ lfs quotachown -i /mnt/lustre
1056 </screen>
1057       <para>Checks quotas for user and group. Turns on quotas after making the
1058       check.</para>
1059       <screen>
1060 $ lfs quotacheck -ug /mnt/lustre
1061 </screen>
1062       <para>Turns on quotas of user and group.</para>
1063       <screen>
1064 $ lfs quotaon -ug /mnt/lustre
1065 </screen>
1066       <para>Turns off quotas of user and group.</para>
1067       <screen>
1068 $ lfs quotaoff -ug /mnt/lustre
1069 </screen>
1070       <para>Sets quotas of user 'bob', with a 1 GB block quota hardlimit and a
1071       2 GB block quota softlimit.</para>
1072       <screen>
1073 $ lfs setquota -u bob --block-softlimit 2000000 --block-hardlimit 1000000
1074 /mnt/lustre
1075 </screen>
1076       <para>Sets grace times for user quotas: 1000 seconds for block quotas, 1
1077       week and 4 days for inode quotas.</para>
1078       <screen>
1079 $ lfs setquota -t -u --block-grace 1000 --inode-grace 1w4d /mnt/lustre
1080 </screen>
1081       <para>Checks the status of all servers (MDT, OST)</para>
1082       <screen>
1083 $ lfs check servers
1084 </screen>
1085       <para>Creates a file striped on two OSTs from the pool
1086       <literal>my_pool</literal></para>
1087       <screen>
1088 $ lfs setstripe --pool my_pool -c 2 /mnt/lustre/file
1089 </screen>
1090       <para>Lists the pools defined for the mounted Lustre file system
1091       <literal>/mnt/lustre</literal></para>
1092       <screen>
1093 $ lfs pool_list /mnt/lustre/
1094 </screen>
1095       <para>Lists the OSTs which are members of the pool
1096       <literal>my_pool</literal> in file system
1097       <literal>my_fs</literal></para>
1098       <screen>
1099 $ lfs pool_list my_fs.my_pool
1100 </screen>
1101       <para>Finds all directories/files associated with
1102       <literal>poolA</literal>.</para>
1103       <screen>
1104 $ lfs find /mnt/lustre --pool poolA
1105 </screen>
1106       <para>Finds all directories/files not associated with a pool.</para>
1107       <screen>
1108 $ lfs find /mnt//lustre --pool ""
1109 </screen>
1110       <para>Finds all directories/files associated with pool.</para>
1111       <screen>
1112 $ lfs find /mnt/lustre ! --pool ""
1113 </screen>
1114       <para>Associates a directory with the pool
1115       <literal>my_pool</literal>, so all new files and directories are created
1116       in the pool.</para>
1117       <screen>
1118 $ lfs setstripe --pool my_pool /mnt/lustre/dir
1119 </screen>
1120     </section>
1121     <section remap="h5">
1122       <title>See Also</title>
1123       <para>
1124         <xref linkend="dbdoclet.50438219_38274" />
1125       </para>
1126     </section>
1127   </section>
1128   <section xml:id="dbdoclet.lfs_migrate">
1129     <title>
1130       <indexterm>
1131         <primary>lfs_migrate</primary>
1132       </indexterm>
1133       <literal>lfs_migrate</literal>
1134     </title>
1135     <para>The
1136     <literal>lfs_migrate</literal> utility is a simple to migrate file
1137     <emphasis>data</emphasis> between OSTs.</para>
1138     <section remap="h5">
1139       <title>Synopsis</title>
1140       <screen>
1141 lfs_migrate [<replaceable>lfs_setstripe_options</replaceable>]
1142         [-h] [-n] [-q] [-R] [-s] [-y] [-0] [file|directory ...]
1143 </screen>
1144     </section>
1145     <section remap="h5">
1146       <title>Description</title>
1147       <para>The
1148       <literal>lfs_migrate</literal> utility is a tool to assist migration
1149       of file data between Lustre OSTs. The utility copies each specified
1150       file to a temporary file using supplied <literal>lfs setstripe</literal>
1151       options, if any, optionally verifies the file contents have not changed,
1152       and then swaps the layout (OST objects) from the temporary file and the
1153       original file (for Lustre 2.5 and later), or renames the temporary file
1154       to the original filename. This allows the user/administrator to balance
1155       space usage between OSTs, or move files off OSTs that are starting to show
1156       hardware problems (though are still functional) or will be removed.</para>
1157       <warning>
1158         <para>For versions of Lustre before 2.5,
1159         <literal>lfs_migrate</literal> was not integrated with the MDS at all.
1160         That made it UNSAFE for use on files that were being modified by other
1161         applications, since the file was migrated through a copy and rename of
1162         the file.  With Lustre 2.5 and later, the new file layout is swapped
1163         with the existing file layout, which ensures that the user-visible
1164         inode number is kept, and open file handles and locks on the file are
1165         kept.</para>
1166       </warning>
1167       <para>Files to be migrated can be specified as command-line arguments. If
1168       a directory is specified on the command-line then all files within the
1169       directory are migrated. If no files are specified on the command-line,
1170       then a list of files is read from the standard input, making
1171       <literal>lfs_migrate</literal> suitable for use with
1172       <literal>lfs find</literal> to locate files on specific OSTs and/or
1173       matching other file attributes, and other tools that generate a list
1174       of files on standard output.</para>
1175       <para>Unless otherwise specified through command-line options, the
1176       file allocation policies on the MDS dictate where the new files
1177       are placed, taking into account whether specific OSTs have been
1178       disabled on the MDS via <literal>lctl</literal> (preventing new
1179       files from being allocated there), whether some OSTs are overly full
1180       (reducing the number of files placed on those OSTs), or if there is
1181       a specific default file striping for the parent directory (potentially
1182       changing the stripe count, stripe size, OST pool, or OST index of a
1183       new file).</para>
1184       <note>
1185         <para>The
1186         <literal>lfs_migrate</literal> utility can also be used in some cases to
1187         reduce file
1188         <indexterm>
1189           <primary>fragmentation</primary>
1190         </indexterm>fragmentation. File fragmentation will typically reduce
1191         Lustre file system performance. File fragmentation may be observed on
1192         an aged file system and will commonly occur if the file was written by
1193         many threads. Provided there is sufficient free space (or if it was
1194         written when the file system was nearly full) that is less fragmented
1195         than the file being copied, re-writing a file with
1196         <literal>lfs_migrate</literal> will result in a migrated file with
1197         reduced fragmentation. The tool
1198         <literal>filefrag</literal> can be used to report file fragmentation.
1199         See
1200         <xref linkend="dbdoclet.50438206_75125" /></para>
1201       </note>
1202       <note>
1203         <para>As long as a file has extent lengths of tens of megabytes (
1204         <replaceable>read_bandwidth * seek_time</replaceable>) or more, the
1205         read performance for the file will not be significantly impacted by
1206         fragmentation, since the read pipeline can be filled by large reads
1207         from disk even with an occasional disk seek.</para>
1208       </note>
1209     </section>
1210     <section remap="h5">
1211       <title>Options</title>
1212       <para>Options supporting
1213       <literal>lfs_migrate</literal> are described below.</para>
1214       <informaltable frame="all">
1215         <tgroup cols="2">
1216           <colspec colname="c1" colwidth="50*" />
1217           <colspec colname="c2" colwidth="50*" />
1218           <thead>
1219             <row>
1220               <entry>
1221                 <para>
1222                   <emphasis role="bold">Option</emphasis>
1223                 </para>
1224               </entry>
1225               <entry>
1226                 <para>
1227                   <emphasis role="bold">Description</emphasis>
1228                 </para>
1229               </entry>
1230             </row>
1231           </thead>
1232           <tbody>
1233             <row>
1234               <entry>
1235                 <para>
1236                   <literal>-c
1237                   <replaceable>stripecount</replaceable></literal>
1238                 </para>
1239               </entry>
1240               <entry>
1241                 <para>Restripe file using the specified stripe count. This
1242                 option may not be specified at the same time as the
1243                 <literal>-R</literal> option.</para>
1244               </entry>
1245             </row>
1246             <row>
1247               <entry>
1248                 <para>
1249                   <literal>-h</literal>
1250                 </para>
1251               </entry>
1252               <entry>
1253                 <para>Display help information.</para>
1254               </entry>
1255             </row>
1256             <row>
1257               <entry>
1258                 <literal>-l</literal>
1259               </entry>
1260               <entry>
1261                 <para>Migrate files with hard links (skips, by default). Files
1262                 with multiple hard links are split into multiple separate files
1263                 by
1264                 <literal>lfs_migrate</literal>, so they are skipped, by
1265                 default, to avoid breaking the hard links.</para>
1266               </entry>
1267             </row>
1268             <row>
1269               <entry>
1270                 <literal>-n</literal>
1271               </entry>
1272               <entry>
1273                 <para>Only print the names of files to be migrated.</para>
1274               </entry>
1275             </row>
1276             <row>
1277               <entry>
1278                 <para>
1279                   <literal>-q</literal>
1280                 </para>
1281               </entry>
1282               <entry>
1283                 <para>Run quietly (does not print filenames or status).</para>
1284               </entry>
1285             </row>
1286             <row>
1287               <entry>
1288                 <literal>-R</literal>
1289               </entry>
1290               <entry>Restripe file using default directory striping instead of
1291               keeping striping. This option may not be specified at the same
1292               time as the
1293               <literal>-c</literal> option.</entry>
1294             </row>
1295             <row>
1296               <entry>
1297                 <literal>-s</literal>
1298               </entry>
1299               <entry>Skip file data comparison after migrate. Default is to
1300               compare migrated file against original to verify
1301               correctness.</entry>
1302             </row>
1303             <row>
1304               <entry>
1305                 <para>
1306                   <literal>-y</literal>
1307                 </para>
1308               </entry>
1309               <entry>
1310                 <para>Answer '
1311                 <literal>y</literal>' to usage warning without prompting (for
1312                 scripts, use with caution).</para>
1313               </entry>
1314             </row>
1315             <row>
1316               <entry>
1317                 <literal>-0</literal>
1318               </entry>
1319               <entry>Expect NUL-terminated filenames on standard input, as
1320               generated by <literal>lfs find -print0</literal> or
1321               <literal>find -print0</literal>.  This allows filenames with
1322               embedded newlines to be handled correctly.
1323               </entry>
1324             </row>
1325           </tbody>
1326         </tgroup>
1327       </informaltable>
1328     </section>
1329     <section remap="h5">
1330       <title>Examples</title>
1331       <para>Rebalance all files in
1332       <literal>/mnt/lustre/dir</literal>:</para>
1333       <screen>
1334 $ lfs_migrate /mnt/lustre/dir
1335 </screen>
1336       <para>Migrate files in /test filesystem on OST0004 larger than 4 GB in
1337       size and older than a day old:</para>
1338       <screen>
1339 $ lfs find /test -obd test-OST0004 -size +4G -mtime +1 | lfs_migrate -y
1340 </screen>
1341     </section>
1342     <section remap="h5">
1343       <title>See Also</title>
1344       <para>
1345         <xref linkend="dbdoclet.50438206_94597" />
1346       </para>
1347     </section>
1348   </section>
1349   <section xml:id="dbdoclet.50438206_75125">
1350     <title>
1351       <indexterm>
1352         <primary>filefrag</primary>
1353       </indexterm>
1354       <literal>filefrag</literal>
1355     </title>
1356     <para>The
1357     <literal>e2fsprogs</literal> package contains the
1358     <literal>filefrag</literal> tool which reports the extent of file
1359     fragmentation.</para>
1360     <section remap="h5">
1361       <title>Synopsis</title>
1362       <screen>
1363 filefrag [ -belsv ] [ files...  ]
1364 </screen>
1365     </section>
1366     <section remap="h5">
1367       <title>Description</title>
1368       <para>The
1369       <literal>filefrag</literal> utility reports the extent of fragmentation in
1370       a given file. The
1371       <literal>filefrag</literal> utility obtains the extent information from
1372       Lustre files using the
1373       <literal>FIEMAP ioctl</literal>, which is efficient and fast, even for
1374       very large files.</para>
1375       <para>In default mode
1376       <footnote>
1377         <para>The default mode is faster than the verbose/extent mode since it
1378         only counts the number of extents.</para>
1379       </footnote>,
1380       <literal>filefrag</literal> prints the number of physically discontiguous
1381       extents in the file. In extent or verbose mode, each extent is printed
1382       with details such as the blocks allocated on each OST. For a Lustre file
1383       system, the extents are printed in device offset order (i.e. all of the
1384       extents for one OST first, then the next OST, etc.), not file logical
1385       offset order. If the file logical offset order was used, the Lustre
1386       striping would make the output very verbose and difficult to see if there
1387       was file fragmentation or not.</para>
1388       <note>
1389         <para>Note that as long as a file has extent lengths of tens of
1390         megabytes or more (i.e.
1391         <replaceable>read_bandwidth * seek_time &gt;
1392         extent_length</replaceable>), the read performance for the file will
1393         not be significantly impacted by fragmentation, since file readahead
1394         can fully utilize the disk disk bandwidth even with occasional
1395         seeks.</para>
1396       </note>
1397       <para>In default mode
1398       <footnote>
1399         <para>The default mode is faster than the verbose/extent mode.</para>
1400       </footnote>,
1401       <literal>filefrag</literal> returns the number of physically discontiguous
1402       extents in the file. In extent or verbose mode, each extent is printed
1403       with details. For a Lustre file system, the extents are printed in device
1404       offset order, not logical offset order.</para>
1405     </section>
1406     <section remap="h5">
1407       <title>Options</title>
1408       <para>The options and descriptions for the
1409       <literal>filefrag</literal> utility are listed below.</para>
1410       <informaltable frame="all">
1411         <tgroup cols="2">
1412           <colspec colname="c1" colwidth="50*" />
1413           <colspec colname="c2" colwidth="50*" />
1414           <thead>
1415             <row>
1416               <entry>
1417                 <para>
1418                   <emphasis role="bold">Option</emphasis>
1419                 </para>
1420               </entry>
1421               <entry>
1422                 <para>
1423                   <emphasis role="bold">Description</emphasis>
1424                 </para>
1425               </entry>
1426             </row>
1427           </thead>
1428           <tbody>
1429             <row>
1430               <entry>
1431                 <para>
1432                   <literal>-b</literal>
1433                 </para>
1434               </entry>
1435               <entry>
1436                 <para>Uses the 1024-byte blocksize for the output. By default,
1437                 this blocksize is used by the Lustre file system, since OSTs
1438                 may use different block sizes.</para>
1439               </entry>
1440             </row>
1441             <row>
1442               <entry>
1443                 <para>
1444                   <literal>-e</literal>
1445                 </para>
1446               </entry>
1447               <entry>
1448                 <para>Uses the extent mode when printing the output. This is
1449                 the default for Lustre files in verbose mode.</para>
1450               </entry>
1451             </row>
1452             <row>
1453               <entry>
1454                 <para>
1455                   <literal>-l</literal>
1456                 </para>
1457               </entry>
1458               <entry>
1459                 <para>Displays extents in LUN offset order. This is the only
1460                 available mode for Lustre.</para>
1461               </entry>
1462             </row>
1463             <row>
1464               <entry>
1465                 <para>
1466                   <literal>-s</literal>
1467                 </para>
1468               </entry>
1469               <entry>
1470                 <para>Synchronizes any unwritten file data to disk before
1471                 requesting the mapping.</para>
1472               </entry>
1473             </row>
1474             <row>
1475               <entry>
1476                 <para>
1477                   <literal>-v</literal>
1478                 </para>
1479               </entry>
1480               <entry>
1481                 <para>Prints the file's layout in verbose mode when checking
1482                 file fragmentation, including the logical to physical mapping
1483                 for each extent in the file and the OST index.</para>
1484               </entry>
1485             </row>
1486           </tbody>
1487         </tgroup>
1488       </informaltable>
1489     </section>
1490     <section remap="h5">
1491       <title>Examples</title>
1492       <para>Lists default output.</para>
1493       <screen>
1494 $ filefrag /mnt/lustre/foo
1495 /mnt/lustre/foo: 13 extents found
1496 </screen>
1497       <para>Lists verbose output in extent format.</para>
1498       <screen>
1499 $ filefrag -v /mnt/lustre/foo
1500 Filesystem type is: bd00bd0
1501 File size of /mnt/lustre/foo is 1468297786 (1433888 blocks of 1024 bytes)
1502  ext:     device_logical:        physical_offset: length:  dev: flags:
1503    0:        0..  122879: 2804679680..2804802559: 122880: 0002: network
1504    1:   122880..  245759: 2804817920..2804940799: 122880: 0002: network
1505    2:   245760..  278527: 2804948992..2804981759:  32768: 0002: network
1506    3:   278528..  360447: 2804982784..2805064703:  81920: 0002: network
1507    4:   360448..  483327: 2805080064..2805202943: 122880: 0002: network
1508    5:   483328..  606207: 2805211136..2805334015: 122880: 0002: network
1509    6:   606208..  729087: 2805342208..2805465087: 122880: 0002: network
1510    7:   729088..  851967: 2805473280..2805596159: 122880: 0002: network
1511    8:   851968..  974847: 2805604352..2805727231: 122880: 0002: network
1512    9:   974848.. 1097727: 2805735424..2805858303: 122880: 0002: network
1513   10:  1097728.. 1220607: 2805866496..2805989375: 122880: 0002: network
1514   11:  1220608.. 1343487: 2805997568..2806120447: 122880: 0002: network
1515   12:  1343488.. 1433599: 2806128640..2806218751:  90112: 0002: network
1516 /mnt/lustre/foo: 13 extents found
1517 </screen>
1518     </section>
1519   </section>
1520   <section xml:id="dbdoclet.50438206_86244">
1521     <title>
1522       <indexterm>
1523         <primary>mount</primary>
1524       </indexterm>
1525       <literal>mount</literal>
1526     </title>
1527     <para>The standard
1528     <literal>mount(8)</literal> Linux command is used to mount a Lustre file
1529     system. When mounting a Lustre file system, mount(8) executes the
1530     <literal>/sbin/mount.lustre</literal> command to complete the mount. The
1531     mount command supports these options specific to a Lustre file
1532     system:</para>
1533     <informaltable frame="all">
1534       <tgroup cols="2">
1535         <colspec colname="c1" colwidth="50*" />
1536         <colspec colname="c2" colwidth="50*" />
1537         <thead>
1538           <row>
1539             <entry>
1540               <para>
1541                 <emphasis role="bold">Server options</emphasis>
1542               </para>
1543             </entry>
1544             <entry>
1545               <para>
1546                 <emphasis role="bold">Description</emphasis>
1547               </para>
1548             </entry>
1549           </row>
1550         </thead>
1551         <tbody>
1552           <row>
1553             <entry>
1554               <para>
1555                 <literal>abort_recov</literal>
1556               </para>
1557             </entry>
1558             <entry>
1559               <para>Aborts recovery when starting a target</para>
1560             </entry>
1561           </row>
1562           <row>
1563             <entry>
1564               <para>
1565                 <literal>nosvc</literal>
1566               </para>
1567             </entry>
1568             <entry>
1569               <para>Starts only MGS/MGC servers</para>
1570             </entry>
1571           </row>
1572           <row>
1573             <entry>
1574               <para>
1575                 <literal>nomgs</literal>
1576               </para>
1577             </entry>
1578             <entry>
1579               <para>Start a MDT with a co-located MGS without starting the
1580               MGS</para>
1581             </entry>
1582           </row>
1583           <row>
1584             <entry>
1585               <para>
1586                 <literal>exclude</literal>
1587               </para>
1588             </entry>
1589             <entry>
1590               <para>Starts with a dead OST</para>
1591             </entry>
1592           </row>
1593           <row>
1594             <entry>
1595               <para>
1596                 <literal>md_stripe_cache_size</literal>
1597               </para>
1598             </entry>
1599             <entry>
1600               <para>Sets the stripe cache size for server side disk with a
1601               striped raid configuration</para>
1602             </entry>
1603           </row>
1604         </tbody>
1605       </tgroup>
1606     </informaltable>
1607     <informaltable frame="all">
1608       <tgroup cols="2">
1609         <colspec colname="c1" colwidth="50*" />
1610         <colspec colname="c2" colwidth="50*" />
1611         <thead>
1612           <row>
1613             <entry>
1614               <para>
1615                 <emphasis role="bold">Client options</emphasis>
1616               </para>
1617             </entry>
1618             <entry>
1619               <para>
1620                 <emphasis role="bold">Description</emphasis>
1621               </para>
1622             </entry>
1623           </row>
1624         </thead>
1625         <tbody>
1626           <row>
1627             <entry>
1628               <para>
1629                 <literal>flock/noflock/localflock</literal>
1630               </para>
1631             </entry>
1632             <entry>
1633               <para>Enables/disables global flock or local flock support</para>
1634             </entry>
1635           </row>
1636           <row>
1637             <entry>
1638               <para>
1639                 <literal>user_xattr/nouser_xattr</literal>
1640               </para>
1641             </entry>
1642             <entry>
1643               <para>Enables/disables user-extended attributes</para>
1644             </entry>
1645           </row>
1646           <row>
1647             <entry>
1648               <para>
1649                 <literal>user_fid2path/nouser_fid2path</literal>
1650               </para>
1651             </entry>
1652             <entry>
1653               <para>Enables/disables FID to path translation by
1654               regular users</para>
1655             </entry>
1656           </row>
1657           <row>
1658             <entry>
1659               <para>
1660                 <literal>retry=</literal>
1661               </para>
1662             </entry>
1663             <entry>
1664               <para>Number of times a client will retry to mount the file
1665               system</para>
1666             </entry>
1667           </row>
1668         </tbody>
1669       </tgroup>
1670     </informaltable>
1671   </section>
1672   <section xml:id="dbdoclet.50438206_56217">
1673     <title>Handling Timeouts</title>
1674     <para>Timeouts are the most common cause of hung applications. After a
1675     timeout involving an MDS or failover OST, applications attempting to access
1676     the disconnected resource wait until the connection gets
1677     established.</para>
1678     <para>When a client performs any remote operation, it gives the server a
1679     reasonable amount of time to respond. If a server does not reply either due
1680     to a down network, hung server, or any other reason, a timeout occurs which
1681     requires a recovery.</para>
1682     <para>If a timeout occurs, a message (similar to this one), appears on the
1683     console of the client, and in
1684     <literal>/var/log/messages</literal>:</para>
1685     <screen>
1686 LustreError: 26597:(client.c:810:ptlrpc_expire_one_request()) @@@ timeout
1687
1688 req@a2d45200 x5886/t0 o38-&gt;mds_svc_UUID@NID_mds_UUID:12 lens 168/64 ref 1 fl
1689
1690 RPC:/0/0 rc 0
1691 </screen>
1692   </section>
1693 </chapter>