maintenance/storage/compressOld.php

   1 <?php
   2 /**
   3  * Compress the text of a wiki.
   4  *
   5  * Usage:
   6  *
   7  * Non-wikimedia
   8  * php compressOld.php [options...]
   9  *
  10  * Wikimedia
  11  * php compressOld.php <database> [options...]
  12  *
  13  * Options are:
  14  *  -t <type>           set compression type to either:
  15  *                          gzip: compress revisions independently
  16  *                          concat: concatenate revisions and compress in chunks (default)
  17  *  -c <chunk-size>     maximum number of revisions in a concat chunk
  18  *  -b <begin-date>     earliest date to check for uncompressed revisions
  19  *  -e <end-date>       latest revision date to compress
  20  *  -s <startid>        the old_id to start from
  21  *  -n <endid>          the old_id to stop at
  22  *  --extdb <cluster>   store specified revisions in an external cluster (untested)
  23  *
  24  * This program is free software; you can redistribute it and/or modify
  25  * it under the terms of the GNU General Public License as published by
  26  * the Free Software Foundation; either version 2 of the License, or
  27  * (at your option) any later version.
  28  *
  29  * This program is distributed in the hope that it will be useful,
  30  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  31  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  32  * GNU General Public License for more details.
  33  *
  34  * You should have received a copy of the GNU General Public License along
  35  * with this program; if not, write to the Free Software Foundation, Inc.,
  36  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  37  * http://www.gnu.org/copyleft/gpl.html
  38  *
  39  * @file
  40  * @ingroup Maintenance ExternalStorage
  41  */
  42
  43 require_once( dirname( __FILE__ ) . '/../Maintenance.php' );
  44
  45 class CompressOld extends Maintenance {
  46         /**
  47          * @todo document
  48          */
  49         const LS_INDIVIDUAL = 0;
  50         const LS_CHUNKED    = 1;
  51
  52         public function __construct() {
  53                 parent::__construct();
  54                 $this->mDescription = 'Compress the text of a wiki';
  55                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  56                 $this->addOption( 'chunksize', 'Maximum number of revisions in a concat chunk', false, true, 'c' );
  57                 $this->addOption( 'begin-date', 'Earliest date to check for uncompressed revisions', false, true, 'b' );
  58                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  59                 $this->addOption( 'startid', 'The old_id to start from', false, true, 's' );
  60                 $this->addOption( 'extdb', 'Store specified revisions in an external cluster (untested)', false, true );
  61                 $this->addOption( 'endid', 'Stop at this old_id', false, true, 'n' );
  62         }
  63
  64         public function execute() {
  65                 global $wgDBname;
  66                 if ( !function_exists( "gzdeflate" ) ) {
  67                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
  68                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
  69                 }
  70
  71                 $type = $this->getOption( 'type', 'concat' );
  72                 $chunkSize = $this->getOption( 'chunksize', 20 );
  73                 $startId = $this->getOption( 'start-id', 0 );
  74                 $beginDate = $this->getOption( 'begin-date', '' );
  75                 $endDate = $this->getOption( 'end-date', '' );
  76                 $extDB = $this->getOption( 'extdb', '' );
  77                 $endId = $this->getOption( 'endid', false );
  78
  79                 if ( $type != 'concat' && $type != 'gzip' ) {
  80                         $this->error( "Type \"{$type}\" not supported" );
  81                 }
  82
  83                 if ( $extDB != '' ) {
  84                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
  85                                 . str_repeat( '-', 76 ) . "\n\n" );
  86                 } else {
  87                         $this->output( "Compressing database {$wgDBname}\n"
  88                                 . str_repeat( '-', 76 ) . "\n\n" );
  89                 }
  90
  91                 $success = true;
  92                 if ( $type == 'concat' ) {
  93                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
  94                                 $endDate, $extDB, $endId );
  95                 } else {
  96                         $this->compressOldPages( $startId, $extDB );
  97                 }
  98
  99                 if ( $success ) {
 100                         $this->output( "Done.\n" );
 101                 }
 102         }
 103
 104         /** @todo document */
 105         private function compressOldPages( $start = 0, $extdb = '' ) {
 106                 $chunksize = 50;
 107                 $this->output( "Starting from old_id $start...\n" );
 108                 $dbw = wfGetDB( DB_MASTER );
 109                 do {
 110                         $res = $dbw->select( 'text', array( 'old_id','old_flags','old_text' ),
 111                                 "old_id>=$start", __METHOD__, array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ) );
 112                         if( $dbw->numRows( $res ) == 0 ) {
 113                                 break;
 114                         }
 115                         $last = $start;
 116                         foreach ( $res as $row ) {
 117                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 118                                 $this->compressPage( $row, $extdb );
 119                                 $last = $row->old_id;
 120                         }
 121                         $start = $last + 1; # Deletion may leave long empty stretches
 122                         $this->output( "$start...\n" );
 123                 } while( true );
 124         }
 125
 126         /** @todo document */
 127         private function compressPage( $row, $extdb ) {
 128                 if ( false !== strpos( $row->old_flags, 'gzip' ) || false !== strpos( $row->old_flags, 'object' ) ) {
 129                         #print "Already compressed row {$row->old_id}\n";
 130                         return false;
 131                 }
 132                 $dbw = wfGetDB( DB_MASTER );
 133                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 134                 $compress = gzdeflate( $row->old_text );
 135
 136                 # Store in external storage if required
 137                 if ( $extdb !== '' ) {
 138                         $storeObj = new ExternalStoreDB;
 139                         $compress = $storeObj->store( $extdb, $compress );
 140                         if ( $compress === false ) {
 141                                 $this->error( "Unable to store object" );
 142                                 return false;
 143                         }
 144                 }
 145
 146                 # Update text row
 147                 $dbw->update( 'text',
 148                         array( /* SET */
 149                                 'old_flags' => $flags,
 150                                 'old_text' => $compress
 151                         ), array( /* WHERE */
 152                                 'old_id' => $row->old_id
 153                         ), __METHOD__,
 154                         array( 'LIMIT' => 1 )
 155                 );
 156                 return true;
 157         }
 158
 159         /** @todo document */
 160         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 161                 $endDate, $extdb = "", $maxPageId = false )
 162         {
 163                 $loadStyle = self::LS_CHUNKED;
 164
 165                 $dbr = wfGetDB( DB_SLAVE );
 166                 $dbw = wfGetDB( DB_MASTER );
 167
 168                 # Set up external storage
 169                 if ( $extdb != '' ) {
 170                         $storeObj = new ExternalStoreDB;
 171                 }
 172
 173                 # Get all articles by page_id
 174                 if ( !$maxPageId ) {
 175                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 176                 }
 177                 $this->output( "Starting from $startId of $maxPageId\n" );
 178                 $pageConds = array();
 179
 180                 /*
 181                 if ( $exclude_ns0 ) {
 182                         print "Excluding main namespace\n";
 183                         $pageConds[] = 'page_namespace<>0';
 184                 }
 185                 if ( $queryExtra ) {
 186                                         $pageConds[] = $queryExtra;
 187                 }
 188                  */
 189
 190                 # For each article, get a list of revisions which fit the criteria
 191
 192                 # No recompression, use a condition on old_flags
 193                 # Don't compress object type entities, because that might produce data loss when
 194                 # overwriting bulk storage concat rows. Don't compress external references, because
 195                 # the script doesn't yet delete rows from external storage.
 196                 $conds = array(
 197                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() ) . ' AND old_flags NOT '
 198                                 . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ) );
 199
 200                 if ( $beginDate ) {
 201                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 202                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 203                                 return false;
 204                         }
 205                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 206                 }
 207                 if ( $endDate )  {
 208                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 209                                 $this->error( "Invalid end date \"$endDate\"\n" );
 210                                 return false;
 211                         }
 212                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 213                 }
 214                 if ( $loadStyle == self::LS_CHUNKED ) {
 215                         $tables = array( 'revision', 'text' );
 216                         $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 217                         $conds[] = 'rev_text_id=old_id';
 218                         $revLoadOptions = 'FOR UPDATE';
 219                 } else {
 220                         $tables = array( 'revision' );
 221                         $fields = array( 'rev_id', 'rev_text_id' );
 222                         $revLoadOptions = array();
 223                 }
 224
 225                 # Don't work with current revisions
 226                 # Don't lock the page table for update either -- TS 2006-04-04
 227                 #$tables[] = 'page';
 228                 #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 229
 230                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 231                         wfWaitForSlaves();
 232
 233                         # Wake up
 234                         $dbr->ping();
 235
 236                         # Get the page row
 237                         $pageRes = $dbr->select( 'page',
 238                                 array('page_id', 'page_namespace', 'page_title','page_latest'),
 239                                 $pageConds + array('page_id' => $pageId), __METHOD__ );
 240                         if ( $dbr->numRows( $pageRes ) == 0 ) {
 241                                 continue;
 242                         }
 243                         $pageRow = $dbr->fetchObject( $pageRes );
 244
 245                         # Display progress
 246                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 247                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 248
 249                         # Load revisions
 250                         $revRes = $dbw->select( $tables, $fields,
 251                                 array_merge( array(
 252                                         'rev_page' => $pageRow->page_id,
 253                                         # Don't operate on the current revision
 254                                         # Use < instead of <> in case the current revision has changed
 255                                         # since the page select, which wasn't locking
 256                                         'rev_id < ' . $pageRow->page_latest
 257                                 ), $conds ),
 258                                 __METHOD__,
 259                                 $revLoadOptions
 260                         );
 261                         $revs = array();
 262                         foreach ( $revRes as $revRow ) {
 263                                 $revs[] = $revRow;
 264                         }
 265
 266                         if ( count( $revs ) < 2) {
 267                                 # No revisions matching, no further processing
 268                                 $this->output( "\n" );
 269                                 continue;
 270                         }
 271
 272                         # For each chunk
 273                         $i = 0;
 274                         while ( $i < count( $revs ) ) {
 275                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 276                                         $thisChunkSize = $maxChunkSize;
 277                                 } else {
 278                                         $thisChunkSize = count( $revs ) - $i;
 279                                 }
 280
 281                                 $chunk = new ConcatenatedGzipHistoryBlob();
 282                                 $stubs = array();
 283                                 $dbw->begin();
 284                                 $usedChunk = false;
 285                                 $primaryOldid = $revs[$i]->rev_text_id;
 286
 287                                 # Get the text of each revision and add it to the object
 288                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 289                                         $oldid = $revs[$i + $j]->rev_text_id;
 290
 291                                         # Get text
 292                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 293                                                 $textRow = $dbw->selectRow( 'text',
 294                                                         array( 'old_flags', 'old_text' ),
 295                                                         array( 'old_id' => $oldid ),
 296                                                         __METHOD__,
 297                                                         'FOR UPDATE'
 298                                                 );
 299                                                 $text = Revision::getRevisionText( $textRow );
 300                                         } else {
 301                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 302                                         }
 303
 304                                         if ( $text === false ) {
 305                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 306                                                 #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 307                                         }
 308
 309                                         if ( $extdb == "" && $j == 0 ) {
 310                                                 $chunk->setText( $text );
 311                                                 $this->output( '.' );
 312                                         } else {
 313                                                 # Don't make a stub if it's going to be longer than the article
 314                                                 # Stubs are typically about 100 bytes
 315                                                 if ( strlen( $text ) < 120 ) {
 316                                                         $stub = false;
 317                                                         $this->output( 'x' );
 318                                                 } else {
 319                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 320                                                         $stub->setLocation( $primaryOldid );
 321                                                         $stub->setReferrer( $oldid );
 322                                                         $this->output( '.' );
 323                                                         $usedChunk = true;
 324                                                 }
 325                                                 $stubs[$j] = $stub;
 326                                         }
 327                                 }
 328                                 $thisChunkSize = $j;
 329
 330                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 331                                 if ( $usedChunk ) {
 332                                         if ( $extdb != "" ) {
 333                                                 # Move blob objects to External Storage
 334                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ));
 335                                                 if ($stored === false) {
 336                                                         $this->error(  "Unable to store object" );
 337                                                         return false;
 338                                                 }
 339                                                 # Store External Storage URLs instead of Stub placeholders
 340                                                 foreach ($stubs as $stub) {
 341                                                         if ($stub===false)
 342                                                                 continue;
 343                                                         # $stored should provide base path to a BLOB
 344                                                         $url = $stored."/".$stub->getHash();
 345                                                         $dbw->update( 'text',
 346                                                                 array( /* SET */
 347                                                                         'old_text' => $url,
 348                                                                         'old_flags' => 'external,utf-8',
 349                                                                 ), array ( /* WHERE */
 350                                                                         'old_id' => $stub->getReferrer(),
 351                                                                 )
 352                                                         );
 353                                                 }
 354                                         } else {
 355                                                 # Store the main object locally
 356                                                 $dbw->update( 'text',
 357                                                         array( /* SET */
 358                                                                 'old_text' => serialize( $chunk ),
 359                                                                 'old_flags' => 'object,utf-8',
 360                                                         ), array( /* WHERE */
 361                                                                 'old_id' => $primaryOldid
 362                                                         )
 363                                                 );
 364
 365                                                 # Store the stub objects
 366                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 367                                                         # Skip if not compressing and don't overwrite the first revision
 368                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 369                                                                 $dbw->update( 'text',
 370                                                                         array( /* SET */
 371                                                                                 'old_text' => serialize($stubs[$j]),
 372                                                                                 'old_flags' => 'object,utf-8',
 373                                                                         ), array( /* WHERE */
 374                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 375                                                                         )
 376                                                                 );
 377                                                         }
 378                                                 }
 379                                         }
 380                                 }
 381                                 # Done, next
 382                                 $this->output( "/" );
 383                                 $dbw->commit();
 384                                 $i += $thisChunkSize;
 385                                 wfWaitForSlaves();
 386                         }
 387                         $this->output( "\n" );
 388                 }
 389                 return true;
 390         }
 391
 392 }
 393
 394 $maintClass = 'CompressOld';
 395 require_once( RUN_MAINTENANCE_IF_MAIN );