maintenance/storage/checkStorage.php

   1 <?php
   2 /**
   3  * Fsck for MediaWiki
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Maintenance ExternalStorage
  22  */
  23
  24 if ( !defined( 'MEDIAWIKI' ) ) {
  25         $optionsWithoutArgs = [ 'fix' ];
  26         require_once __DIR__ . '/../commandLine.inc';
  27
  28         $cs = new CheckStorage;
  29         $fix = isset( $options['fix'] );
  30         if ( isset( $args[0] ) ) {
  31                 $xml = $args[0];
  32         } else {
  33                 $xml = false;
  34         }
  35         $cs->check( $fix, $xml );
  36 }
  37
  38 // ----------------------------------------------------------------------------------
  39
  40 /**
  41  * Maintenance script to do various checks on external storage.
  42  *
  43  * @fixme this should extend the base Maintenance class
  44  * @ingroup Maintenance ExternalStorage
  45  */
  46 class CheckStorage {
  47         const CONCAT_HEADER = 'O:27:"concatenatedgziphistoryblob"';
  48         public $oldIdMap, $errors;
  49         public $dbStore = null;
  50
  51         public $errorDescriptions = [
  52                 'restore text' => 'Damaged text, need to be restored from a backup',
  53                 'restore revision' => 'Damaged revision row, need to be restored from a backup',
  54                 'unfixable' => 'Unexpected errors with no automated fixing method',
  55                 'fixed' => 'Errors already fixed',
  56                 'fixable' => 'Errors which would already be fixed if --fix was specified',
  57         ];
  58
  59         function check( $fix = false, $xml = '' ) {
  60                 $dbr = wfGetDB( DB_REPLICA );
  61                 if ( $fix ) {
  62                         print "Checking, will fix errors if possible...\n";
  63                 } else {
  64                         print "Checking...\n";
  65                 }
  66                 $maxRevId = $dbr->selectField( 'revision', 'MAX(rev_id)', false, __METHOD__ );
  67                 $chunkSize = 1000;
  68                 $flagStats = [];
  69                 $objectStats = [];
  70                 $knownFlags = [ 'external', 'gzip', 'object', 'utf-8' ];
  71                 $this->errors = [
  72                         'restore text' => [],
  73                         'restore revision' => [],
  74                         'unfixable' => [],
  75                         'fixed' => [],
  76                         'fixable' => [],
  77                 ];
  78
  79                 for ( $chunkStart = 1; $chunkStart < $maxRevId; $chunkStart += $chunkSize ) {
  80                         $chunkEnd = $chunkStart + $chunkSize - 1;
  81                         // print "$chunkStart of $maxRevId\n";
  82
  83                         // Fetch revision rows
  84                         $this->oldIdMap = [];
  85                         $dbr->ping();
  86                         $res = $dbr->select( 'revision', [ 'rev_id', 'rev_text_id' ],
  87                                 [ "rev_id BETWEEN $chunkStart AND $chunkEnd" ], __METHOD__ );
  88                         foreach ( $res as $row ) {
  89                                 $this->oldIdMap[$row->rev_id] = $row->rev_text_id;
  90                         }
  91                         $dbr->freeResult( $res );
  92
  93                         if ( !count( $this->oldIdMap ) ) {
  94                                 continue;
  95                         }
  96
  97                         // Fetch old_flags
  98                         $missingTextRows = array_flip( $this->oldIdMap );
  99                         $externalRevs = [];
 100                         $objectRevs = [];
 101                         $res = $dbr->select( 'text', [ 'old_id', 'old_flags' ],
 102                                 'old_id IN (' . implode( ',', $this->oldIdMap ) . ')', __METHOD__ );
 103                         foreach ( $res as $row ) {
 104                                 /**
 105                                  * @var $flags int
 106                                  */
 107                                 $flags = $row->old_flags;
 108                                 $id = $row->old_id;
 109
 110                                 // Create flagStats row if it doesn't exist
 111                                 $flagStats = $flagStats + [ $flags => 0 ];
 112                                 // Increment counter
 113                                 $flagStats[$flags]++;
 114
 115                                 // Not missing
 116                                 unset( $missingTextRows[$row->old_id] );
 117
 118                                 // Check for external or object
 119                                 if ( $flags == '' ) {
 120                                         $flagArray = [];
 121                                 } else {
 122                                         $flagArray = explode( ',', $flags );
 123                                 }
 124                                 if ( in_array( 'external', $flagArray ) ) {
 125                                         $externalRevs[] = $id;
 126                                 } elseif ( in_array( 'object', $flagArray ) ) {
 127                                         $objectRevs[] = $id;
 128                                 }
 129
 130                                 // Check for unrecognised flags
 131                                 if ( $flags == '0' ) {
 132                                         // This is a known bug from 2004
 133                                         // It's safe to just erase the old_flags field
 134                                         if ( $fix ) {
 135                                                 $this->error( 'fixed', "Warning: old_flags set to 0", $id );
 136                                                 $dbw = wfGetDB( DB_MASTER );
 137                                                 $dbw->ping();
 138                                                 $dbw->update( 'text', [ 'old_flags' => '' ],
 139                                                         [ 'old_id' => $id ], __METHOD__ );
 140                                                 echo "Fixed\n";
 141                                         } else {
 142                                                 $this->error( 'fixable', "Warning: old_flags set to 0", $id );
 143                                         }
 144                                 } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) {
 145                                         $this->error( 'unfixable', "Error: invalid flags field \"$flags\"", $id );
 146                                 }
 147                         }
 148                         $dbr->freeResult( $res );
 149
 150                         // Output errors for any missing text rows
 151                         foreach ( $missingTextRows as $oldId => $revId ) {
 152                                 $this->error( 'restore revision', "Error: missing text row", $oldId );
 153                         }
 154
 155                         // Verify external revisions
 156                         $externalConcatBlobs = [];
 157                         $externalNormalBlobs = [];
 158                         if ( count( $externalRevs ) ) {
 159                                 $res = $dbr->select( 'text', [ 'old_id', 'old_flags', 'old_text' ],
 160                                         [ 'old_id IN (' . implode( ',', $externalRevs ) . ')' ], __METHOD__ );
 161                                 foreach ( $res as $row ) {
 162                                         $urlParts = explode( '://', $row->old_text, 2 );
 163                                         if ( count( $urlParts ) !== 2 || $urlParts[1] == '' ) {
 164                                                 $this->error( 'restore text', "Error: invalid URL \"{$row->old_text}\"", $row->old_id );
 165                                                 continue;
 166                                         }
 167                                         list( $proto, ) = $urlParts;
 168                                         if ( $proto != 'DB' ) {
 169                                                 $this->error( 'restore text', "Error: invalid external protocol \"$proto\"", $row->old_id );
 170                                                 continue;
 171                                         }
 172                                         $path = explode( '/', $row->old_text );
 173                                         $cluster = $path[2];
 174                                         $id = $path[3];
 175                                         if ( isset( $path[4] ) ) {
 176                                                 $externalConcatBlobs[$cluster][$id][] = $row->old_id;
 177                                         } else {
 178                                                 $externalNormalBlobs[$cluster][$id][] = $row->old_id;
 179                                         }
 180                                 }
 181                                 $dbr->freeResult( $res );
 182                         }
 183
 184                         // Check external concat blobs for the right header
 185                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 186
 187                         // Check external normal blobs for existence
 188                         if ( count( $externalNormalBlobs ) ) {
 189                                 if ( is_null( $this->dbStore ) ) {
 190                                         $this->dbStore = new ExternalStoreDB;
 191                                 }
 192                                 foreach ( $externalConcatBlobs as $cluster => $xBlobIds ) {
 193                                         $blobIds = array_keys( $xBlobIds );
 194                                         $extDb =& $this->dbStore->getSlave( $cluster );
 195                                         $blobsTable = $this->dbStore->getTable( $extDb );
 196                                         $res = $extDb->select( $blobsTable,
 197                                                 [ 'blob_id' ],
 198                                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 199                                         foreach ( $res as $row ) {
 200                                                 unset( $xBlobIds[$row->blob_id] );
 201                                         }
 202                                         $extDb->freeResult( $res );
 203                                         // Print errors for missing blobs rows
 204                                         foreach ( $xBlobIds as $blobId => $oldId ) {
 205                                                 $this->error( 'restore text', "Error: missing target $blobId for one-part ES URL", $oldId );
 206                                         }
 207                                 }
 208                         }
 209
 210                         // Check local objects
 211                         $dbr->ping();
 212                         $concatBlobs = [];
 213                         $curIds = [];
 214                         if ( count( $objectRevs ) ) {
 215                                 $headerLength = 300;
 216                                 $res = $dbr->select(
 217                                         'text',
 218                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 219                                         [ 'old_id IN (' . implode( ',', $objectRevs ) . ')' ],
 220                                         __METHOD__
 221                                 );
 222                                 foreach ( $res as $row ) {
 223                                         $oldId = $row->old_id;
 224                                         $matches = [];
 225                                         if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header, $matches ) ) {
 226                                                 $this->error( 'restore text', "Error: invalid object header", $oldId );
 227                                                 continue;
 228                                         }
 229
 230                                         $className = strtolower( $matches[2] );
 231                                         if ( strlen( $className ) != $matches[1] ) {
 232                                                 $this->error(
 233                                                         'restore text',
 234                                                         "Error: invalid object header, wrong class name length",
 235                                                         $oldId
 236                                                 );
 237                                                 continue;
 238                                         }
 239
 240                                         $objectStats = $objectStats + [ $className => 0 ];
 241                                         $objectStats[$className]++;
 242
 243                                         switch ( $className ) {
 244                                                 case 'concatenatedgziphistoryblob':
 245                                                         // Good
 246                                                         break;
 247                                                 case 'historyblobstub':
 248                                                 case 'historyblobcurstub':
 249                                                         if ( strlen( $row->header ) == $headerLength ) {
 250                                                                 $this->error( 'unfixable', "Error: overlong stub header", $oldId );
 251                                                                 continue;
 252                                                         }
 253                                                         $stubObj = unserialize( $row->header );
 254                                                         if ( !is_object( $stubObj ) ) {
 255                                                                 $this->error( 'restore text', "Error: unable to unserialize stub object", $oldId );
 256                                                                 continue;
 257                                                         }
 258                                                         if ( $className == 'historyblobstub' ) {
 259                                                                 $concatBlobs[$stubObj->mOldId][] = $oldId;
 260                                                         } else {
 261                                                                 $curIds[$stubObj->mCurId][] = $oldId;
 262                                                         }
 263                                                         break;
 264                                                 default:
 265                                                         $this->error( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
 266                                         }
 267                                 }
 268                                 $dbr->freeResult( $res );
 269                         }
 270
 271                         // Check local concat blob validity
 272                         $externalConcatBlobs = [];
 273                         if ( count( $concatBlobs ) ) {
 274                                 $headerLength = 300;
 275                                 $res = $dbr->select(
 276                                         'text',
 277                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 278                                         [ 'old_id IN (' . implode( ',', array_keys( $concatBlobs ) ) . ')' ],
 279                                         __METHOD__
 280                                 );
 281                                 foreach ( $res as $row ) {
 282                                         $flags = explode( ',', $row->old_flags );
 283                                         if ( in_array( 'external', $flags ) ) {
 284                                                 // Concat blob is in external storage?
 285                                                 if ( in_array( 'object', $flags ) ) {
 286                                                         $urlParts = explode( '/', $row->header );
 287                                                         if ( $urlParts[0] != 'DB:' ) {
 288                                                                 $this->error(
 289                                                                         'unfixable',
 290                                                                         "Error: unrecognised external storage type \"{$urlParts[0]}",
 291                                                                         $row->old_id
 292                                                                 );
 293                                                         } else {
 294                                                                 $cluster = $urlParts[2];
 295                                                                 $id = $urlParts[3];
 296                                                                 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
 297                                                                         $externalConcatBlobs[$cluster][$id] = [];
 298                                                                 }
 299                                                                 $externalConcatBlobs[$cluster][$id] = array_merge(
 300                                                                         $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id]
 301                                                                 );
 302                                                         }
 303                                                 } else {
 304                                                         $this->error(
 305                                                                 'unfixable',
 306                                                                 "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
 307                                                                 $concatBlobs[$row->old_id] );
 308                                                 }
 309                                         } elseif ( strcasecmp(
 310                                                 substr( $row->header, 0, strlen( self::CONCAT_HEADER ) ),
 311                                                 self::CONCAT_HEADER
 312                                         ) ) {
 313                                                 $this->error(
 314                                                         'restore text',
 315                                                         "Error: Incorrect object header for concat bulk row {$row->old_id}",
 316                                                         $concatBlobs[$row->old_id]
 317                                                 );
 318                                         } # else good
 319
 320                                         unset( $concatBlobs[$row->old_id] );
 321                                 }
 322                                 $dbr->freeResult( $res );
 323                         }
 324
 325                         // Check targets of unresolved stubs
 326                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 327                         // next chunk
 328                 }
 329
 330                 print "\n\nErrors:\n";
 331                 foreach ( $this->errors as $name => $errors ) {
 332                         if ( count( $errors ) ) {
 333                                 $description = $this->errorDescriptions[$name];
 334                                 echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n";
 335                         }
 336                 }
 337
 338                 if ( count( $this->errors['restore text'] ) && $fix ) {
 339                         if ( (string)$xml !== '' ) {
 340                                 $this->restoreText( array_keys( $this->errors['restore text'] ), $xml );
 341                         } else {
 342                                 echo "Can't fix text, no XML backup specified\n";
 343                         }
 344                 }
 345
 346                 print "\nFlag statistics:\n";
 347                 $total = array_sum( $flagStats );
 348                 foreach ( $flagStats as $flag => $count ) {
 349                         printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 );
 350                 }
 351                 print "\nLocal object statistics:\n";
 352                 $total = array_sum( $objectStats );
 353                 foreach ( $objectStats as $className => $count ) {
 354                         printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 );
 355                 }
 356         }
 357
 358         function error( $type, $msg, $ids ) {
 359                 if ( is_array( $ids ) && count( $ids ) == 1 ) {
 360                         $ids = reset( $ids );
 361                 }
 362                 if ( is_array( $ids ) ) {
 363                         $revIds = [];
 364                         foreach ( $ids as $id ) {
 365                                 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap, $id ) );
 366                         }
 367                         print "$msg in text rows " . implode( ', ', $ids ) .
 368                                 ", revisions " . implode( ', ', $revIds ) . "\n";
 369                 } else {
 370                         $id = $ids;
 371                         $revIds = array_keys( $this->oldIdMap, $id );
 372                         if ( count( $revIds ) == 1 ) {
 373                                 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
 374                         } else {
 375                                 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
 376                         }
 377                 }
 378                 $this->errors[$type] = $this->errors[$type] + array_flip( $revIds );
 379         }
 380
 381         function checkExternalConcatBlobs( $externalConcatBlobs ) {
 382                 if ( !count( $externalConcatBlobs ) ) {
 383                         return;
 384                 }
 385
 386                 if ( is_null( $this->dbStore ) ) {
 387                         $this->dbStore = new ExternalStoreDB;
 388                 }
 389
 390                 foreach ( $externalConcatBlobs as $cluster => $oldIds ) {
 391                         $blobIds = array_keys( $oldIds );
 392                         $extDb =& $this->dbStore->getSlave( $cluster );
 393                         $blobsTable = $this->dbStore->getTable( $extDb );
 394                         $headerLength = strlen( self::CONCAT_HEADER );
 395                         $res = $extDb->select( $blobsTable,
 396                                 [ 'blob_id', "LEFT(blob_text, $headerLength) AS header" ],
 397                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 398                         foreach ( $res as $row ) {
 399                                 if ( strcasecmp( $row->header, self::CONCAT_HEADER ) ) {
 400                                         $this->error(
 401                                                 'restore text',
 402                                                 "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
 403                                                 $oldIds[$row->blob_id]
 404                                         );
 405                                 }
 406                                 unset( $oldIds[$row->blob_id] );
 407                         }
 408                         $extDb->freeResult( $res );
 409
 410                         // Print errors for missing blobs rows
 411                         foreach ( $oldIds as $blobId => $oldIds2 ) {
 412                                 $this->error(
 413                                         'restore text',
 414                                         "Error: missing target $cluster/$blobId for two-part ES URL",
 415                                         $oldIds2
 416                                 );
 417                         }
 418                 }
 419         }
 420
 421         function restoreText( $revIds, $xml ) {
 422                 global $wgDBname;
 423                 $tmpDir = wfTempDir();
 424
 425                 if ( !count( $revIds ) ) {
 426                         return;
 427                 }
 428
 429                 print "Restoring text from XML backup...\n";
 430
 431                 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
 432                 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
 433
 434                 // Write revision list
 435                 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
 436                         echo "Error writing revision list, can't restore text\n";
 437
 438                         return;
 439                 }
 440
 441                 // Run mwdumper
 442                 echo "Filtering XML dump...\n";
 443                 $exitStatus = 0;
 444                 passthru( 'mwdumper ' .
 445                         wfEscapeShellArg(
 446                                 "--output=file:$filteredXmlFileName",
 447                                 "--filter=revlist:$revFileName",
 448                                 $xml
 449                         ), $exitStatus
 450                 );
 451
 452                 if ( $exitStatus ) {
 453                         echo "mwdumper died with exit status $exitStatus\n";
 454
 455                         return;
 456                 }
 457
 458                 $file = fopen( $filteredXmlFileName, 'r' );
 459                 if ( !$file ) {
 460                         echo "Unable to open filtered XML file\n";
 461
 462                         return;
 463                 }
 464
 465                 $dbr = wfGetDB( DB_REPLICA );
 466                 $dbw = wfGetDB( DB_MASTER );
 467                 $dbr->ping();
 468                 $dbw->ping();
 469
 470                 $source = new ImportStreamSource( $file );
 471                 $importer = new WikiImporter(
 472                         $source,
 473                         ConfigFactory::getDefaultInstance()->makeConfig( 'main' )
 474                 );
 475                 $importer->setRevisionCallback( [ $this, 'importRevision' ] );
 476                 $importer->doImport();
 477         }
 478
 479         function importRevision( &$revision, &$importer ) {
 480                 $id = $revision->getID();
 481                 $content = $revision->getContent( Revision::RAW );
 482                 $id = $id ? $id : '';
 483
 484                 if ( $content === null ) {
 485                         echo "Revision $id is broken, we have no content available\n";
 486
 487                         return;
 488                 }
 489
 490                 $text = $content->serialize();
 491                 if ( $text === '' ) {
 492                         // This is what happens if the revision was broken at the time the
 493                         // dump was made. Unfortunately, it also happens if the revision was
 494                         // legitimately blank, so there's no way to tell the difference. To
 495                         // be safe, we'll skip it and leave it broken
 496
 497                         echo "Revision $id is blank in the dump, may have been broken before export\n";
 498
 499                         return;
 500                 }
 501
 502                 if ( !$id ) {
 503                         // No ID, can't import
 504                         echo "No id tag in revision, can't import\n";
 505
 506                         return;
 507                 }
 508
 509                 // Find text row again
 510                 $dbr = wfGetDB( DB_REPLICA );
 511                 $oldId = $dbr->selectField( 'revision', 'rev_text_id', [ 'rev_id' => $id ], __METHOD__ );
 512                 if ( !$oldId ) {
 513                         echo "Missing revision row for rev_id $id\n";
 514
 515                         return;
 516                 }
 517
 518                 // Compress the text
 519                 $flags = Revision::compressRevisionText( $text );
 520
 521                 // Update the text row
 522                 $dbw = wfGetDB( DB_MASTER );
 523                 $dbw->update( 'text',
 524                         [ 'old_flags' => $flags, 'old_text' => $text ],
 525                         [ 'old_id' => $oldId ],
 526                         __METHOD__, [ 'LIMIT' => 1 ]
 527                 );
 528
 529                 // Remove it from the unfixed list and add it to the fixed list
 530                 unset( $this->errors['restore text'][$id] );
 531                 $this->errors['fixed'][$id] = true;
 532         }
 533 }