maintenance/backup.inc

   1 <?php
   2 /**
   3  * Base classes for database dumpers
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   6  * https://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Dump Maintenance
  25  */
  26
  27 require_once __DIR__ . '/Maintenance.php';
  28 require_once __DIR__ . '/../includes/export/DumpFilter.php';
  29
  30 /**
  31  * @ingroup Dump Maintenance
  32  */
  33 class BackupDumper extends Maintenance {
  34         public $reporting = true;
  35         public $pages = null; // all pages
  36         public $skipHeader = false; // don't output <mediawiki> and <siteinfo>
  37         public $skipFooter = false; // don't output </mediawiki>
  38         public $startId = 0;
  39         public $endId = 0;
  40         public $revStartId = 0;
  41         public $revEndId = 0;
  42         public $dumpUploads = false;
  43         public $dumpUploadFileContents = false;
  44
  45         protected $reportingInterval = 100;
  46         protected $pageCount = 0;
  47         protected $revCount = 0;
  48         protected $server = null; // use default
  49         protected $sink = null; // Output filters
  50         protected $lastTime = 0;
  51         protected $pageCountLast = 0;
  52         protected $revCountLast = 0;
  53
  54         protected $outputTypes = [];
  55         protected $filterTypes = [];
  56
  57         protected $ID = 0;
  58
  59         /**
  60          * The dependency-injected database to use.
  61          *
  62          * @var DatabaseBase|null
  63          *
  64          * @see self::setDB
  65          */
  66         protected $forcedDb = null;
  67
  68         /** @var LoadBalancer */
  69         protected $lb;
  70
  71         // @todo Unused?
  72         private $stubText = false; // include rev_text_id instead of text; for 2-pass dump
  73
  74         /**
  75          * @param array $args For backward compatibility
  76          */
  77         function __construct( $args = null ) {
  78                 parent::__construct();
  79                 $this->stderr = fopen( "php://stderr", "wt" );
  80
  81                 // Built-in output and filter plugins
  82                 $this->registerOutput( 'file', 'DumpFileOutput' );
  83                 $this->registerOutput( 'gzip', 'DumpGZipOutput' );
  84                 $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
  85                 $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
  86                 $this->registerOutput( '7zip', 'Dump7ZipOutput' );
  87
  88                 $this->registerFilter( 'latest', 'DumpLatestFilter' );
  89                 $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
  90                 $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
  91
  92                 // These three can be specified multiple times
  93                 $this->addOption( 'plugin', 'Load a dump plugin class. Specify as <class>[:<file>].',
  94                         false, true, false, true );
  95                 $this->addOption( 'output', 'Begin a filtered output stream; Specify as <type>:<file>. ' .
  96                         '<type>s: file, gzip, bzip2, 7zip, dbzip2', false, true, false, true );
  97                 $this->addOption( 'filter', 'Add a filter on an output branch. Specify as ' .
  98                         '<type>[:<options>]. <types>s: latest, notalk, namespace', false, true, false, true );
  99                 $this->addOption( 'report', 'Report position and speed after every n pages processed. ' .
 100                         'Default: 100.', false, true );
 101                 $this->addOption( 'server', 'Force reading from MySQL server', false, true );
 102                 $this->addOption( '7ziplevel', '7zip compression level for all 7zip outputs. Used for ' .
 103                         '-mx option to 7za command.', false, true );
 104
 105                 if ( $args ) {
 106                         // Args should be loaded and processed so that dump() can be called directly
 107                         // instead of execute()
 108                         $this->loadWithArgv( $args );
 109                         $this->processOptions();
 110                 }
 111         }
 112
 113         /**
 114          * @param string $name
 115          * @param string $class Name of output filter plugin class
 116          */
 117         function registerOutput( $name, $class ) {
 118                 $this->outputTypes[$name] = $class;
 119         }
 120
 121         /**
 122          * @param string $name
 123          * @param string $class Name of filter plugin class
 124          */
 125         function registerFilter( $name, $class ) {
 126                 $this->filterTypes[$name] = $class;
 127         }
 128
 129         /**
 130          * Load a plugin and register it
 131          *
 132          * @param string $class Name of plugin class; must have a static 'register'
 133          *   method that takes a BackupDumper as a parameter.
 134          * @param string $file Full or relative path to the PHP file to load, or empty
 135          */
 136         function loadPlugin( $class, $file ) {
 137                 if ( $file != '' ) {
 138                         require_once $file;
 139                 }
 140                 $register = [ $class, 'register' ];
 141                 call_user_func_array( $register, [ $this ] );
 142         }
 143
 144         function execute() {
 145                 throw new MWException( 'execute() must be overridden in subclasses' );
 146         }
 147
 148         /**
 149          * Processes arguments and sets $this->$sink accordingly
 150          */
 151         function processOptions() {
 152                 $sink = null;
 153                 $sinks = [];
 154
 155                 $options = $this->orderedOptions;
 156                 foreach ( $options as $arg ) {
 157                         $opt = $arg[0];
 158                         $param = $arg[1];
 159
 160                         switch ( $opt ) {
 161                                 case 'plugin':
 162                                         $val = explode( ':', $param );
 163
 164                                         if ( count( $val ) === 1 ) {
 165                                                 $this->loadPlugin( $val[0] );
 166                                         } elseif ( count( $val ) === 2 ) {
 167                                                 $this->loadPlugin( $val[0], $val[1] );
 168                                         } else {
 169                                                 $this->fatalError( 'Invalid plugin parameter' );
 170                                                 return;
 171                                         }
 172
 173                                         break;
 174                                 case 'output':
 175                                         $split = explode( ':', $param, 2 );
 176                                         if ( count( $split ) !== 2 ) {
 177                                                 $this->fatalError( 'Invalid output parameter' );
 178                                         }
 179                                         list( $type, $file ) = $split;
 180                                         if ( !is_null( $sink ) ) {
 181                                                 $sinks[] = $sink;
 182                                         }
 183                                         if ( !isset( $this->outputTypes[$type] ) ) {
 184                                                 $this->fatalError( "Unrecognized output sink type '$type'" );
 185                                         }
 186                                         $class = $this->outputTypes[$type];
 187                                         if ( $type === "7zip" ) {
 188                                                 $sink = new $class( $file, intval( $this->getOption( '7ziplevel' ) ) );
 189                                         } else {
 190                                                 $sink = new $class( $file );
 191                                         }
 192
 193                                         break;
 194                                 case 'filter':
 195                                         if ( is_null( $sink ) ) {
 196                                                 $sink = new DumpOutput();
 197                                         }
 198
 199                                         $split = explode( ':', $param );
 200                                         $key = $split[0];
 201
 202                                         if ( !isset( $this->filterTypes[$key] ) ) {
 203                                                 $this->fatalError( "Unrecognized filter type '$key'" );
 204                                         }
 205
 206                                         $type = $this->filterTypes[$key];
 207
 208                                         if ( count( $split ) === 1 ) {
 209                                                 $filter = new $type( $sink );
 210                                         } elseif ( count( $split ) === 2 ) {
 211                                                 $filter = new $type( $sink, $split[1] );
 212                                         } else {
 213                                                 $this->fatalError( 'Invalid filter parameter' );
 214                                         }
 215
 216                                         // references are lame in php...
 217                                         unset( $sink );
 218                                         $sink = $filter;
 219
 220                                         break;
 221                         }
 222                 }
 223
 224                 if ( $this->hasOption( 'report' ) ) {
 225                         $this->reportingInterval = intval( $this->getOption( 'report' ) );
 226                 }
 227
 228                 if ( $this->hasOption( 'server' ) ) {
 229                         $this->server = $this->getOption( 'server' );
 230                 }
 231
 232                 if ( is_null( $sink ) ) {
 233                         $sink = new DumpOutput();
 234                 }
 235                 $sinks[] = $sink;
 236
 237                 if ( count( $sinks ) > 1 ) {
 238                         $this->sink = new DumpMultiWriter( $sinks );
 239                 } else {
 240                         $this->sink = $sink;
 241                 }
 242         }
 243
 244         function dump( $history, $text = WikiExporter::TEXT ) {
 245                 # Notice messages will foul up your XML output even if they're
 246                 # relatively harmless.
 247                 if ( ini_get( 'display_errors' ) ) {
 248                         ini_set( 'display_errors', 'stderr' );
 249                 }
 250
 251                 $this->initProgress( $history );
 252
 253                 $db = $this->backupDb();
 254                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 255                 $exporter->dumpUploads = $this->dumpUploads;
 256                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 257
 258                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 259                 $exporter->setOutputSink( $wrapper );
 260
 261                 if ( !$this->skipHeader ) {
 262                         $exporter->openStream();
 263                 }
 264                 # Log item dumps: all or by range
 265                 if ( $history & WikiExporter::LOGS ) {
 266                         if ( $this->startId || $this->endId ) {
 267                                 $exporter->logsByRange( $this->startId, $this->endId );
 268                         } else {
 269                                 $exporter->allLogs();
 270                         }
 271                 } elseif ( is_null( $this->pages ) ) {
 272                         # Page dumps: all or by page ID range
 273                         if ( $this->startId || $this->endId ) {
 274                                 $exporter->pagesByRange( $this->startId, $this->endId );
 275                         } elseif ( $this->revStartId || $this->revEndId ) {
 276                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 277                         } else {
 278                                 $exporter->allPages();
 279                         }
 280                 } else {
 281                         # Dump of specific pages
 282                         $exporter->pagesByName( $this->pages );
 283                 }
 284
 285                 if ( !$this->skipFooter ) {
 286                         $exporter->closeStream();
 287                 }
 288
 289                 $this->report( true );
 290         }
 291
 292         /**
 293          * Initialise starting time and maximum revision count.
 294          * We'll make ETA calculations based an progress, assuming relatively
 295          * constant per-revision rate.
 296          * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
 297          */
 298         function initProgress( $history = WikiExporter::FULL ) {
 299                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 300                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 301
 302                 $dbr = $this->forcedDb;
 303                 if ( $this->forcedDb === null ) {
 304                         $dbr = wfGetDB( DB_SLAVE );
 305                 }
 306                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 307                 $this->startTime = microtime( true );
 308                 $this->lastTime = $this->startTime;
 309                 $this->ID = getmypid();
 310         }
 311
 312         /**
 313          * @todo Fixme: the --server parameter is currently not respected, as it
 314          * doesn't seem terribly easy to ask the load balancer for a particular
 315          * connection by name.
 316          * @return DatabaseBase
 317          */
 318         function backupDb() {
 319                 if ( $this->forcedDb !== null ) {
 320                         return $this->forcedDb;
 321                 }
 322
 323                 $this->lb = wfGetLBFactory()->newMainLB();
 324                 $db = $this->lb->getConnection( DB_SLAVE, 'dump' );
 325
 326                 // Discourage the server from disconnecting us if it takes a long time
 327                 // to read out the big ol' batch query.
 328                 $db->setSessionOptions( [ 'connTimeout' => 3600 * 24 ] );
 329
 330                 return $db;
 331         }
 332
 333         /**
 334          * Force the dump to use the provided database connection for database
 335          * operations, wherever possible.
 336          *
 337          * @param DatabaseBase|null $db (Optional) the database connection to use. If null, resort to
 338          *   use the globally provided ways to get database connections.
 339          */
 340         function setDB( IDatabase $db = null ) {
 341                 parent::setDB( $db );
 342                 $this->forcedDb = $db;
 343         }
 344
 345         function __destruct() {
 346                 if ( isset( $this->lb ) ) {
 347                         $this->lb->closeAll();
 348                 }
 349         }
 350
 351         function backupServer() {
 352                 global $wgDBserver;
 353
 354                 return $this->server
 355                         ? $this->server
 356                         : $wgDBserver;
 357         }
 358
 359         function reportPage() {
 360                 $this->pageCount++;
 361         }
 362
 363         function revCount() {
 364                 $this->revCount++;
 365                 $this->report();
 366         }
 367
 368         function report( $final = false ) {
 369                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 370                         $this->showReport();
 371                 }
 372         }
 373
 374         function showReport() {
 375                 if ( $this->reporting ) {
 376                         $now = wfTimestamp( TS_DB );
 377                         $nowts = microtime( true );
 378                         $deltaAll = $nowts - $this->startTime;
 379                         $deltaPart = $nowts - $this->lastTime;
 380                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 381                         $this->revCountPart = $this->revCount - $this->revCountLast;
 382
 383                         if ( $deltaAll ) {
 384                                 $portion = $this->revCount / $this->maxCount;
 385                                 $eta = $this->startTime + $deltaAll / $portion;
 386                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 387                                 $pageRate = $this->pageCount / $deltaAll;
 388                                 $revRate = $this->revCount / $deltaAll;
 389                         } else {
 390                                 $pageRate = '-';
 391                                 $revRate = '-';
 392                                 $etats = '-';
 393                         }
 394                         if ( $deltaPart ) {
 395                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 396                                 $revRatePart = $this->revCountPart / $deltaPart;
 397                         } else {
 398                                 $pageRatePart = '-';
 399                                 $revRatePart = '-';
 400                         }
 401                         $this->progress( sprintf(
 402                                 "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
 403                                         . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 404                                 $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate,
 405                                 $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
 406                                 $this->maxCount
 407                         ) );
 408                         $this->lastTime = $nowts;
 409                         $this->revCountLast = $this->revCount;
 410                 }
 411         }
 412
 413         function progress( $string ) {
 414                 if ( $this->reporting ) {
 415                         fwrite( $this->stderr, $string . "\n" );
 416                 }
 417         }
 418
 419         function fatalError( $msg ) {
 420                 $this->error( "$msg\n", 1 );
 421         }
 422 }
 423
 424 class ExportProgressFilter extends DumpFilter {
 425         function __construct( &$sink, &$progress ) {
 426                 parent::__construct( $sink );
 427                 $this->progress = $progress;
 428         }
 429
 430         function writeClosePage( $string ) {
 431                 parent::writeClosePage( $string );
 432                 $this->progress->reportPage();
 433         }
 434
 435         function writeRevision( $rev, $string ) {
 436                 parent::writeRevision( $rev, $string );
 437                 $this->progress->revCount();
 438         }
 439 }