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