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