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
  55         function BackupDumper( $args ) {
  56                 $this->stderr = fopen( "php://stderr", "wt" );
  57
  58                 // Built-in output and filter plugins
  59                 $this->registerOutput( 'file', 'DumpFileOutput' );
  60                 $this->registerOutput( 'gzip', 'DumpGZipOutput' );
  61                 $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
  62                 $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
  63                 $this->registerOutput( '7zip', 'Dump7ZipOutput' );
  64
  65                 $this->registerFilter( 'latest', 'DumpLatestFilter' );
  66                 $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
  67                 $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
  68
  69                 $this->sink = $this->processArgs( $args );
  70         }
  71
  72         /**
  73          * @param $name String
  74          * @param $class String: name of output filter plugin class
  75          */
  76         function registerOutput( $name, $class ) {
  77                 $this->outputTypes[$name] = $class;
  78         }
  79
  80         /**
  81          * @param $name String
  82          * @param $class String: name of filter plugin class
  83          */
  84         function registerFilter( $name, $class ) {
  85                 $this->filterTypes[$name] = $class;
  86         }
  87
  88         /**
  89          * Load a plugin and register it
  90          *
  91          * @param $class String: name of plugin class; must have a static 'register'
  92          *               method that takes a BackupDumper as a parameter.
  93          * @param $file String: full or relative path to the PHP file to load, or empty
  94          */
  95         function loadPlugin( $class, $file ) {
  96                 if ( $file != '' ) {
  97                         require_once( $file );
  98                 }
  99                 $register = array( $class, 'register' );
 100                 call_user_func_array( $register, array( &$this ) );
 101         }
 102
 103         /**
 104          * @param $args Array
 105          * @return Array
 106          */
 107         function processArgs( $args ) {
 108                 $sink = null;
 109                 $sinks = array();
 110                 foreach ( $args as $arg ) {
 111                         $matches = array();
 112                         if ( preg_match( '/^--(.+?)(?:=(.+?)(?::(.+?))?)?$/', $arg, $matches ) ) {
 113                                 @list( /* $full */ , $opt, $val, $param ) = $matches;
 114                                 switch( $opt ) {
 115                                 case "plugin":
 116                                         $this->loadPlugin( $val, $param );
 117                                         break;
 118                                 case "output":
 119                                         if ( !is_null( $sink ) ) {
 120                                                 $sinks[] = $sink;
 121                                         }
 122                                         if ( !isset( $this->outputTypes[$val] ) ) {
 123                                                 $this->fatalError( "Unrecognized output sink type '$val'" );
 124                                         }
 125                                         $type = $this->outputTypes[$val];
 126                                         $sink = new $type( $param );
 127                                         break;
 128                                 case "filter":
 129                                         if ( is_null( $sink ) ) {
 130                                                 $sink = new DumpOutput();
 131                                         }
 132                                         if ( !isset( $this->filterTypes[$val] ) ) {
 133                                                 $this->fatalError( "Unrecognized filter type '$val'" );
 134                                         }
 135                                         $type = $this->filterTypes[$val];
 136                                         $filter = new $type( $sink, $param );
 137
 138                                         // references are lame in php...
 139                                         unset( $sink );
 140                                         $sink = $filter;
 141
 142                                         break;
 143                                 case "report":
 144                                         $this->reportingInterval = intval( $val );
 145                                         break;
 146                                 case "server":
 147                                         $this->server = $val;
 148                                         break;
 149                                 case "force-normal":
 150                                         if ( !function_exists( 'utf8_normalize' ) ) {
 151                                                 wfDl( "php_utfnormal.so" );
 152                                                 if ( !function_exists( 'utf8_normalize' ) ) {
 153                                                         $this->fatalError( "Failed to load UTF-8 normalization extension. " .
 154                                                                 "Install or remove --force-normal parameter to use slower code." );
 155                                                 }
 156                                         }
 157                                         break;
 158                                 default:
 159                                         $this->processOption( $opt, $val, $param );
 160                                 }
 161                         }
 162                 }
 163
 164                 if ( is_null( $sink ) ) {
 165                         $sink = new DumpOutput();
 166                 }
 167                 $sinks[] = $sink;
 168
 169                 if ( count( $sinks ) > 1 ) {
 170                         return new DumpMultiWriter( $sinks );
 171                 } else {
 172                         return $sink;
 173                 }
 174         }
 175
 176         function processOption( $opt, $val, $param ) {
 177                 // extension point for subclasses to add options
 178         }
 179
 180         function dump( $history, $text = WikiExporter::TEXT ) {
 181                 # Notice messages will foul up your XML output even if they're
 182                 # relatively harmless.
 183                 if ( ini_get( 'display_errors' ) )
 184                         ini_set( 'display_errors', 'stderr' );
 185
 186                 $this->initProgress( $history );
 187
 188                 $db = $this->backupDb();
 189                 $exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
 190                 $exporter->dumpUploads = $this->dumpUploads;
 191                 $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
 192
 193                 $wrapper = new ExportProgressFilter( $this->sink, $this );
 194                 $exporter->setOutputSink( $wrapper );
 195
 196                 if ( !$this->skipHeader )
 197                         $exporter->openStream();
 198                 # Log item dumps: all or by range
 199                 if ( $history & WikiExporter::LOGS ) {
 200                         if ( $this->startId || $this->endId ) {
 201                                 $exporter->logsByRange( $this->startId, $this->endId );
 202                         } else {
 203                                 $exporter->allLogs();
 204                         }
 205                 # Page dumps: all or by page ID range
 206                 } else if ( is_null( $this->pages ) ) {
 207                         if ( $this->startId || $this->endId ) {
 208                                 $exporter->pagesByRange( $this->startId, $this->endId );
 209                         } else {
 210                                 $exporter->allPages();
 211                         }
 212                 # Dump of specific pages
 213                 } else {
 214                         $exporter->pagesByName( $this->pages );
 215                 }
 216
 217                 if ( !$this->skipFooter )
 218                         $exporter->closeStream();
 219
 220                 $this->report( true );
 221         }
 222
 223         /**
 224          * Initialise starting time and maximum revision count.
 225          * We'll make ETA calculations based an progress, assuming relatively
 226          * constant per-revision rate.
 227          * @param $history Integer: WikiExporter::CURRENT or WikiExporter::FULL
 228          */
 229         function initProgress( $history = WikiExporter::FULL ) {
 230                 $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
 231                 $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 232
 233                 $dbr = wfGetDB( DB_SLAVE );
 234                 $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
 235                 $this->startTime = wfTime();
 236         }
 237
 238         /**
 239          * @todo Fixme: the --server parameter is currently not respected, as it
 240          * doesn't seem terribly easy to ask the load balancer for a particular
 241          * connection by name.
 242          */
 243         function backupDb() {
 244                 $this->lb = wfGetLBFactory()->newMainLB();
 245                 $db = $this->lb->getConnection( DB_SLAVE, 'backup' );
 246
 247                 // Discourage the server from disconnecting us if it takes a long time
 248                 // to read out the big ol' batch query.
 249                 $db->setTimeout( 3600 * 24 );
 250
 251                 return $db;
 252         }
 253
 254         function __destruct() {
 255                 if ( isset( $this->lb ) ) {
 256                         $this->lb->closeAll();
 257                 }
 258         }
 259
 260         function backupServer() {
 261                 global $wgDBserver;
 262                 return $this->server
 263                         ? $this->server
 264                         : $wgDBserver;
 265         }
 266
 267         function reportPage() {
 268                 $this->pageCount++;
 269         }
 270
 271         function revCount() {
 272                 $this->revCount++;
 273                 $this->report();
 274         }
 275
 276         function report( $final = false ) {
 277                 if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
 278                         $this->showReport();
 279                 }
 280         }
 281
 282         function showReport() {
 283                 if ( $this->reporting ) {
 284                         $delta = wfTime() - $this->startTime;
 285                         $now = wfTimestamp( TS_DB );
 286                         if ( $delta ) {
 287                                 $rate = $this->pageCount / $delta;
 288                                 $revrate = $this->revCount / $delta;
 289                                 $portion = $this->revCount / $this->maxCount;
 290                                 $eta = $this->startTime + $delta / $portion;
 291                                 $etats = wfTimestamp( TS_DB, intval( $eta ) );
 292                         } else {
 293                                 $rate = '-';
 294                                 $revrate = '-';
 295                                 $etats = '-';
 296                         }
 297                         $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), ETA %s [max %d]",
 298                                 $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $etats, $this->maxCount ) );
 299                 }
 300         }
 301
 302         function progress( $string ) {
 303                 fwrite( $this->stderr, $string . "\n" );
 304         }
 305
 306         function fatalError( $msg ) {
 307                 $this->progress( "$msg\n" );
 308                 die(1);
 309         }
 310 }
 311
 312 class ExportProgressFilter extends DumpFilter {
 313         function ExportProgressFilter( &$sink, &$progress ) {
 314                 parent::__construct( $sink );
 315                 $this->progress = $progress;
 316         }
 317
 318         function writeClosePage( $string ) {
 319                 parent::writeClosePage( $string );
 320                 $this->progress->reportPage();
 321         }
 322
 323         function writeRevision( $rev, $string ) {
 324                 parent::writeRevision( $rev, $string );
 325                 $this->progress->revCount();
 326         }
 327 }