maintenance/backup.inc

   1 <?php
   2 /**
   3  * Base classes for database dumpers
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   6  * http://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         var $reportingInterval = 100;
  41         var $reporting = true;
  42         var $pageCount = 0;
  43         var $revCount = 0;
  44         var $server = null; // use default
  45         var $pages = null; // all pages
  46         var $skipHeader = false; // don't output <mediawiki> and <siteinfo>
  47         var $skipFooter = false; // don't output </mediawiki>
  48         var $startId = 0;
  49         var $endId = 0;
  50         var $revStartId = 0;
  51         var $revEndId = 0;
  52         var $sink = null; // Output filters
  53         var $stubText = false; // include rev_text_id instead of text; for 2-pass dump
  54         var $dumpUploads = false;
  55         var $dumpUploadFileContents = false;
  56         var $lastTime = 0;
  57         var $pageCountLast = 0;
  58         var $revCountLast = 0;
  59         var $ID = 0;
  60
  61         var $outputTypes = array(), $filterTypes = array();
  62
  63         /**
  64          * The dependency-injected database to use.
  65          *
  66          * @var DatabaseBase|null
  67          *
  68          * @see self::setDb
  69          */
  70         protected $forcedDb = null;
  71
  72         /**
  73          * @var LoadBalancer
  74          */
  75         protected $lb;
  76
  77         function __construct( $args ) {
  78                 $this->stderr = fopen( "php://stderr", "wt" );
  79
  80                 // Built-in output and filter plugins
  81                 $this->registerOutput( 'file', 'DumpFileOutput' );
  82                 $this->registerOutput( 'gzip', 'DumpGZipOutput' );
  83                 $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
  84                 $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
  85                 $this->registerOutput( '7zip', 'Dump7ZipOutput' );
  86
  87                 $this->registerFilter( 'latest', 'DumpLatestFilter' );
  88                 $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
  89                 $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
  90
  91                 $this->sink = $this->processArgs( $args );
  92         }
  93
  94         /**
  95          * @param $name String
  96          * @param $class String: name of output filter plugin class
  97          */
  98         function registerOutput( $name, $class ) {
  99                 $this->outputTypes[$name] = $class;
 100         }
 101
 102         /**
 103          * @param $name String
 104          * @param $class String: name of filter plugin class
 105          */
 106         function registerFilter( $name, $class ) {
 107                 $this->filterTypes[$name] = $class;
 108         }
 109
 110         /**
 111          * Load a plugin and register it
 112          *
 113          * @param $class String: name of plugin class; must have a static 'register'
 114          *               method that takes a BackupDumper as a parameter.
 115          * @param $file String: full or relative path to the PHP file to load, or empty
 116          */
 117         function loadPlugin( $class, $file ) {
 118                 if ( $file != '' ) {
 119                         require_once $file;
 120                 }
 121                 $register = array( $class, 'register' );
 122                 call_user_func_array( $register, array( &$this ) );
 123         }
 124
 125         /**
 126          * @param $args Array
 127          * @return Array
 128          */
 129         function processArgs( $args ) {
 130                 $sink = null;
 131                 $sinks = array();
 132                 foreach ( $args as $arg ) {
 133                         $matches = array();
 134                         if ( preg_match( '/^--(.+?)(?:=(.+?)(?::(.+?))?)?$/', $arg, $matches ) ) {
 135                                 @list( /* $full */ , $opt, $val, $param ) = $matches;
 136                                 switch ( $opt ) {
 137                                 case "plugin":
 138                                         $this->loadPlugin( $val, $param );
 139                                         break;
 140                                 case "output":
 141                                         if ( !is_null( $sink ) ) {
 142                                                 $sinks[] = $sink;
 143                                         }
 144                                         if ( !isset( $this->outputTypes[$val] ) ) {
 145                                                 $this->fatalError( "Unrecognized output sink type '$val'" );
 146                                         }
 147                                         $type = $this->outputTypes[$val];
 148                                         $sink = new $type( $param );
 149                                         break;
 150                                 case "filter":
 151                                         if ( is_null( $sink ) ) {
 152                                                 $sink = new DumpOutput();
 153                                         }
 154                                         if ( !isset( $this->filterTypes[$val] ) ) {
 155                                                 $this->fatalError( "Unrecognized filter type '$val'" );
 156                                         }
 157                                         $type = $this->filterTypes[$val];
 158                                         $filter = new $type( $sink, $param );
 159
 160                                         // references are lame in php...
 161                                         unset( $sink );
 162                                         $sink = $filter;
 163
 164                                         break;
 165                                 case "report":
 166                                         $this->reportingInterval = intval( $val );
 167                                         break;
 168                                 case "server":
 169                                         $this->server = $val;
 170                                         break;
 171                                 case "force-normal":
 172                                         if ( !function_exists( 'utf8_normalize' ) ) {
 173                                                 wfDl( "php_utfnormal.so" );
 174                                                 if ( !function_exists( 'utf8_normalize' ) ) {
 175                                                         $this->fatalError( "Failed to load UTF-8 normalization extension. " .
 176                                                                 "Install or remove --force-normal parameter to use slower code." );
 177                                                 }
 178                                         }
 179                                         break;
 180                                 default:
 181                                         $this->processOption( $opt, $val, $param );
 182                                 }
 183                         }
 184                 }
 185
 186                 if ( is_null( $sink ) ) {
 187                         $sink = new DumpOutput();
 188                 }
 189                 $sinks[] = $sink;
 190
 191                 if ( count( $sinks ) > 1 ) {
 192                         return new DumpMultiWriter( $sinks );
 193                 } else {
 194                         return $sink;
 195                 }
 196         }
 197
 198         function processOption( $opt, $val, $param ) {
 199                 // extension point for subclasses to add options
 200         }
 201
 202         function dump( $history, $text = WikiExporter::TEXT ) {
 203                 # Notice messages will foul up your XML output even if they're
 204                 # relatively harmless.
 205                 if ( ini_get( 'display_errors' ) ) {
 206                         ini_set( 'display_errors', 'stderr' );
 207                 }
 208
 209                 $this->initProgress( $history );
 210
 211                 $db = $this->backupDb();
 212                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 213                 $exporter->dumpUploads = $this->dumpUploads;
 214                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 215
 216                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 217                 $exporter->setOutputSink( $wrapper );
 218
 219                 if ( !$this->skipHeader ) {
 220                         $exporter->openStream();
 221                 }
 222                 # Log item dumps: all or by range
 223                 if ( $history & WikiExporter::LOGS ) {
 224                         if ( $this->startId || $this->endId ) {
 225                                 $exporter->logsByRange( $this->startId, $this->endId );
 226                         } else {
 227                                 $exporter->allLogs();
 228                         }
 229                 # Page dumps: all or by page ID range
 230                 } elseif ( is_null( $this->pages ) ) {
 231                         if ( $this->startId || $this->endId ) {
 232                                 $exporter->pagesByRange( $this->startId, $this->endId );
 233                         } elseif ( $this->revStartId || $this->revEndId ) {
 234                                 $exporter->revsByRange( $this->revStartId, $this->revEndId );
 235                         } else {
 236                                 $exporter->allPages();
 237                         }
 238                 # Dump of specific pages
 239                 } else {
 240                         $exporter->pagesByName( $this->pages );
 241                 }
 242
 243                 if ( !$this->skipFooter ) {
 244                         $exporter->closeStream();
 245                 }
 246
 247                 $this->report( true );
 248         }
 249
 250         /**
 251          * Initialise starting time and maximum revision count.
 252          * We'll make ETA calculations based an progress, assuming relatively
 253          * constant per-revision rate.
 254          * @param $history Integer: WikiExporter::CURRENT or WikiExporter::FULL
 255          */
 256         function initProgress( $history = WikiExporter::FULL ) {
 257                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 258                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 259
 260                 $dbr = $this->forcedDb;
 261                 if ( $this->forcedDb === null ) {
 262                         $dbr = wfGetDB( DB_SLAVE );
 263                 }
 264                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 265                 $this->startTime = microtime( true );
 266                 $this->lastTime = $this->startTime;
 267                 $this->ID = getmypid();
 268         }
 269
 270         /**
 271          * @todo Fixme: the --server parameter is currently not respected, as it
 272          * doesn't seem terribly easy to ask the load balancer for a particular
 273          * connection by name.
 274          * @return DatabaseBase
 275          */
 276         function backupDb() {
 277                 if ( $this->forcedDb !== null ) {
 278                         return $this->forcedDb;
 279                 }
 280
 281                 $this->lb = wfGetLBFactory()->newMainLB();
 282                 $db = $this->lb->getConnection( DB_SLAVE, 'backup' );
 283
 284                 // Discourage the server from disconnecting us if it takes a long time
 285                 // to read out the big ol' batch query.
 286                 $db->setSessionOptions( array( 'connTimeout' => 3600 * 24 ) );
 287
 288                 return $db;
 289         }
 290
 291         /**
 292          * Force the dump to use the provided database connection for database
 293          * operations, wherever possible.
 294          *
 295          * @param $db DatabaseBase|null: (Optional) the database connection to
 296          *            use. If null, resort to use the globally provided ways to
 297          *            get database connections.
 298          */
 299         function setDb( DatabaseBase $db = null ) {
 300                 $this->forcedDb = $db;
 301         }
 302
 303         function __destruct() {
 304                 if ( isset( $this->lb ) ) {
 305                         $this->lb->closeAll();
 306                 }
 307         }
 308
 309         function backupServer() {
 310                 global $wgDBserver;
 311                 return $this->server
 312                         ? $this->server
 313                         : $wgDBserver;
 314         }
 315
 316         function reportPage() {
 317                 $this->pageCount++;
 318         }
 319
 320         function revCount() {
 321                 $this->revCount++;
 322                 $this->report();
 323         }
 324
 325         function report( $final = false ) {
 326                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 327                         $this->showReport();
 328                 }
 329         }
 330
 331         function showReport() {
 332                 if ( $this->reporting ) {
 333                         $now = wfTimestamp( TS_DB );
 334                         $nowts = microtime( true );
 335                         $deltaAll = $nowts - $this->startTime;
 336                         $deltaPart = $nowts - $this->lastTime;
 337                         $this->pageCountPart = $this->pageCount - $this->pageCountLast;
 338                         $this->revCountPart = $this->revCount - $this->revCountLast;
 339
 340                         if ( $deltaAll ) {
 341                                 $portion = $this->revCount / $this->maxCount;
 342                                 $eta = $this->startTime + $deltaAll / $portion;
 343                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 344                                 $pageRate = $this->pageCount / $deltaAll;
 345                                 $revRate = $this->revCount / $deltaAll;
 346                         } else {
 347                                 $pageRate = '-';
 348                                 $revRate = '-';
 349                                 $etats = '-';
 350                         }
 351                         if ( $deltaPart ) {
 352                                 $pageRatePart = $this->pageCountPart / $deltaPart;
 353                                 $revRatePart = $this->revCountPart / $deltaPart;
 354                         } else {
 355                                 $pageRatePart = '-';
 356                                 $revRatePart = '-';
 357                         }
 358                         $this->progress( sprintf( "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), %d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
 359                                         $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate, $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats, $this->maxCount ) );
 360                         $this->lastTime = $nowts;
 361                         $this->revCountLast = $this->revCount;
 362                 }
 363         }
 364
 365         function progress( $string ) {
 366                 fwrite( $this->stderr, $string . "\n" );
 367         }
 368
 369         function fatalError( $msg ) {
 370                 $this->progress( "$msg\n" );
 371                 die( 1 );
 372         }
 373 }
 374
 375 class ExportProgressFilter extends DumpFilter {
 376         function __construct( &$sink, &$progress ) {
 377                 parent::__construct( $sink );
 378                 $this->progress = $progress;
 379         }
 380
 381         function writeClosePage( $string ) {
 382                 parent::writeClosePage( $string );
 383                 $this->progress->reportPage();
 384         }
 385
 386         function writeRevision( $rev, $string ) {
 387                 parent::writeRevision( $rev, $string );
 388                 $this->progress->revCount();
 389         }
 390 }