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