maintenance/includes/BackupDumper.php

   1 <?php
   2 /**
   3  * Base classes for database-dumping maintenance scripts.
   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
  25  * @ingroup Maintenance
  26  */
  27
  28 require_once __DIR__ . '/../Maintenance.php';
  29 require_once __DIR__ . '/../../includes/export/WikiExporter.php';
  30
  31 use MediaWiki\MediaWikiServices;
  32 use Wikimedia\Rdbms\LoadBalancer;
  33 use Wikimedia\Rdbms\IMaintainableDatabase;
  34
  35 /**
  36  * @ingroup Dump
  37  * @ingroup Maintenance
  38  */
  39 abstract class BackupDumper extends Maintenance {
  40         public $reporting = true;
  41         public $pages = null; // all pages
  42         public $skipHeader = false; // don't output <mediawiki> and <siteinfo>
  43         public $skipFooter = false; // don't output </mediawiki>
  44         public $startId = 0;
  45         public $endId = 0;
  46         public $revStartId = 0;
  47         public $revEndId = 0;
  48         public $dumpUploads = false;
  49         public $dumpUploadFileContents = false;
  50         public $orderRevs = false;
  51         public $limitNamespaces = [];
  52         /** @var bool|resource */
  53         public $stderr;
  54
  55         protected $reportingInterval = 100;
  56         protected $pageCount = 0;
  57         protected $revCount = 0;
  58         protected $schemaVersion = null; // use default
  59         protected $server = null; // use default
  60         protected $sink = null; // Output filters
  61         protected $lastTime = 0;
  62         protected $pageCountLast = 0;
  63         protected $revCountLast = 0;
  64
  65         protected $outputTypes = [];
  66         protected $filterTypes = [];
  67
  68         protected $ID = 0;
  69
  70         /** @var int */
  71         protected $startTime;
  72         /** @var int */
  73         protected $pageCountPart;
  74         /** @var int */
  75         protected $revCountPart;
  76         /** @var int */
  77         protected $maxCount;
  78         /** @var int */
  79         protected $timeOfCheckpoint;
  80         /** @var ExportProgressFilter */
  81         protected $egress;
  82         /** @var string */
  83         protected $buffer;
  84         /** @var array|false */
  85         protected $openElement;
  86         /** @var bool */
  87         protected $atStart;
  88         /** @var string|null */
  89         protected $thisRevModel;
  90         /** @var string|null */
  91         protected $thisRevFormat;
  92         /** @var string */
  93         protected $lastName;
  94         /** @var string */
  95         protected $state;
  96
  97         /**
  98          * The dependency-injected database to use.
  99          *
 100          * @var IMaintainableDatabase|null
 101          *
 102          * @see self::setDB
 103          */
 104         protected $forcedDb = null;
 105
 106         /** @var LoadBalancer */
 107         protected $lb;
 108
 109         /**
 110          * @param array|null $args For backward compatibility
 111          */
 112         function __construct( $args = null ) {
 113                 parent::__construct();
 114                 $this->stderr = fopen( "php://stderr", "wt" );
 115
 116                 // Built-in output and filter plugins
 117                 $this->registerOutput( 'file', DumpFileOutput::class );
 118                 $this->registerOutput( 'gzip', DumpGZipOutput::class );
 119                 $this->registerOutput( 'bzip2', DumpBZip2Output::class );
 120                 $this->registerOutput( 'dbzip2', DumpDBZip2Output::class );
 121                 $this->registerOutput( 'lbzip2', DumpLBZip2Output::class );
 122                 $this->registerOutput( '7zip', Dump7ZipOutput::class );
 123
 124                 $this->registerFilter( 'latest', DumpLatestFilter::class );
 125                 $this->registerFilter( 'notalk', DumpNotalkFilter::class );
 126                 $this->registerFilter( 'namespace', DumpNamespaceFilter::class );
 127
 128                 // These three can be specified multiple times
 129                 $this->addOption( 'plugin', 'Load a dump plugin class. Specify as <class>[:<file>].',
 130                         false, true, false, true );
 131                 $this->addOption( 'output', 'Begin a filtered output stream; Specify as <type>:<file>. ' .
 132                         '<type>s: file, gzip, bzip2, 7zip, dbzip2, lbzip2', false, true, false, true );
 133                 $this->addOption( 'filter', 'Add a filter on an output branch. Specify as ' .
 134                         '<type>[:<options>]. <types>s: latest, notalk, namespace', false, true, false, true );
 135                 $this->addOption( 'report', 'Report position and speed after every n pages processed. ' .
 136                         'Default: 100.', false, true );
 137                 $this->addOption( 'schema-version', 'Schema version to use for output. ' .
 138                         'Default: ' . WikiExporter::schemaVersion(), false, true );
 139                 $this->addOption( 'server', 'Force reading from MySQL server', false, true );
 140                 $this->addOption( '7ziplevel', '7zip compression level for all 7zip outputs. Used for ' .
 141                         '-mx option to 7za command.', false, true );
 142
 143                 if ( $args ) {
 144                         // Args should be loaded and processed so that dump() can be called directly
 145                         // instead of execute()
 146                         $this->loadWithArgv( $args );
 147                         $this->processOptions();
 148                 }
 149         }
 150
 151         /**
 152          * @param string $name
 153          * @param string $class Name of output filter plugin class
 154          */
 155         function registerOutput( $name, $class ) {
 156                 $this->outputTypes[$name] = $class;
 157         }
 158
 159         /**
 160          * @param string $name
 161          * @param string $class Name of filter plugin class
 162          */
 163         function registerFilter( $name, $class ) {
 164                 $this->filterTypes[$name] = $class;
 165         }
 166
 167         /**
 168          * Load a plugin and register it
 169          *
 170          * @param string $class Name of plugin class; must have a static 'register'
 171          *   method that takes a BackupDumper as a parameter.
 172          * @param string $file Full or relative path to the PHP file to load, or empty
 173          */
 174         function loadPlugin( $class, $file ) {
 175                 if ( $file != '' ) {
 176                         require_once $file;
 177                 }
 178                 $register = [ $class, 'register' ];
 179                 $register( $this );
 180         }
 181
 182         function execute() {
 183                 throw new MWException( 'execute() must be overridden in subclasses' );
 184         }
 185
 186         /**
 187          * Processes arguments and sets $this->$sink accordingly
 188          */
 189         function processOptions() {
 190                 $sink = null;
 191                 $sinks = [];
 192
 193                 $this->schemaVersion = WikiExporter::schemaVersion();
 194
 195                 $options = $this->orderedOptions;
 196                 foreach ( $options as $arg ) {
 197                         list( $opt, $param ) = $arg;
 198
 199                         switch ( $opt ) {
 200                                 case 'plugin':
 201                                         $val = explode( ':', $param, 2 );
 202
 203                                         if ( count( $val ) === 1 ) {
 204                                                 $this->loadPlugin( $val[0], '' );
 205                                         } elseif ( count( $val ) === 2 ) {
 206                                                 $this->loadPlugin( $val[0], $val[1] );
 207                                         }
 208
 209                                         break;
 210                                 case 'output':
 211                                         $split = explode( ':', $param, 2 );
 212                                         if ( count( $split ) !== 2 ) {
 213                                                 $this->fatalError( 'Invalid output parameter' );
 214                                         }
 215                                         list( $type, $file ) = $split;
 216                                         if ( !is_null( $sink ) ) {
 217                                                 $sinks[] = $sink;
 218                                         }
 219                                         if ( !isset( $this->outputTypes[$type] ) ) {
 220                                                 $this->fatalError( "Unrecognized output sink type '$type'" );
 221                                         }
 222                                         $class = $this->outputTypes[$type];
 223                                         if ( $type === "7zip" ) {
 224                                                 $sink = new $class( $file, intval( $this->getOption( '7ziplevel' ) ) );
 225                                         } else {
 226                                                 $sink = new $class( $file );
 227                                         }
 228
 229                                         break;
 230                                 case 'filter':
 231                                         if ( is_null( $sink ) ) {
 232                                                 $sink = new DumpOutput();
 233                                         }
 234
 235                                         $split = explode( ':', $param, 2 );
 236                                         $key = $split[0];
 237
 238                                         if ( !isset( $this->filterTypes[$key] ) ) {
 239                                                 $this->fatalError( "Unrecognized filter type '$key'" );
 240                                         }
 241
 242                                         $type = $this->filterTypes[$key];
 243
 244                                         if ( count( $split ) === 1 ) {
 245                                                 $filter = new $type( $sink );
 246                                         } elseif ( count( $split ) === 2 ) {
 247                                                 $filter = new $type( $sink, $split[1] );
 248                                         }
 249
 250                                         // references are lame in php...
 251                                         unset( $sink );
 252                                         $sink = $filter;
 253
 254                                         break;
 255                                 case 'schema-version':
 256                                         if ( !in_array( $param, XmlDumpWriter::$supportedSchemas ) ) {
 257                                                 $this->fatalError(
 258                                                         "Unsupported schema version $param. Supported versions: " .
 259                                                         implode( ', ', XmlDumpWriter::$supportedSchemas )
 260                                                 );
 261                                         }
 262                                         $this->schemaVersion = $param;
 263                                         break;
 264                         }
 265                 }
 266
 267                 if ( $this->hasOption( 'report' ) ) {
 268                         $this->reportingInterval = intval( $this->getOption( 'report' ) );
 269                 }
 270
 271                 if ( $this->hasOption( 'server' ) ) {
 272                         $this->server = $this->getOption( 'server' );
 273                 }
 274
 275                 if ( is_null( $sink ) ) {
 276                         $sink = new DumpOutput();
 277                 }
 278                 $sinks[] = $sink;
 279
 280                 if ( count( $sinks ) > 1 ) {
 281                         $this->sink = new DumpMultiWriter( $sinks );
 282                 } else {
 283                         $this->sink = $sink;
 284                 }
 285         }
 286
 287         function dump( $history, $text = WikiExporter::TEXT ) {
 288                 # Notice messages will foul up your XML output even if they're
 289                 # relatively harmless.
 290                 if ( ini_get( 'display_errors' ) ) {
 291                         ini_set( 'display_errors', 'stderr' );
 292                 }
 293
 294                 $this->initProgress( $history );
 295
 296                 $db = $this->backupDb();
 297                 $exporter = new WikiExporter( $db, $history, $text, $this->limitNamespaces );
 298                 $exporter->setSchemaVersion( $this->schemaVersion );
 299                 $exporter->dumpUploads = $this->dumpUploads;
 300                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 301
 302                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 303                 $exporter->setOutputSink( $wrapper );
 304
 305                 if ( !$this->skipHeader ) {
 306                         $exporter->openStream();
 307                 }
 308                 # Log item dumps: all or by range
 309                 if ( $history & WikiExporter::LOGS ) {
 310                         if ( $this->startId || $this->endId ) {
 311                                 $exporter->logsByRange( $this->startId, $this->endId );
 312                         } else {
 313                                 $exporter->allLogs();
 314                         }
 315                 } elseif ( is_null( $this->pages ) ) {
 316                         # Page dumps: all or by page ID range
 317                         if ( $this->startId || $this->endId ) {
 318                                 $exporter->pagesByRange( $this->startId, $this->endId, $this->orderRevs );
 319                         } elseif ( $this->revStartId || $this->revEndId ) {
 320                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 321                         } else {
 322                                 $exporter->allPages();
 323                         }
 324                 } else {
 325                         # Dump of specific pages
 326                         $exporter->pagesByName( $this->pages );
 327                 }
 328
 329                 if ( !$this->skipFooter ) {
 330                         $exporter->closeStream();
 331                 }
 332
 333                 $this->report( true );
 334         }
 335
 336         /**
 337          * Initialise starting time and maximum revision count.
 338          * We'll make ETA calculations based an progress, assuming relatively
 339          * constant per-revision rate.
 340          * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
 341          */
 342         function initProgress( $history = WikiExporter::FULL ) {
 343                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 344                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 345
 346                 $dbr = $this->forcedDb;
 347                 if ( $this->forcedDb === null ) {
 348                         $dbr = $this->getDB( DB_REPLICA );
 349                 }
 350                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 351                 $this->startTime = microtime( true );
 352                 $this->lastTime = $this->startTime;
 353                 $this->ID = getmypid();
 354         }
 355
 356         /**
 357          * @todo Fixme: the --server parameter is currently not respected, as it
 358          * doesn't seem terribly easy to ask the load balancer for a particular
 359          * connection by name.
 360          * @return IMaintainableDatabase
 361          */
 362         function backupDb() {
 363                 if ( $this->forcedDb !== null ) {
 364                         return $this->forcedDb;
 365                 }
 366
 367                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 368                 $this->lb = $lbFactory->newMainLB();
 369                 $db = $this->lb->getMaintenanceConnectionRef( DB_REPLICA, 'dump' );
 370
 371                 // Discourage the server from disconnecting us if it takes a long time
 372                 // to read out the big ol' batch query.
 373                 $db->setSessionOptions( [ 'connTimeout' => 3600 * 24 ] );
 374
 375                 return $db;
 376         }
 377
 378         /**
 379          * Force the dump to use the provided database connection for database
 380          * operations, wherever possible.
 381          *
 382          * @param IMaintainableDatabase $db The database connection to use
 383          */
 384         function setDB( IMaintainableDatabase $db ) {
 385                 parent::setDB( $db );
 386                 $this->forcedDb = $db;
 387         }
 388
 389         function __destruct() {
 390                 if ( isset( $this->lb ) ) {
 391                         $this->lb->closeAll();
 392                 }
 393         }
 394
 395         function backupServer() {
 396                 global $wgDBserver;
 397
 398                 return $this->server ?: $wgDBserver;
 399         }
 400
 401         function reportPage() {
 402                 $this->pageCount++;
 403         }
 404
 405         function revCount() {
 406                 $this->revCount++;
 407                 $this->report();
 408         }
 409
 410         function report( $final = false ) {
 411                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 412                         $this->showReport();
 413                 }
 414         }
 415
 416         function showReport() {
 417                 if ( $this->reporting ) {
 418                         $now = wfTimestamp( TS_DB );
 419                         $nowts = microtime( true );
 420                         $deltaAll = $nowts - $this->startTime;
 421                         $deltaPart = $nowts - $this->lastTime;
 422                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 423                         $this->revCountPart = $this->revCount - $this->revCountLast;
 424
 425                         if ( $deltaAll ) {
 426                                 $portion = $this->revCount / $this->maxCount;
 427                                 $eta = $this->startTime + $deltaAll / $portion;
 428                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 429                                 $pageRate = $this->pageCount / $deltaAll;
 430                                 $revRate = $this->revCount / $deltaAll;
 431                         } else {
 432                                 $pageRate = '-';
 433                                 $revRate = '-';
 434                                 $etats = '-';
 435                         }
 436                         if ( $deltaPart ) {
 437                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 438                                 $revRatePart = $this->revCountPart / $deltaPart;
 439                         } else {
 440                                 $pageRatePart = '-';
 441                                 $revRatePart = '-';
 442                         }
 443
 444                         $dbDomain = WikiMap::getCurrentWikiDbDomain()->getId();
 445                         $this->progress( sprintf(
 446                                 "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
 447                                         . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 448                                 $now, $dbDomain, $this->ID, $this->pageCount, $pageRate,
 449                                 $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
 450                                 $this->maxCount
 451                         ) );
 452                         $this->lastTime = $nowts;
 453                         $this->revCountLast = $this->revCount;
 454                 }
 455         }
 456
 457         function progress( $string ) {
 458                 if ( $this->reporting ) {
 459                         fwrite( $this->stderr, $string . "\n" );
 460                 }
 461         }
 462 }