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 = array();
  55         protected $filterTypes = array();
  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
 103                 if ( $args ) {
 104                         // Args should be loaded and processed so that dump() can be called directly
 105                         // instead of execute()
 106                         $this->loadWithArgv( $args );
 107                         $this->processOptions();
 108                 }
 109         }
 110
 111         /**
 112          * @param string $name
 113          * @param string $class Name of output filter plugin class
 114          */
 115         function registerOutput( $name, $class ) {
 116                 $this->outputTypes[$name] = $class;
 117         }
 118
 119         /**
 120          * @param string $name
 121          * @param string $class Name of filter plugin class
 122          */
 123         function registerFilter( $name, $class ) {
 124                 $this->filterTypes[$name] = $class;
 125         }
 126
 127         /**
 128          * Load a plugin and register it
 129          *
 130          * @param string $class Name of plugin class; must have a static 'register'
 131          *   method that takes a BackupDumper as a parameter.
 132          * @param string $file Full or relative path to the PHP file to load, or empty
 133          */
 134         function loadPlugin( $class, $file ) {
 135                 if ( $file != '' ) {
 136                         require_once $file;
 137                 }
 138                 $register = array( $class, 'register' );
 139                 call_user_func_array( $register, array( &$this ) );
 140         }
 141
 142         function execute() {
 143                 throw new MWException( 'execute() must be overridden in subclasses' );
 144         }
 145
 146         /**
 147          * Processes arguments and sets $this->$sink accordingly
 148          */
 149         function processOptions() {
 150                 $sink = null;
 151                 $sinks = array();
 152
 153                 $options = $this->orderedOptions;
 154                 foreach ( $options as $arg ) {
 155                         $opt = $arg[0];
 156                         $param = $arg[1];
 157
 158                         switch ( $opt ) {
 159                                 case 'plugin':
 160                                         $val = explode( ':', $param );
 161
 162                                         if ( count( $val ) === 1 ) {
 163                                                 $this->loadPlugin( $val[0] );
 164                                         } elseif ( count( $val ) === 2 ) {
 165                                                 $this->loadPlugin( $val[0], $val[1] );
 166                                         } else {
 167                                                 $this->fatalError( 'Invalid plugin parameter' );
 168                                                 return;
 169                                         }
 170
 171                                         break;
 172                                 case 'output':
 173                                         $split = explode( ':', $param, 2 );
 174                                         if ( count( $split ) !== 2 ) {
 175                                                 $this->fatalError( 'Invalid output parameter' );
 176                                         }
 177                                         list( $type, $file ) = $split;
 178                                         if ( !is_null( $sink ) ) {
 179                                                 $sinks[] = $sink;
 180                                         }
 181                                         if ( !isset( $this->outputTypes[$type] ) ) {
 182                                                 $this->fatalError( "Unrecognized output sink type '$type'" );
 183                                         }
 184                                         $class = $this->outputTypes[$type];
 185                                         $sink = new $class( $file );
 186
 187                                         break;
 188                                 case 'filter':
 189                                         if ( is_null( $sink ) ) {
 190                                                 $sink = new DumpOutput();
 191                                         }
 192
 193                                         $split = explode( ':', $param );
 194                                         $key = $split[0];
 195
 196                                         if ( !isset( $this->filterTypes[$key] ) ) {
 197                                                 $this->fatalError( "Unrecognized filter type '$key'" );
 198                                         }
 199
 200                                         $type = $this->filterTypes[$key];
 201
 202                                         if ( count( $split ) === 1 ) {
 203                                                 $filter = new $type( $sink );
 204                                         } elseif ( count( $split ) === 2 ) {
 205                                                 $filter = new $type( $sink, $split[1] );
 206                                         } else {
 207                                                 $this->fatalError( 'Invalid filter parameter' );
 208                                         }
 209
 210                                         // references are lame in php...
 211                                         unset( $sink );
 212                                         $sink = $filter;
 213
 214                                         break;
 215                         }
 216                 }
 217
 218                 if ( $this->hasOption( 'report' ) ) {
 219                         $this->reportingInterval = intval( $this->getOption( 'report' ) );
 220                 }
 221
 222                 if ( $this->hasOption( 'server' ) ) {
 223                         $this->server = $this->getOption( 'server' );
 224                 }
 225
 226                 if ( is_null( $sink ) ) {
 227                         $sink = new DumpOutput();
 228                 }
 229                 $sinks[] = $sink;
 230
 231                 if ( count( $sinks ) > 1 ) {
 232                         $this->sink = new DumpMultiWriter( $sinks );
 233                 } else {
 234                         $this->sink = $sink;
 235                 }
 236         }
 237
 238         function dump( $history, $text = WikiExporter::TEXT ) {
 239                 # Notice messages will foul up your XML output even if they're
 240                 # relatively harmless.
 241                 if ( ini_get( 'display_errors' ) ) {
 242                         ini_set( 'display_errors', 'stderr' );
 243                 }
 244
 245                 $this->initProgress( $history );
 246
 247                 $db = $this->backupDb();
 248                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 249                 $exporter->dumpUploads = $this->dumpUploads;
 250                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 251
 252                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 253                 $exporter->setOutputSink( $wrapper );
 254
 255                 if ( !$this->skipHeader ) {
 256                         $exporter->openStream();
 257                 }
 258                 # Log item dumps: all or by range
 259                 if ( $history & WikiExporter::LOGS ) {
 260                         if ( $this->startId || $this->endId ) {
 261                                 $exporter->logsByRange( $this->startId, $this->endId );
 262                         } else {
 263                                 $exporter->allLogs();
 264                         }
 265                 } elseif ( is_null( $this->pages ) ) {
 266                         # Page dumps: all or by page ID range
 267                         if ( $this->startId || $this->endId ) {
 268                                 $exporter->pagesByRange( $this->startId, $this->endId );
 269                         } elseif ( $this->revStartId || $this->revEndId ) {
 270                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 271                         } else {
 272                                 $exporter->allPages();
 273                         }
 274                 } else {
 275                         # Dump of specific pages
 276                         $exporter->pagesByName( $this->pages );
 277                 }
 278
 279                 if ( !$this->skipFooter ) {
 280                         $exporter->closeStream();
 281                 }
 282
 283                 $this->report( true );
 284         }
 285
 286         /**
 287          * Initialise starting time and maximum revision count.
 288          * We'll make ETA calculations based an progress, assuming relatively
 289          * constant per-revision rate.
 290          * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
 291          */
 292         function initProgress( $history = WikiExporter::FULL ) {
 293                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 294                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 295
 296                 $dbr = $this->forcedDb;
 297                 if ( $this->forcedDb === null ) {
 298                         $dbr = wfGetDB( DB_SLAVE );
 299                 }
 300                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 301                 $this->startTime = microtime( true );
 302                 $this->lastTime = $this->startTime;
 303                 $this->ID = getmypid();
 304         }
 305
 306         /**
 307          * @todo Fixme: the --server parameter is currently not respected, as it
 308          * doesn't seem terribly easy to ask the load balancer for a particular
 309          * connection by name.
 310          * @return DatabaseBase
 311          */
 312         function backupDb() {
 313                 if ( $this->forcedDb !== null ) {
 314                         return $this->forcedDb;
 315                 }
 316
 317                 $this->lb = wfGetLBFactory()->newMainLB();
 318                 $db = $this->lb->getConnection( DB_SLAVE, 'dump' );
 319
 320                 // Discourage the server from disconnecting us if it takes a long time
 321                 // to read out the big ol' batch query.
 322                 $db->setSessionOptions( array( 'connTimeout' => 3600 * 24 ) );
 323
 324                 return $db;
 325         }
 326
 327         /**
 328          * Force the dump to use the provided database connection for database
 329          * operations, wherever possible.
 330          *
 331          * @param DatabaseBase|null $db (Optional) the database connection to use. If null, resort to
 332          *   use the globally provided ways to get database connections.
 333          */
 334         function setDB( IDatabase $db = null ) {
 335                 parent::setDB( $db );
 336                 $this->forcedDb = $db;
 337         }
 338
 339         function __destruct() {
 340                 if ( isset( $this->lb ) ) {
 341                         $this->lb->closeAll();
 342                 }
 343         }
 344
 345         function backupServer() {
 346                 global $wgDBserver;
 347
 348                 return $this->server
 349                         ? $this->server
 350                         : $wgDBserver;
 351         }
 352
 353         function reportPage() {
 354                 $this->pageCount++;
 355         }
 356
 357         function revCount() {
 358                 $this->revCount++;
 359                 $this->report();
 360         }
 361
 362         function report( $final = false ) {
 363                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 364                         $this->showReport();
 365                 }
 366         }
 367
 368         function showReport() {
 369                 if ( $this->reporting ) {
 370                         $now = wfTimestamp( TS_DB );
 371                         $nowts = microtime( true );
 372                         $deltaAll = $nowts - $this->startTime;
 373                         $deltaPart = $nowts - $this->lastTime;
 374                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 375                         $this->revCountPart = $this->revCount - $this->revCountLast;
 376
 377                         if ( $deltaAll ) {
 378                                 $portion = $this->revCount / $this->maxCount;
 379                                 $eta = $this->startTime + $deltaAll / $portion;
 380                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 381                                 $pageRate = $this->pageCount / $deltaAll;
 382                                 $revRate = $this->revCount / $deltaAll;
 383                         } else {
 384                                 $pageRate = '-';
 385                                 $revRate = '-';
 386                                 $etats = '-';
 387                         }
 388                         if ( $deltaPart ) {
 389                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 390                                 $revRatePart = $this->revCountPart / $deltaPart;
 391                         } else {
 392                                 $pageRatePart = '-';
 393                                 $revRatePart = '-';
 394                         }
 395                         $this->progress( sprintf(
 396                                 "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
 397                                         . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 398                                 $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate,
 399                                 $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
 400                                 $this->maxCount
 401                         ) );
 402                         $this->lastTime = $nowts;
 403                         $this->revCountLast = $this->revCount;
 404                 }
 405         }
 406
 407         function progress( $string ) {
 408                 if ( $this->reporting ) {
 409                         fwrite( $this->stderr, $string . "\n" );
 410                 }
 411         }
 412
 413         function fatalError( $msg ) {
 414                 $this->error( "$msg\n", 1 );
 415         }
 416 }
 417
 418 class ExportProgressFilter extends DumpFilter {
 419         function __construct( &$sink, &$progress ) {
 420                 parent::__construct( $sink );
 421                 $this->progress = $progress;
 422         }
 423
 424         function writeClosePage( $string ) {
 425                 parent::writeClosePage( $string );
 426                 $this->progress->reportPage();
 427         }
 428
 429         function writeRevision( $rev, $string ) {
 430                 parent::writeRevision( $rev, $string );
 431                 $this->progress->revCount();
 432         }
 433 }