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 use MediaWiki\MediaWikiServices;
  44 use MediaWiki\Revision\SlotRecord;
  45
  46 require_once __DIR__ . '/../Maintenance.php';
  47
  48 /**
  49  * Maintenance script that compress the text of a wiki.
  50  *
  51  * @ingroup Maintenance ExternalStorage
  52  */
  53 class CompressOld extends Maintenance {
  54         public function __construct() {
  55                 parent::__construct();
  56                 $this->addDescription( 'Compress the text of a wiki' );
  57                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  58                 $this->addOption(
  59                         'chunksize',
  60                         'Maximum number of revisions in a concat chunk',
  61                         false,
  62                         true,
  63                         'c'
  64                 );
  65                 $this->addOption(
  66                         'begin-date',
  67                         'Earliest date to check for uncompressed revisions',
  68                         false,
  69                         true,
  70                         'b'
  71                 );
  72                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  73                 $this->addOption(
  74                         'startid',
  75                         'The id to start from (gzip -> text table, concat -> page table)',
  76                         false,
  77                         true,
  78                         's'
  79                 );
  80                 $this->addOption(
  81                         'extdb',
  82                         'Store specified revisions in an external cluster (untested)',
  83                         false,
  84                         true
  85                 );
  86                 $this->addOption(
  87                         'endid',
  88                         'The page_id to stop at (only when using concat compression type)',
  89                         false,
  90                         true,
  91                         'n'
  92                 );
  93         }
  94
  95         public function execute() {
  96                 global $wgDBname;
  97                 if ( !function_exists( "gzdeflate" ) ) {
  98                         $this->fatalError( "You must enable zlib support in PHP to compress old revisions!\n" .
  99                                 "Please see https://www.php.net/manual/en/ref.zlib.php\n" );
 100                 }
 101
 102                 $type = $this->getOption( 'type', 'concat' );
 103                 $chunkSize = $this->getOption( 'chunksize', 20 );
 104                 $startId = $this->getOption( 'startid', 0 );
 105                 $beginDate = $this->getOption( 'begin-date', '' );
 106                 $endDate = $this->getOption( 'end-date', '' );
 107                 $extDB = $this->getOption( 'extdb', '' );
 108                 $endId = $this->getOption( 'endid', false );
 109
 110                 if ( $type != 'concat' && $type != 'gzip' ) {
 111                         $this->error( "Type \"{$type}\" not supported" );
 112                 }
 113
 114                 if ( $extDB != '' ) {
 115                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
 116                                 . str_repeat( '-', 76 ) . "\n\n" );
 117                 } else {
 118                         $this->output( "Compressing database {$wgDBname}\n"
 119                                 . str_repeat( '-', 76 ) . "\n\n" );
 120                 }
 121
 122                 $success = true;
 123                 if ( $type == 'concat' ) {
 124                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
 125                                 $endDate, $extDB, $endId );
 126                 } else {
 127                         $this->compressOldPages( $startId, $extDB );
 128                 }
 129
 130                 if ( $success ) {
 131                         $this->output( "Done.\n" );
 132                 }
 133         }
 134
 135         /**
 136          * Fetch the text row-by-row to 'compressPage' function for compression.
 137          *
 138          * @param int $start
 139          * @param string $extdb
 140          */
 141         private function compressOldPages( $start = 0, $extdb = '' ) {
 142                 $chunksize = 50;
 143                 $this->output( "Starting from old_id $start...\n" );
 144                 $dbw = $this->getDB( DB_MASTER );
 145                 do {
 146                         $res = $dbw->select(
 147                                 'text',
 148                                 [ 'old_id', 'old_flags', 'old_text' ],
 149                                 "old_id>=$start",
 150                                 __METHOD__,
 151                                 [ 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ]
 152                         );
 153
 154                         if ( $res->numRows() == 0 ) {
 155                                 break;
 156                         }
 157
 158                         $last = $start;
 159
 160                         foreach ( $res as $row ) {
 161                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 162                                 $this->compressPage( $row, $extdb );
 163                                 $last = $row->old_id;
 164                         }
 165
 166                         $start = $last + 1; # Deletion may leave long empty stretches
 167                         $this->output( "$start...\n" );
 168                 } while ( true );
 169         }
 170
 171         /**
 172          * Compress the text in gzip format.
 173          *
 174          * @param stdClass $row
 175          * @param string $extdb
 176          * @return bool
 177          */
 178         private function compressPage( $row, $extdb ) {
 179                 if ( strpos( $row->old_flags, 'gzip' ) !== false
 180                         || strpos( $row->old_flags, 'object' ) !== false
 181                 ) {
 182                         # print "Already compressed row {$row->old_id}\n";
 183                         return false;
 184                 }
 185                 $dbw = $this->getDB( DB_MASTER );
 186                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 187                 $compress = gzdeflate( $row->old_text );
 188
 189                 # Store in external storage if required
 190                 if ( $extdb !== '' ) {
 191                         $esFactory = MediaWikiServices::getInstance()->getExternalStoreFactory();
 192                         /** @var ExternalStoreDB $storeObj */
 193                         $storeObj = $esFactory->getStore( 'DB' );
 194                         $compress = $storeObj->store( $extdb, $compress );
 195                         if ( $compress === false ) {
 196                                 $this->error( "Unable to store object" );
 197
 198                                 return false;
 199                         }
 200                 }
 201
 202                 # Update text row
 203                 $dbw->update( 'text',
 204                         [ /* SET */
 205                                 'old_flags' => $flags,
 206                                 'old_text' => $compress
 207                         ], [ /* WHERE */
 208                                 'old_id' => $row->old_id
 209                         ], __METHOD__,
 210                         [ 'LIMIT' => 1 ]
 211                 );
 212
 213                 return true;
 214         }
 215
 216         /**
 217          * Compress the text in chunks after concatenating the revisions.
 218          *
 219          * @param int $startId
 220          * @param int $maxChunkSize
 221          * @param string $beginDate
 222          * @param string $endDate
 223          * @param string $extdb
 224          * @param bool|int $maxPageId
 225          * @return bool
 226          * @suppress PhanTypeInvalidDimOffset
 227          */
 228         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 229                 $endDate, $extdb = "", $maxPageId = false
 230         ) {
 231                 global $wgMultiContentRevisionSchemaMigrationStage;
 232
 233                 $dbr = $this->getDB( DB_REPLICA );
 234                 $dbw = $this->getDB( DB_MASTER );
 235
 236                 # Set up external storage
 237                 if ( $extdb != '' ) {
 238                         $esFactory = MediaWikiServices::getInstance()->getExternalStoreFactory();
 239                         /** @var ExternalStoreDB $storeObj */
 240                         $storeObj = $esFactory->getStore( 'DB' );
 241                 }
 242                 // @phan-suppress-next-line PhanAccessMethodInternal
 243                 $blobStore = MediaWikiServices::getInstance()
 244                         ->getBlobStoreFactory()
 245                         ->newSqlBlobStore();
 246
 247                 # Get all articles by page_id
 248                 if ( !$maxPageId ) {
 249                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 250                 }
 251                 $this->output( "Starting from $startId of $maxPageId\n" );
 252                 $pageConds = [];
 253
 254                 /*
 255                 if ( $exclude_ns0 ) {
 256                         print "Excluding main namespace\n";
 257                         $pageConds[] = 'page_namespace<>0';
 258                 }
 259                 if ( $queryExtra ) {
 260                                         $pageConds[] = $queryExtra;
 261                 }
 262                  */
 263
 264                 # For each article, get a list of revisions which fit the criteria
 265
 266                 # No recompression, use a condition on old_flags
 267                 # Don't compress object type entities, because that might produce data loss when
 268                 # overwriting bulk storage concat rows. Don't compress external references, because
 269                 # the script doesn't yet delete rows from external storage.
 270                 $conds = [
 271                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() )
 272                         . ' AND old_flags NOT '
 273                         . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() )
 274                 ];
 275
 276                 if ( $beginDate ) {
 277                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 278                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 279
 280                                 return false;
 281                         }
 282                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 283                 }
 284                 if ( $endDate ) {
 285                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 286                                 $this->error( "Invalid end date \"$endDate\"\n" );
 287
 288                                 return false;
 289                         }
 290                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 291                 }
 292
 293                 if ( $wgMultiContentRevisionSchemaMigrationStage & SCHEMA_COMPAT_READ_OLD ) {
 294                         $tables = [ 'revision', 'text' ];
 295                         $conds[] = 'rev_text_id=old_id';
 296                 } else {
 297                         $slotRoleStore = MediaWikiServices::getInstance()->getSlotRoleStore();
 298                         $tables = [ 'revision', 'slots', 'content', 'text' ];
 299                         $conds = array_merge( [
 300                                 'rev_id=slot_revision_id',
 301                                 'slot_role_id=' . $slotRoleStore->getId( SlotRecord::MAIN ),
 302                                 'content_id=slot_content_id',
 303                                 'SUBSTRING(content_address, 1, 3)=' . $dbr->addQuotes( 'tt:' ),
 304                                 'SUBSTRING(content_address, 4)=old_id',
 305                         ], $conds );
 306                 }
 307
 308                 $fields = [ 'rev_id', 'old_id', 'old_flags', 'old_text' ];
 309                 $revLoadOptions = 'FOR UPDATE';
 310
 311                 # Don't work with current revisions
 312                 # Don't lock the page table for update either -- TS 2006-04-04
 313                 # $tables[] = 'page';
 314                 # $conds[] = 'page_id=rev_page AND rev_id != page_latest';
 315
 316                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 317                         wfWaitForSlaves();
 318
 319                         # Wake up
 320                         $dbr->ping();
 321
 322                         # Get the page row
 323                         $pageRes = $dbr->select( 'page',
 324                                 [ 'page_id', 'page_namespace', 'page_title', 'page_latest' ],
 325                                 $pageConds + [ 'page_id' => $pageId ], __METHOD__ );
 326                         if ( $pageRes->numRows() == 0 ) {
 327                                 continue;
 328                         }
 329                         $pageRow = $dbr->fetchObject( $pageRes );
 330
 331                         # Display progress
 332                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 333                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 334
 335                         # Load revisions
 336                         $revRes = $dbw->select( $tables, $fields,
 337                                 array_merge( [
 338                                         'rev_page' => $pageRow->page_id,
 339                                         # Don't operate on the current revision
 340                                         # Use < instead of <> in case the current revision has changed
 341                                         # since the page select, which wasn't locking
 342                                         'rev_id < ' . $pageRow->page_latest
 343                                 ], $conds ),
 344                                 __METHOD__,
 345                                 $revLoadOptions
 346                         );
 347                         $revs = [];
 348                         foreach ( $revRes as $revRow ) {
 349                                 $revs[] = $revRow;
 350                         }
 351
 352                         if ( count( $revs ) < 2 ) {
 353                                 # No revisions matching, no further processing
 354                                 $this->output( "\n" );
 355                                 continue;
 356                         }
 357
 358                         # For each chunk
 359                         $i = 0;
 360                         while ( $i < count( $revs ) ) {
 361                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 362                                         $thisChunkSize = $maxChunkSize;
 363                                 } else {
 364                                         $thisChunkSize = count( $revs ) - $i;
 365                                 }
 366
 367                                 $chunk = new ConcatenatedGzipHistoryBlob();
 368                                 $stubs = [];
 369                                 $this->beginTransaction( $dbw, __METHOD__ );
 370                                 $usedChunk = false;
 371                                 $primaryOldid = $revs[$i]->old_id;
 372
 373                                 # Get the text of each revision and add it to the object
 374                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 375                                         $oldid = $revs[$i + $j]->old_id;
 376
 377                                         # Get text. We do not need the full `extractBlob` since the query is built
 378                                         # to fetch non-externalstore blobs.
 379                                         $text = $blobStore->decompressData(
 380                                                 $revs[$i + $j]->old_text,
 381                                                 explode( ',', $revs[$i + $j]->old_flags )
 382                                         );
 383
 384                                         if ( $text === false ) {
 385                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 386                                                 # $dbw->delete( 'old', [ 'old_id' => $oldid ] );
 387                                         }
 388
 389                                         if ( $extdb == "" && $j == 0 ) {
 390                                                 $chunk->setText( $text );
 391                                                 $this->output( '.' );
 392                                         } else {
 393                                                 # Don't make a stub if it's going to be longer than the article
 394                                                 # Stubs are typically about 100 bytes
 395                                                 if ( strlen( $text ) < 120 ) {
 396                                                         $stub = false;
 397                                                         $this->output( 'x' );
 398                                                 } else {
 399                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 400                                                         $stub->setLocation( $primaryOldid );
 401                                                         $stub->setReferrer( $oldid );
 402                                                         $this->output( '.' );
 403                                                         $usedChunk = true;
 404                                                 }
 405                                                 $stubs[$j] = $stub;
 406                                         }
 407                                 }
 408                                 $thisChunkSize = $j;
 409
 410                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 411                                 if ( $usedChunk ) {
 412                                         if ( $extdb != "" ) {
 413                                                 # Move blob objects to External Storage
 414                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ) );
 415                                                 if ( $stored === false ) {
 416                                                         $this->error( "Unable to store object" );
 417
 418                                                         return false;
 419                                                 }
 420                                                 # Store External Storage URLs instead of Stub placeholders
 421                                                 foreach ( $stubs as $stub ) {
 422                                                         if ( $stub === false ) {
 423                                                                 continue;
 424                                                         }
 425                                                         # $stored should provide base path to a BLOB
 426                                                         $url = $stored . "/" . $stub->getHash();
 427                                                         $dbw->update( 'text',
 428                                                                 [ /* SET */
 429                                                                         'old_text' => $url,
 430                                                                         'old_flags' => 'external,utf-8',
 431                                                                 ], [ /* WHERE */
 432                                                                         'old_id' => $stub->getReferrer(),
 433                                                                 ]
 434                                                         );
 435                                                 }
 436                                         } else {
 437                                                 # Store the main object locally
 438                                                 $dbw->update( 'text',
 439                                                         [ /* SET */
 440                                                                 'old_text' => serialize( $chunk ),
 441                                                                 'old_flags' => 'object,utf-8',
 442                                                         ], [ /* WHERE */
 443                                                                 'old_id' => $primaryOldid
 444                                                         ]
 445                                                 );
 446
 447                                                 # Store the stub objects
 448                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 449                                                         # Skip if not compressing and don't overwrite the first revision
 450                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->old_id != $primaryOldid ) {
 451                                                                 $dbw->update( 'text',
 452                                                                         [ /* SET */
 453                                                                                 'old_text' => serialize( $stubs[$j] ),
 454                                                                                 'old_flags' => 'object,utf-8',
 455                                                                         ], [ /* WHERE */
 456                                                                                 'old_id' => $revs[$i + $j]->old_id
 457                                                                         ]
 458                                                                 );
 459                                                         }
 460                                                 }
 461                                         }
 462                                 }
 463                                 # Done, next
 464                                 $this->output( "/" );
 465                                 $this->commitTransaction( $dbw, __METHOD__ );
 466                                 $i += $thisChunkSize;
 467                         }
 468                         $this->output( "\n" );
 469                 }
 470
 471                 return true;
 472         }
 473 }
 474
 475 $maintClass = CompressOld::class;
 476 require_once RUN_MAINTENANCE_IF_MAIN;