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