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                                         default:
 178                                                 $this->processOption( $opt, $val, $param );
 179                                 }
 180                         }
 181                 }
 182
 183                 if ( is_null( $sink ) ) {
 184                         $sink = new DumpOutput();
 185                 }
 186                 $sinks[] = $sink;
 187
 188                 if ( count( $sinks ) > 1 ) {
 189                         return new DumpMultiWriter( $sinks );
 190                 } else {
 191                         return $sink;
 192                 }
 193         }
 194
 195         function processOption( $opt, $val, $param ) {
 196                 // extension point for subclasses to add options
 197         }
 198
 199         function dump( $history, $text = WikiExporter::TEXT ) {
 200                 # Notice messages will foul up your XML output even if they're
 201                 # relatively harmless.
 202                 if ( ini_get( 'display_errors' ) ) {
 203                         ini_set( 'display_errors', 'stderr' );
 204                 }
 205
 206                 $this->initProgress( $history );
 207
 208                 $db = $this->backupDb();
 209                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 210                 $exporter->dumpUploads = $this->dumpUploads;
 211                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 212
 213                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 214                 $exporter->setOutputSink( $wrapper );
 215
 216                 if ( !$this->skipHeader ) {
 217                         $exporter->openStream();
 218                 }
 219                 # Log item dumps: all or by range
 220                 if ( $history & WikiExporter::LOGS ) {
 221                         if ( $this->startId || $this->endId ) {
 222                                 $exporter->logsByRange( $this->startId, $this->endId );
 223                         } else {
 224                                 $exporter->allLogs();
 225                         }
 226                 } elseif ( is_null( $this->pages ) ) {
 227                         # Page dumps: all or by page ID range
 228                         if ( $this->startId || $this->endId ) {
 229                                 $exporter->pagesByRange( $this->startId, $this->endId );
 230                         } elseif ( $this->revStartId || $this->revEndId ) {
 231                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 232                         } else {
 233                                 $exporter->allPages();
 234                         }
 235                 } else {
 236                         # Dump of specific pages
 237                         $exporter->pagesByName( $this->pages );
 238                 }
 239
 240                 if ( !$this->skipFooter ) {
 241                         $exporter->closeStream();
 242                 }
 243
 244                 $this->report( true );
 245         }
 246
 247         /**
 248          * Initialise starting time and maximum revision count.
 249          * We'll make ETA calculations based an progress, assuming relatively
 250          * constant per-revision rate.
 251          * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
 252          */
 253         function initProgress( $history = WikiExporter::FULL ) {
 254                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 255                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 256
 257                 $dbr = $this->forcedDb;
 258                 if ( $this->forcedDb === null ) {
 259                         $dbr = wfGetDB( DB_SLAVE );
 260                 }
 261                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 262                 $this->startTime = microtime( true );
 263                 $this->lastTime = $this->startTime;
 264                 $this->ID = getmypid();
 265         }
 266
 267         /**
 268          * @todo Fixme: the --server parameter is currently not respected, as it
 269          * doesn't seem terribly easy to ask the load balancer for a particular
 270          * connection by name.
 271          * @return DatabaseBase
 272          */
 273         function backupDb() {
 274                 if ( $this->forcedDb !== null ) {
 275                         return $this->forcedDb;
 276                 }
 277
 278                 $this->lb = wfGetLBFactory()->newMainLB();
 279                 $db = $this->lb->getConnection( DB_SLAVE, 'dump' );
 280
 281                 // Discourage the server from disconnecting us if it takes a long time
 282                 // to read out the big ol' batch query.
 283                 $db->setSessionOptions( array( 'connTimeout' => 3600 * 24 ) );
 284
 285                 return $db;
 286         }
 287
 288         /**
 289          * Force the dump to use the provided database connection for database
 290          * operations, wherever possible.
 291          *
 292          * @param DatabaseBase|null $db (Optional) the database connection to use. If null, resort to
 293          *   use the globally provided ways to get database connections.
 294          */
 295         function setDb( DatabaseBase $db = null ) {
 296                 $this->forcedDb = $db;
 297         }
 298
 299         function __destruct() {
 300                 if ( isset( $this->lb ) ) {
 301                         $this->lb->closeAll();
 302                 }
 303         }
 304
 305         function backupServer() {
 306                 global $wgDBserver;
 307
 308                 return $this->server
 309                         ? $this->server
 310                         : $wgDBserver;
 311         }
 312
 313         function reportPage() {
 314                 $this->pageCount++;
 315         }
 316
 317         function revCount() {
 318                 $this->revCount++;
 319                 $this->report();
 320         }
 321
 322         function report( $final = false ) {
 323                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 324                         $this->showReport();
 325                 }
 326         }
 327
 328         function showReport() {
 329                 if ( $this->reporting ) {
 330                         $now = wfTimestamp( TS_DB );
 331                         $nowts = microtime( true );
 332                         $deltaAll = $nowts - $this->startTime;
 333                         $deltaPart = $nowts - $this->lastTime;
 334                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 335                         $this->revCountPart = $this->revCount - $this->revCountLast;
 336
 337                         if ( $deltaAll ) {
 338                                 $portion = $this->revCount / $this->maxCount;
 339                                 $eta = $this->startTime + $deltaAll / $portion;
 340                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 341                                 $pageRate = $this->pageCount / $deltaAll;
 342                                 $revRate = $this->revCount / $deltaAll;
 343                         } else {
 344                                 $pageRate = '-';
 345                                 $revRate = '-';
 346                                 $etats = '-';
 347                         }
 348                         if ( $deltaPart ) {
 349                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 350                                 $revRatePart = $this->revCountPart / $deltaPart;
 351                         } else {
 352                                 $pageRatePart = '-';
 353                                 $revRatePart = '-';
 354                         }
 355                         $this->progress( sprintf(
 356                                 "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
 357                                         . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 358                                 $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate,
 359                                 $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
 360                                 $this->maxCount
 361                         ) );
 362                         $this->lastTime = $nowts;
 363                         $this->revCountLast = $this->revCount;
 364                 }
 365         }
 366
 367         function progress( $string ) {
 368                 fwrite( $this->stderr, $string . "\n" );
 369         }
 370
 371         function fatalError( $msg ) {
 372                 $this->progress( "$msg\n" );
 373                 die( 1 );
 374         }
 375 }
 376
 377 class ExportProgressFilter extends DumpFilter {
 378         function __construct( &$sink, &$progress ) {
 379                 parent::__construct( $sink );
 380                 $this->progress = $progress;
 381         }
 382
 383         function writeClosePage( $string ) {
 384                 parent::writeClosePage( $string );
 385                 $this->progress->reportPage();
 386         }
 387
 388         function writeRevision( $rev, $string ) {
 389                 parent::writeRevision( $rev, $string );
 390                 $this->progress->revCount();
 391         }
 392 }