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 id to start from (referring to the text table for
  21  *                      type gzip, and to the page table for type concat)
  22  *  -n <endid>          the page_id to stop at (only when using concat compression type)
  23  *  --extdb <cluster>   store specified revisions in an external cluster (untested)
  24  *
  25  * This program is free software; you can redistribute it and/or modify
  26  * it under the terms of the GNU General Public License as published by
  27  * the Free Software Foundation; either version 2 of the License, or
  28  * (at your option) any later version.
  29  *
  30  * This program is distributed in the hope that it will be useful,
  31  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  32  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  33  * GNU General Public License for more details.
  34  *
  35  * You should have received a copy of the GNU General Public License along
  36  * with this program; if not, write to the Free Software Foundation, Inc.,
  37  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  38  * http://www.gnu.org/copyleft/gpl.html
  39  *
  40  * @file
  41  * @ingroup Maintenance ExternalStorage
  42  */
  43
  44 require_once __DIR__ . '/../Maintenance.php';
  45
  46 /**
  47  * Maintenance script that compress the text of a wiki.
  48  *
  49  * @ingroup Maintenance ExternalStorage
  50  */
  51 class CompressOld extends Maintenance {
  52         /**
  53          * Option to load each revision individually.
  54          *
  55          */
  56         const LS_INDIVIDUAL = 0;
  57
  58         /**
  59          * Option to load revisions in chunks.
  60          *
  61          */
  62         const LS_CHUNKED = 1;
  63
  64         public function __construct() {
  65                 parent::__construct();
  66                 $this->addDescription( 'Compress the text of a wiki' );
  67                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  68                 $this->addOption(
  69                         'chunksize',
  70                         'Maximum number of revisions in a concat chunk',
  71                         false,
  72                         true,
  73                         'c'
  74                 );
  75                 $this->addOption(
  76                         'begin-date',
  77                         'Earliest date to check for uncompressed revisions',
  78                         false,
  79                         true,
  80                         'b'
  81                 );
  82                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  83                 $this->addOption(
  84                         'startid',
  85                         'The id to start from (gzip -> text table, concat -> page table)',
  86                         false,
  87                         true,
  88                         's'
  89                 );
  90                 $this->addOption(
  91                         'extdb',
  92                         'Store specified revisions in an external cluster (untested)',
  93                         false,
  94                         true
  95                 );
  96                 $this->addOption(
  97                         'endid',
  98                         'The page_id to stop at (only when using concat compression type)',
  99                         false,
 100                         true,
 101                         'n'
 102                 );
 103         }
 104
 105         public function execute() {
 106                 global $wgDBname;
 107                 if ( !function_exists( "gzdeflate" ) ) {
 108                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
 109                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
 110                 }
 111
 112                 $type = $this->getOption( 'type', 'concat' );
 113                 $chunkSize = $this->getOption( 'chunksize', 20 );
 114                 $startId = $this->getOption( 'startid', 0 );
 115                 $beginDate = $this->getOption( 'begin-date', '' );
 116                 $endDate = $this->getOption( 'end-date', '' );
 117                 $extDB = $this->getOption( 'extdb', '' );
 118                 $endId = $this->getOption( 'endid', false );
 119
 120                 if ( $type != 'concat' && $type != 'gzip' ) {
 121                         $this->error( "Type \"{$type}\" not supported" );
 122                 }
 123
 124                 if ( $extDB != '' ) {
 125                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
 126                                 . str_repeat( '-', 76 ) . "\n\n" );
 127                 } else {
 128                         $this->output( "Compressing database {$wgDBname}\n"
 129                                 . str_repeat( '-', 76 ) . "\n\n" );
 130                 }
 131
 132                 $success = true;
 133                 if ( $type == 'concat' ) {
 134                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
 135                                 $endDate, $extDB, $endId );
 136                 } else {
 137                         $this->compressOldPages( $startId, $extDB );
 138                 }
 139
 140                 if ( $success ) {
 141                         $this->output( "Done.\n" );
 142                 }
 143         }
 144
 145         /**
 146          * Fetch the text row-by-row to 'compressPage' function for compression.
 147          *
 148          * @param int $start
 149          * @param string $extdb
 150          */
 151         private function compressOldPages( $start = 0, $extdb = '' ) {
 152                 $chunksize = 50;
 153                 $this->output( "Starting from old_id $start...\n" );
 154                 $dbw = $this->getDB( DB_MASTER );
 155                 do {
 156                         $res = $dbw->select(
 157                                 'text',
 158                                 [ 'old_id', 'old_flags', 'old_text' ],
 159                                 "old_id>=$start",
 160                                 __METHOD__,
 161                                 [ 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ]
 162                         );
 163
 164                         if ( $res->numRows() == 0 ) {
 165                                 break;
 166                         }
 167
 168                         $last = $start;
 169
 170                         foreach ( $res as $row ) {
 171                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 172                                 $this->compressPage( $row, $extdb );
 173                                 $last = $row->old_id;
 174                         }
 175
 176                         $start = $last + 1; # Deletion may leave long empty stretches
 177                         $this->output( "$start...\n" );
 178                 } while ( true );
 179         }
 180
 181         /**
 182          * Compress the text in gzip format.
 183          *
 184          * @param stdClass $row
 185          * @param string $extdb
 186          * @return bool
 187          */
 188         private function compressPage( $row, $extdb ) {
 189                 if ( false !== strpos( $row->old_flags, 'gzip' )
 190                         || false !== strpos( $row->old_flags, 'object' )
 191                 ) {
 192                         # print "Already compressed row {$row->old_id}\n";
 193                         return false;
 194                 }
 195                 $dbw = $this->getDB( DB_MASTER );
 196                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 197                 $compress = gzdeflate( $row->old_text );
 198
 199                 # Store in external storage if required
 200                 if ( $extdb !== '' ) {
 201                         $storeObj = new ExternalStoreDB;
 202                         $compress = $storeObj->store( $extdb, $compress );
 203                         if ( $compress === false ) {
 204                                 $this->error( "Unable to store object" );
 205
 206                                 return false;
 207                         }
 208                 }
 209
 210                 # Update text row
 211                 $dbw->update( 'text',
 212                         [ /* SET */
 213                                 'old_flags' => $flags,
 214                                 'old_text' => $compress
 215                         ], [ /* WHERE */
 216                                 'old_id' => $row->old_id
 217                         ], __METHOD__,
 218                         [ 'LIMIT' => 1 ]
 219                 );
 220
 221                 return true;
 222         }
 223
 224         /**
 225          * Compress the text in chunks after concatenating the revisions.
 226          *
 227          * @param int $startId
 228          * @param int $maxChunkSize
 229          * @param string $beginDate
 230          * @param string $endDate
 231          * @param string $extdb
 232          * @param bool|int $maxPageId
 233          * @return bool
 234          */
 235         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 236                 $endDate, $extdb = "", $maxPageId = false
 237         ) {
 238                 $loadStyle = self::LS_CHUNKED;
 239
 240                 $dbr = $this->getDB( DB_REPLICA );
 241                 $dbw = $this->getDB( DB_MASTER );
 242
 243                 # Set up external storage
 244                 if ( $extdb != '' ) {
 245                         $storeObj = new ExternalStoreDB;
 246                 }
 247
 248                 # Get all articles by page_id
 249                 if ( !$maxPageId ) {
 250                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 251                 }
 252                 $this->output( "Starting from $startId of $maxPageId\n" );
 253                 $pageConds = [];
 254
 255                 /*
 256                 if ( $exclude_ns0 ) {
 257                         print "Excluding main namespace\n";
 258                         $pageConds[] = 'page_namespace<>0';
 259                 }
 260                 if ( $queryExtra ) {
 261                                         $pageConds[] = $queryExtra;
 262                 }
 263                  */
 264
 265                 # For each article, get a list of revisions which fit the criteria
 266
 267                 # No recompression, use a condition on old_flags
 268                 # Don't compress object type entities, because that might produce data loss when
 269                 # overwriting bulk storage concat rows. Don't compress external references, because
 270                 # the script doesn't yet delete rows from external storage.
 271                 $conds = [
 272                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() )
 273                         . ' AND old_flags NOT '
 274                         . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() )
 275                 ];
 276
 277                 if ( $beginDate ) {
 278                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 279                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 280
 281                                 return false;
 282                         }
 283                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 284                 }
 285                 if ( $endDate ) {
 286                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 287                                 $this->error( "Invalid end date \"$endDate\"\n" );
 288
 289                                 return false;
 290                         }
 291                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 292                 }
 293                 if ( $loadStyle == self::LS_CHUNKED ) {
 294                         $tables = [ 'revision', 'text' ];
 295                         $fields = [ 'rev_id', 'rev_text_id', 'old_flags', 'old_text' ];
 296                         $conds[] = 'rev_text_id=old_id';
 297                         $revLoadOptions = 'FOR UPDATE';
 298                 } else {
 299                         $tables = [ 'revision' ];
 300                         $fields = [ 'rev_id', 'rev_text_id' ];
 301                         $revLoadOptions = [];
 302                 }
 303
 304                 # Don't work with current revisions
 305                 # Don't lock the page table for update either -- TS 2006-04-04
 306                 # $tables[] = 'page';
 307                 # $conds[] = 'page_id=rev_page AND rev_id != page_latest';
 308
 309                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 310                         wfWaitForSlaves();
 311
 312                         # Wake up
 313                         $dbr->ping();
 314
 315                         # Get the page row
 316                         $pageRes = $dbr->select( 'page',
 317                                 [ 'page_id', 'page_namespace', 'page_title', 'page_latest' ],
 318                                 $pageConds + [ 'page_id' => $pageId ], __METHOD__ );
 319                         if ( $pageRes->numRows() == 0 ) {
 320                                 continue;
 321                         }
 322                         $pageRow = $dbr->fetchObject( $pageRes );
 323
 324                         # Display progress
 325                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 326                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 327
 328                         # Load revisions
 329                         $revRes = $dbw->select( $tables, $fields,
 330                                 array_merge( [
 331                                         'rev_page' => $pageRow->page_id,
 332                                         # Don't operate on the current revision
 333                                         # Use < instead of <> in case the current revision has changed
 334                                         # since the page select, which wasn't locking
 335                                         'rev_id < ' . $pageRow->page_latest
 336                                 ], $conds ),
 337                                 __METHOD__,
 338                                 $revLoadOptions
 339                         );
 340                         $revs = [];
 341                         foreach ( $revRes as $revRow ) {
 342                                 $revs[] = $revRow;
 343                         }
 344
 345                         if ( count( $revs ) < 2 ) {
 346                                 # No revisions matching, no further processing
 347                                 $this->output( "\n" );
 348                                 continue;
 349                         }
 350
 351                         # For each chunk
 352                         $i = 0;
 353                         while ( $i < count( $revs ) ) {
 354                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 355                                         $thisChunkSize = $maxChunkSize;
 356                                 } else {
 357                                         $thisChunkSize = count( $revs ) - $i;
 358                                 }
 359
 360                                 $chunk = new ConcatenatedGzipHistoryBlob();
 361                                 $stubs = [];
 362                                 $this->beginTransaction( $dbw, __METHOD__ );
 363                                 $usedChunk = false;
 364                                 $primaryOldid = $revs[$i]->rev_text_id;
 365
 366                                 // @codingStandardsIgnoreStart Ignore avoid function calls in a FOR loop test part warning
 367                                 # Get the text of each revision and add it to the object
 368                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 369                                         // @codingStandardsIgnoreEnd
 370                                         $oldid = $revs[$i + $j]->rev_text_id;
 371
 372                                         # Get text
 373                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 374                                                 $textRow = $dbw->selectRow( 'text',
 375                                                         [ 'old_flags', 'old_text' ],
 376                                                         [ 'old_id' => $oldid ],
 377                                                         __METHOD__,
 378                                                         'FOR UPDATE'
 379                                                 );
 380                                                 $text = Revision::getRevisionText( $textRow );
 381                                         } else {
 382                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 383                                         }
 384
 385                                         if ( $text === false ) {
 386                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 387                                                 # $dbw->delete( 'old', [ 'old_id' => $oldid ] );
 388                                         }
 389
 390                                         if ( $extdb == "" && $j == 0 ) {
 391                                                 $chunk->setText( $text );
 392                                                 $this->output( '.' );
 393                                         } else {
 394                                                 # Don't make a stub if it's going to be longer than the article
 395                                                 # Stubs are typically about 100 bytes
 396                                                 if ( strlen( $text ) < 120 ) {
 397                                                         $stub = false;
 398                                                         $this->output( 'x' );
 399                                                 } else {
 400                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 401                                                         $stub->setLocation( $primaryOldid );
 402                                                         $stub->setReferrer( $oldid );
 403                                                         $this->output( '.' );
 404                                                         $usedChunk = true;
 405                                                 }
 406                                                 $stubs[$j] = $stub;
 407                                         }
 408                                 }
 409                                 $thisChunkSize = $j;
 410
 411                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 412                                 if ( $usedChunk ) {
 413                                         if ( $extdb != "" ) {
 414                                                 # Move blob objects to External Storage
 415                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ) );
 416                                                 if ( $stored === false ) {
 417                                                         $this->error( "Unable to store object" );
 418
 419                                                         return false;
 420                                                 }
 421                                                 # Store External Storage URLs instead of Stub placeholders
 422                                                 foreach ( $stubs as $stub ) {
 423                                                         if ( $stub === false ) {
 424                                                                 continue;
 425                                                         }
 426                                                         # $stored should provide base path to a BLOB
 427                                                         $url = $stored . "/" . $stub->getHash();
 428                                                         $dbw->update( 'text',
 429                                                                 [ /* SET */
 430                                                                         'old_text' => $url,
 431                                                                         'old_flags' => 'external,utf-8',
 432                                                                 ], [ /* WHERE */
 433                                                                         'old_id' => $stub->getReferrer(),
 434                                                                 ]
 435                                                         );
 436                                                 }
 437                                         } else {
 438                                                 # Store the main object locally
 439                                                 $dbw->update( 'text',
 440                                                         [ /* SET */
 441                                                                 'old_text' => serialize( $chunk ),
 442                                                                 'old_flags' => 'object,utf-8',
 443                                                         ], [ /* WHERE */
 444                                                                 'old_id' => $primaryOldid
 445                                                         ]
 446                                                 );
 447
 448                                                 # Store the stub objects
 449                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 450                                                         # Skip if not compressing and don't overwrite the first revision
 451                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 452                                                                 $dbw->update( 'text',
 453                                                                         [ /* SET */
 454                                                                                 'old_text' => serialize( $stubs[$j] ),
 455                                                                                 'old_flags' => 'object,utf-8',
 456                                                                         ], [ /* WHERE */
 457                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 458                                                                         ]
 459                                                                 );
 460                                                         }
 461                                                 }
 462                                         }
 463                                 }
 464                                 # Done, next
 465                                 $this->output( "/" );
 466                                 $this->commitTransaction( $dbw, __METHOD__ );
 467                                 $i += $thisChunkSize;
 468                                 wfWaitForSlaves();
 469                         }
 470                         $this->output( "\n" );
 471                 }
 472
 473                 return true;
 474         }
 475 }
 476
 477 $maintClass = 'CompressOld';
 478 require_once RUN_MAINTENANCE_IF_MAIN;