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
  54         function BackupDumper( $args ) {
  55                 $this->stderr = fopen( "php://stderr", "wt" );
  56
  57                 // Built-in output and filter plugins
  58                 $this->registerOutput( 'file', 'DumpFileOutput' );
  59                 $this->registerOutput( 'gzip', 'DumpGZipOutput' );
  60                 $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
  61                 $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
  62                 $this->registerOutput( '7zip', 'Dump7ZipOutput' );
  63
  64                 $this->registerFilter( 'latest', 'DumpLatestFilter' );
  65                 $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
  66                 $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
  67
  68                 $this->sink = $this->processArgs( $args );
  69         }
  70
  71         /**
  72          * @param $name String
  73          * @param $class String: name of output filter plugin class
  74          */
  75         function registerOutput( $name, $class ) {
  76                 $this->outputTypes[$name] = $class;
  77         }
  78
  79         /**
  80          * @param $name String
  81          * @param $class String: name of filter plugin class
  82          */
  83         function registerFilter( $name, $class ) {
  84                 $this->filterTypes[$name] = $class;
  85         }
  86
  87         /**
  88          * Load a plugin and register it
  89          *
  90          * @param $class String: name of plugin class; must have a static 'register'
  91          *               method that takes a BackupDumper as a parameter.
  92          * @param $file String: full or relative path to the PHP file to load, or empty
  93          */
  94         function loadPlugin( $class, $file ) {
  95                 if ( $file != '' ) {
  96                         require_once( $file );
  97                 }
  98                 $register = array( $class, 'register' );
  99                 call_user_func_array( $register, array( &$this ) );
 100         }
 101
 102         /**
 103          * @param $args Array
 104          * @return Array
 105          */
 106         function processArgs( $args ) {
 107                 $sink = null;
 108                 $sinks = array();
 109                 foreach ( $args as $arg ) {
 110                         $matches = array();
 111                         if ( preg_match( '/^--(.+?)(?:=(.+?)(?::(.+?))?)?$/', $arg, $matches ) ) {
 112                                 @list( /* $full */ , $opt, $val, $param ) = $matches;
 113                                 switch( $opt ) {
 114                                 case "plugin":
 115                                         $this->loadPlugin( $val, $param );
 116                                         break;
 117                                 case "output":
 118                                         if ( !is_null( $sink ) ) {
 119                                                 $sinks[] = $sink;
 120                                         }
 121                                         if ( !isset( $this->outputTypes[$val] ) ) {
 122                                                 wfDie( "Unrecognized output sink type '$val'\n" );
 123                                         }
 124                                         $type = $this->outputTypes[$val];
 125                                         $sink = new $type( $param );
 126                                         break;
 127                                 case "filter":
 128                                         if ( is_null( $sink ) ) {
 129                                                 $sink = new DumpOutput();
 130                                         }
 131                                         if ( !isset( $this->filterTypes[$val] ) ) {
 132                                                 wfDie( "Unrecognized filter type '$val'\n" );
 133                                         }
 134                                         $type = $this->filterTypes[$val];
 135                                         $filter = new $type( $sink, $param );
 136
 137                                         // references are lame in php...
 138                                         unset( $sink );
 139                                         $sink = $filter;
 140
 141                                         break;
 142                                 case "report":
 143                                         $this->reportingInterval = intval( $val );
 144                                         break;
 145                                 case "server":
 146                                         $this->server = $val;
 147                                         break;
 148                                 case "force-normal":
 149                                         if ( !function_exists( 'utf8_normalize' ) ) {
 150                                                 wfDl( "php_utfnormal.so" );
 151                                                 if ( !function_exists( 'utf8_normalize' ) ) {
 152                                                         wfDie( "Failed to load UTF-8 normalization extension. " .
 153                                                                 "Install or remove --force-normal parameter to use slower code.\n" );
 154                                                 }
 155                                         }
 156                                         break;
 157                                 default:
 158                                         $this->processOption( $opt, $val, $param );
 159                                 }
 160                         }
 161                 }
 162
 163                 if ( is_null( $sink ) ) {
 164                         $sink = new DumpOutput();
 165                 }
 166                 $sinks[] = $sink;
 167
 168                 if ( count( $sinks ) > 1 ) {
 169                         return new DumpMultiWriter( $sinks );
 170                 } else {
 171                         return $sink;
 172                 }
 173         }
 174
 175         function processOption( $opt, $val, $param ) {
 176                 // extension point for subclasses to add options
 177         }
 178
 179         function dump( $history, $text = WikiExporter::TEXT ) {
 180                 # Notice messages will foul up your XML output even if they're
 181                 # relatively harmless.
 182                 if ( ini_get( 'display_errors' ) )
 183                         ini_set( 'display_errors', 'stderr' );
 184
 185                 $this->initProgress( $history );
 186
 187                 $db = $this->backupDb();
 188                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 189                 $exporter->dumpUploads = $this->dumpUploads;
 190
 191                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 192                 $exporter->setOutputSink( $wrapper );
 193
 194                 if ( !$this->skipHeader )
 195                         $exporter->openStream();
 196                 # Log item dumps: all or by range
 197                 if ( $history & WikiExporter::LOGS ) {
 198                         if ( $this->startId || $this->endId ) {
 199                                 $exporter->logsByRange( $this->startId, $this->endId );
 200                         } else {
 201                                 $exporter->allLogs();
 202                         }
 203                 # Page dumps: all or by page ID range
 204                 } else if ( is_null( $this->pages ) ) {
 205                         if ( $this->startId || $this->endId ) {
 206                                 $exporter->pagesByRange( $this->startId, $this->endId );
 207                         } else {
 208                                 $exporter->allPages();
 209                         }
 210                 # Dump of specific pages
 211                 } else {
 212                         $exporter->pagesByName( $this->pages );
 213                 }
 214
 215                 if ( !$this->skipFooter )
 216                         $exporter->closeStream();
 217
 218                 $this->report( true );
 219         }
 220
 221         /**
 222          * Initialise starting time and maximum revision count.
 223          * We'll make ETA calculations based an progress, assuming relatively
 224          * constant per-revision rate.
 225          * @param $history Integer: WikiExporter::CURRENT or WikiExporter::FULL
 226          */
 227         function initProgress( $history = WikiExporter::FULL ) {
 228                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 229                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 230
 231                 $dbr = wfGetDB( DB_SLAVE );
 232                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 233                 $this->startTime = wfTime();
 234         }
 235
 236         /**
 237          * @todo Fixme: the --server parameter is currently not respected, as it
 238          * doesn't seem terribly easy to ask the load balancer for a particular
 239          * connection by name.
 240          */
 241         function backupDb() {
 242                 $this->lb = wfGetLBFactory()->newMainLB();
 243                 $db = $this->lb->getConnection( DB_SLAVE, 'backup' );
 244
 245                 // Discourage the server from disconnecting us if it takes a long time
 246                 // to read out the big ol' batch query.
 247                 $db->setTimeout( 3600 * 24 );
 248
 249                 return $db;
 250         }
 251
 252         function __destruct() {
 253                 if ( isset( $this->lb ) ) {
 254                         $this->lb->closeAll();
 255                 }
 256         }
 257
 258         function backupServer() {
 259                 global $wgDBserver;
 260                 return $this->server
 261                         ? $this->server
 262                         : $wgDBserver;
 263         }
 264
 265         function reportPage() {
 266                 $this->pageCount++;
 267         }
 268
 269         function revCount() {
 270                 $this->revCount++;
 271                 $this->report();
 272         }
 273
 274         function report( $final = false ) {
 275                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 276                         $this->showReport();
 277                 }
 278         }
 279
 280         function showReport() {
 281                 if ( $this->reporting ) {
 282                         $delta = wfTime() - $this->startTime;
 283                         $now = wfTimestamp( TS_DB );
 284                         if ( $delta ) {
 285                                 $rate = $this->pageCount / $delta;
 286                                 $revrate = $this->revCount / $delta;
 287                                 $portion = $this->revCount / $this->maxCount;
 288                                 $eta = $this->startTime + $delta / $portion;
 289                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 290                         } else {
 291                                 $rate = '-';
 292                                 $revrate = '-';
 293                                 $etats = '-';
 294                         }
 295                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), ETA %s [max %d]",
 296                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $etats, $this->maxCount ) );
 297                 }
 298         }
 299
 300         function progress( $string ) {
 301                 fwrite( $this->stderr, $string . "\n" );
 302         }
 303 }
 304
 305 class ExportProgressFilter extends DumpFilter {
 306         function ExportProgressFilter( &$sink, &$progress ) {
 307                 parent::__construct( $sink );
 308                 $this->progress = $progress;
 309         }
 310
 311         function writeClosePage( $string ) {
 312                 parent::writeClosePage( $string );
 313                 $this->progress->reportPage();
 314         }
 315
 316         function writeRevision( $rev, $string ) {
 317                 parent::writeRevision( $rev, $string );
 318                 $this->progress->revCount();
 319         }
 320 }