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 /**
  28  * @ingroup Dump Maintenance
  29  */
  30 class DumpDBZip2Output extends DumpPipeOutput {
  31         function __construct( $file ) {
  32                 parent::__construct( "dbzip2", $file );
  33         }
  34 }
  35
  36 /**
  37  * @ingroup Dump Maintenance
  38  */
  39 class BackupDumper {
  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
  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 = array();
  61         protected $filterTypes = array();
  62
  63         protected $ID = 0;
  64
  65         /**
  66          * The dependency-injected database to use.
  67          *
  68          * @var DatabaseBase|null
  69          *
  70          * @see self::setDb
  71          */
  72         protected $forcedDb = null;
  73
  74         /** @var LoadBalancer */
  75         protected $lb;
  76
  77         // @todo Unused?
  78         private $stubText = false; // include rev_text_id instead of text; for 2-pass dump
  79
  80         function __construct( $args ) {
  81                 $this->stderr = fopen( "php://stderr", "wt" );
  82
  83                 // Built-in output and filter plugins
  84                 $this->registerOutput( 'file', 'DumpFileOutput' );
  85                 $this->registerOutput( 'gzip', 'DumpGZipOutput' );
  86                 $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
  87                 $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
  88                 $this->registerOutput( '7zip', 'Dump7ZipOutput' );
  89
  90                 $this->registerFilter( 'latest', 'DumpLatestFilter' );
  91                 $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
  92                 $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
  93
  94                 $this->sink = $this->processArgs( $args );
  95         }
  96
  97         /**
  98          * @param string $name
  99          * @param string $class Name of output filter plugin class
 100          */
 101         function registerOutput( $name, $class ) {
 102                 $this->outputTypes[$name] = $class;
 103         }
 104
 105         /**
 106          * @param string $name
 107          * @param string $class Name of filter plugin class
 108          */
 109         function registerFilter( $name, $class ) {
 110                 $this->filterTypes[$name] = $class;
 111         }
 112
 113         /**
 114          * Load a plugin and register it
 115          *
 116          * @param string $class Name of plugin class; must have a static 'register'
 117          *   method that takes a BackupDumper as a parameter.
 118          * @param string $file Full or relative path to the PHP file to load, or empty
 119          */
 120         function loadPlugin( $class, $file ) {
 121                 if ( $file != '' ) {
 122                         require_once $file;
 123                 }
 124                 $register = array( $class, 'register' );
 125                 call_user_func_array( $register, array( &$this ) );
 126         }
 127
 128         /**
 129          * @param array $args
 130          * @return array
 131          */
 132         function processArgs( $args ) {
 133                 $sink = null;
 134                 $sinks = array();
 135                 foreach ( $args as $arg ) {
 136                         $matches = array();
 137                         if ( preg_match( '/^--(.+?)(?:=(.+?)(?::(.+?))?)?$/', $arg, $matches ) ) {
 138                                 MediaWiki\suppressWarnings();
 139                                 list( /* $full */, $opt, $val, $param ) = $matches;
 140                                 MediaWiki\restoreWarnings();
 141
 142                                 switch ( $opt ) {
 143                                         case "plugin":
 144                                                 $this->loadPlugin( $val, $param );
 145                                                 break;
 146                                         case "output":
 147                                                 if ( !is_null( $sink ) ) {
 148                                                         $sinks[] = $sink;
 149                                                 }
 150                                                 if ( !isset( $this->outputTypes[$val] ) ) {
 151                                                         $this->fatalError( "Unrecognized output sink type '$val'" );
 152                                                 }
 153                                                 $type = $this->outputTypes[$val];
 154                                                 $sink = new $type( $param );
 155                                                 break;
 156                                         case "filter":
 157                                                 if ( is_null( $sink ) ) {
 158                                                         $sink = new DumpOutput();
 159                                                 }
 160                                                 if ( !isset( $this->filterTypes[$val] ) ) {
 161                                                         $this->fatalError( "Unrecognized filter type '$val'" );
 162                                                 }
 163                                                 $type = $this->filterTypes[$val];
 164                                                 $filter = new $type( $sink, $param );
 165
 166                                                 // references are lame in php...
 167                                                 unset( $sink );
 168                                                 $sink = $filter;
 169
 170                                                 break;
 171                                         case "report":
 172                                                 $this->reportingInterval = intval( $val );
 173                                                 break;
 174                                         case "server":
 175                                                 $this->server = $val;
 176                                                 break;
 177                                         case "force-normal":
 178                                                 if ( !function_exists( 'utf8_normalize' ) ) {
 179                                                         $this->fatalError( "UTF-8 normalization extension not loaded. " .
 180                                                                 "Install or remove --force-normal parameter to use slower code." );
 181                                                 }
 182                                                 break;
 183                                         default:
 184                                                 $this->processOption( $opt, $val, $param );
 185                                 }
 186                         }
 187                 }
 188
 189                 if ( is_null( $sink ) ) {
 190                         $sink = new DumpOutput();
 191                 }
 192                 $sinks[] = $sink;
 193
 194                 if ( count( $sinks ) > 1 ) {
 195                         return new DumpMultiWriter( $sinks );
 196                 } else {
 197                         return $sink;
 198                 }
 199         }
 200
 201         function processOption( $opt, $val, $param ) {
 202                 // extension point for subclasses to add options
 203         }
 204
 205         function dump( $history, $text = WikiExporter::TEXT ) {
 206                 # Notice messages will foul up your XML output even if they're
 207                 # relatively harmless.
 208                 if ( ini_get( 'display_errors' ) ) {
 209                         ini_set( 'display_errors', 'stderr' );
 210                 }
 211
 212                 $this->initProgress( $history );
 213
 214                 $db = $this->backupDb();
 215                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 216                 $exporter->dumpUploads = $this->dumpUploads;
 217                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 218
 219                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 220                 $exporter->setOutputSink( $wrapper );
 221
 222                 if ( !$this->skipHeader ) {
 223                         $exporter->openStream();
 224                 }
 225                 # Log item dumps: all or by range
 226                 if ( $history & WikiExporter::LOGS ) {
 227                         if ( $this->startId || $this->endId ) {
 228                                 $exporter->logsByRange( $this->startId, $this->endId );
 229                         } else {
 230                                 $exporter->allLogs();
 231                         }
 232                 } elseif ( is_null( $this->pages ) ) {
 233                         # Page dumps: all or by page ID range
 234                         if ( $this->startId || $this->endId ) {
 235                                 $exporter->pagesByRange( $this->startId, $this->endId );
 236                         } elseif ( $this->revStartId || $this->revEndId ) {
 237                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 238                         } else {
 239                                 $exporter->allPages();
 240                         }
 241                 } else {
 242                         # Dump of specific pages
 243                         $exporter->pagesByName( $this->pages );
 244                 }
 245
 246                 if ( !$this->skipFooter ) {
 247                         $exporter->closeStream();
 248                 }
 249
 250                 $this->report( true );
 251         }
 252
 253         /**
 254          * Initialise starting time and maximum revision count.
 255          * We'll make ETA calculations based an progress, assuming relatively
 256          * constant per-revision rate.
 257          * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
 258          */
 259         function initProgress( $history = WikiExporter::FULL ) {
 260                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 261                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 262
 263                 $dbr = $this->forcedDb;
 264                 if ( $this->forcedDb === null ) {
 265                         $dbr = wfGetDB( DB_SLAVE );
 266                 }
 267                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 268                 $this->startTime = microtime( true );
 269                 $this->lastTime = $this->startTime;
 270                 $this->ID = getmypid();
 271         }
 272
 273         /**
 274          * @todo Fixme: the --server parameter is currently not respected, as it
 275          * doesn't seem terribly easy to ask the load balancer for a particular
 276          * connection by name.
 277          * @return DatabaseBase
 278          */
 279         function backupDb() {
 280                 if ( $this->forcedDb !== null ) {
 281                         return $this->forcedDb;
 282                 }
 283
 284                 $this->lb = wfGetLBFactory()->newMainLB();
 285                 $db = $this->lb->getConnection( DB_SLAVE, 'dump' );
 286
 287                 // Discourage the server from disconnecting us if it takes a long time
 288                 // to read out the big ol' batch query.
 289                 $db->setSessionOptions( array( 'connTimeout' => 3600 * 24 ) );
 290
 291                 return $db;
 292         }
 293
 294         /**
 295          * Force the dump to use the provided database connection for database
 296          * operations, wherever possible.
 297          *
 298          * @param DatabaseBase|null $db (Optional) the database connection to use. If null, resort to
 299          *   use the globally provided ways to get database connections.
 300          */
 301         function setDb( DatabaseBase $db = null ) {
 302                 $this->forcedDb = $db;
 303         }
 304
 305         function __destruct() {
 306                 if ( isset( $this->lb ) ) {
 307                         $this->lb->closeAll();
 308                 }
 309         }
 310
 311         function backupServer() {
 312                 global $wgDBserver;
 313
 314                 return $this->server
 315                         ? $this->server
 316                         : $wgDBserver;
 317         }
 318
 319         function reportPage() {
 320                 $this->pageCount++;
 321         }
 322
 323         function revCount() {
 324                 $this->revCount++;
 325                 $this->report();
 326         }
 327
 328         function report( $final = false ) {
 329                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 330                         $this->showReport();
 331                 }
 332         }
 333
 334         function showReport() {
 335                 if ( $this->reporting ) {
 336                         $now = wfTimestamp( TS_DB );
 337                         $nowts = microtime( true );
 338                         $deltaAll = $nowts - $this->startTime;
 339                         $deltaPart = $nowts - $this->lastTime;
 340                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 341                         $this->revCountPart = $this->revCount - $this->revCountLast;
 342
 343                         if ( $deltaAll ) {
 344                                 $portion = $this->revCount / $this->maxCount;
 345                                 $eta = $this->startTime + $deltaAll / $portion;
 346                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 347                                 $pageRate = $this->pageCount / $deltaAll;
 348                                 $revRate = $this->revCount / $deltaAll;
 349                         } else {
 350                                 $pageRate = '-';
 351                                 $revRate = '-';
 352                                 $etats = '-';
 353                         }
 354                         if ( $deltaPart ) {
 355                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 356                                 $revRatePart = $this->revCountPart / $deltaPart;
 357                         } else {
 358                                 $pageRatePart = '-';
 359                                 $revRatePart = '-';
 360                         }
 361                         $this->progress( sprintf(
 362                                 "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
 363                                         . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 364                                 $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate,
 365                                 $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
 366                                 $this->maxCount
 367                         ) );
 368                         $this->lastTime = $nowts;
 369                         $this->revCountLast = $this->revCount;
 370                 }
 371         }
 372
 373         function progress( $string ) {
 374                 fwrite( $this->stderr, $string . "\n" );
 375         }
 376
 377         function fatalError( $msg ) {
 378                 $this->progress( "$msg\n" );
 379                 die( 1 );
 380         }
 381 }
 382
 383 class ExportProgressFilter extends DumpFilter {
 384         function __construct( &$sink, &$progress ) {
 385                 parent::__construct( $sink );
 386                 $this->progress = $progress;
 387         }
 388
 389         function writeClosePage( $string ) {
 390                 parent::writeClosePage( $string );
 391                 $this->progress->reportPage();
 392         }
 393
 394         function writeRevision( $rev, $string ) {
 395                 parent::writeRevision( $rev, $string );
 396                 $this->progress->revCount();
 397         }
 398 }