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 use MediaWiki\MediaWikiServices;
  25
  26 if ( !defined( 'MEDIAWIKI' ) ) {
  27         $optionsWithoutArgs = [ 'fix' ];
  28         require_once __DIR__ . '/../commandLine.inc';
  29
  30         $cs = new CheckStorage;
  31         $fix = isset( $options['fix'] );
  32         if ( isset( $args[0] ) ) {
  33                 $xml = $args[0];
  34         } else {
  35                 $xml = false;
  36         }
  37         $cs->check( $fix, $xml );
  38 }
  39
  40 // ----------------------------------------------------------------------------------
  41
  42 /**
  43  * Maintenance script to do various checks on external storage.
  44  *
  45  * @fixme this should extend the base Maintenance class
  46  * @ingroup Maintenance ExternalStorage
  47  */
  48 class CheckStorage {
  49         const CONCAT_HEADER = 'O:27:"concatenatedgziphistoryblob"';
  50         public $oldIdMap, $errors;
  51         public $dbStore = null;
  52
  53         public $errorDescriptions = [
  54                 'restore text' => 'Damaged text, need to be restored from a backup',
  55                 'restore revision' => 'Damaged revision row, need to be restored from a backup',
  56                 'unfixable' => 'Unexpected errors with no automated fixing method',
  57                 'fixed' => 'Errors already fixed',
  58                 'fixable' => 'Errors which would already be fixed if --fix was specified',
  59         ];
  60
  61         function check( $fix = false, $xml = '' ) {
  62                 $dbr = wfGetDB( DB_REPLICA );
  63                 if ( $fix ) {
  64                         print "Checking, will fix errors if possible...\n";
  65                 } else {
  66                         print "Checking...\n";
  67                 }
  68                 $maxRevId = $dbr->selectField( 'revision', 'MAX(rev_id)', false, __METHOD__ );
  69                 $chunkSize = 1000;
  70                 $flagStats = [];
  71                 $objectStats = [];
  72                 $knownFlags = [ 'external', 'gzip', 'object', 'utf-8' ];
  73                 $this->errors = [
  74                         'restore text' => [],
  75                         'restore revision' => [],
  76                         'unfixable' => [],
  77                         'fixed' => [],
  78                         'fixable' => [],
  79                 ];
  80
  81                 for ( $chunkStart = 1; $chunkStart < $maxRevId; $chunkStart += $chunkSize ) {
  82                         $chunkEnd = $chunkStart + $chunkSize - 1;
  83                         // print "$chunkStart of $maxRevId\n";
  84
  85                         // Fetch revision rows
  86                         $this->oldIdMap = [];
  87                         $dbr->ping();
  88                         $res = $dbr->select( 'revision', [ 'rev_id', 'rev_text_id' ],
  89                                 [ "rev_id BETWEEN $chunkStart AND $chunkEnd" ], __METHOD__ );
  90                         foreach ( $res as $row ) {
  91                                 $this->oldIdMap[$row->rev_id] = $row->rev_text_id;
  92                         }
  93                         $dbr->freeResult( $res );
  94
  95                         if ( !count( $this->oldIdMap ) ) {
  96                                 continue;
  97                         }
  98
  99                         // Fetch old_flags
 100                         $missingTextRows = array_flip( $this->oldIdMap );
 101                         $externalRevs = [];
 102                         $objectRevs = [];
 103                         $res = $dbr->select( 'text', [ 'old_id', 'old_flags' ],
 104                                 'old_id IN (' . implode( ',', $this->oldIdMap ) . ')', __METHOD__ );
 105                         foreach ( $res as $row ) {
 106                                 /**
 107                                  * @var $flags int
 108                                  */
 109                                 $flags = $row->old_flags;
 110                                 $id = $row->old_id;
 111
 112                                 // Create flagStats row if it doesn't exist
 113                                 $flagStats = $flagStats + [ $flags => 0 ];
 114                                 // Increment counter
 115                                 $flagStats[$flags]++;
 116
 117                                 // Not missing
 118                                 unset( $missingTextRows[$row->old_id] );
 119
 120                                 // Check for external or object
 121                                 if ( $flags == '' ) {
 122                                         $flagArray = [];
 123                                 } else {
 124                                         $flagArray = explode( ',', $flags );
 125                                 }
 126                                 if ( in_array( 'external', $flagArray ) ) {
 127                                         $externalRevs[] = $id;
 128                                 } elseif ( in_array( 'object', $flagArray ) ) {
 129                                         $objectRevs[] = $id;
 130                                 }
 131
 132                                 // Check for unrecognised flags
 133                                 if ( $flags == '0' ) {
 134                                         // This is a known bug from 2004
 135                                         // It's safe to just erase the old_flags field
 136                                         if ( $fix ) {
 137                                                 $this->addError( 'fixed', "Warning: old_flags set to 0", $id );
 138                                                 $dbw = wfGetDB( DB_MASTER );
 139                                                 $dbw->ping();
 140                                                 $dbw->update( 'text', [ 'old_flags' => '' ],
 141                                                         [ 'old_id' => $id ], __METHOD__ );
 142                                                 echo "Fixed\n";
 143                                         } else {
 144                                                 $this->addError( 'fixable', "Warning: old_flags set to 0", $id );
 145                                         }
 146                                 } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) {
 147                                         $this->addError( 'unfixable', "Error: invalid flags field \"$flags\"", $id );
 148                                 }
 149                         }
 150                         $dbr->freeResult( $res );
 151
 152                         // Output errors for any missing text rows
 153                         foreach ( $missingTextRows as $oldId => $revId ) {
 154                                 $this->addError( 'restore revision', "Error: missing text row", $oldId );
 155                         }
 156
 157                         // Verify external revisions
 158                         $externalConcatBlobs = [];
 159                         $externalNormalBlobs = [];
 160                         if ( count( $externalRevs ) ) {
 161                                 $res = $dbr->select( 'text', [ 'old_id', 'old_flags', 'old_text' ],
 162                                         [ 'old_id IN (' . implode( ',', $externalRevs ) . ')' ], __METHOD__ );
 163                                 foreach ( $res as $row ) {
 164                                         $urlParts = explode( '://', $row->old_text, 2 );
 165                                         if ( count( $urlParts ) !== 2 || $urlParts[1] == '' ) {
 166                                                 $this->addError( 'restore text', "Error: invalid URL \"{$row->old_text}\"", $row->old_id );
 167                                                 continue;
 168                                         }
 169                                         list( $proto, ) = $urlParts;
 170                                         if ( $proto != 'DB' ) {
 171                                                 $this->addError(
 172                                                         'restore text',
 173                                                         "Error: invalid external protocol \"$proto\"",
 174                                                         $row->old_id );
 175                                                 continue;
 176                                         }
 177                                         $path = explode( '/', $row->old_text );
 178                                         $cluster = $path[2];
 179                                         $id = $path[3];
 180                                         if ( isset( $path[4] ) ) {
 181                                                 $externalConcatBlobs[$cluster][$id][] = $row->old_id;
 182                                         } else {
 183                                                 $externalNormalBlobs[$cluster][$id][] = $row->old_id;
 184                                         }
 185                                 }
 186                                 $dbr->freeResult( $res );
 187                         }
 188
 189                         // Check external concat blobs for the right header
 190                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 191
 192                         // Check external normal blobs for existence
 193                         if ( count( $externalNormalBlobs ) ) {
 194                                 if ( is_null( $this->dbStore ) ) {
 195                                         $this->dbStore = new ExternalStoreDB;
 196                                 }
 197                                 foreach ( $externalConcatBlobs as $cluster => $xBlobIds ) {
 198                                         $blobIds = array_keys( $xBlobIds );
 199                                         $extDb =& $this->dbStore->getSlave( $cluster );
 200                                         $blobsTable = $this->dbStore->getTable( $extDb );
 201                                         $res = $extDb->select( $blobsTable,
 202                                                 [ 'blob_id' ],
 203                                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 204                                         foreach ( $res as $row ) {
 205                                                 unset( $xBlobIds[$row->blob_id] );
 206                                         }
 207                                         $extDb->freeResult( $res );
 208                                         // Print errors for missing blobs rows
 209                                         foreach ( $xBlobIds as $blobId => $oldId ) {
 210                                                 $this->addError(
 211                                                         'restore text',
 212                                                         "Error: missing target $blobId for one-part ES URL",
 213                                                         $oldId );
 214                                         }
 215                                 }
 216                         }
 217
 218                         // Check local objects
 219                         $dbr->ping();
 220                         $concatBlobs = [];
 221                         $curIds = [];
 222                         if ( count( $objectRevs ) ) {
 223                                 $headerLength = 300;
 224                                 $res = $dbr->select(
 225                                         'text',
 226                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 227                                         [ 'old_id IN (' . implode( ',', $objectRevs ) . ')' ],
 228                                         __METHOD__
 229                                 );
 230                                 foreach ( $res as $row ) {
 231                                         $oldId = $row->old_id;
 232                                         $matches = [];
 233                                         if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header, $matches ) ) {
 234                                                 $this->addError( 'restore text', "Error: invalid object header", $oldId );
 235                                                 continue;
 236                                         }
 237
 238                                         $className = strtolower( $matches[2] );
 239                                         if ( strlen( $className ) != $matches[1] ) {
 240                                                 $this->addError(
 241                                                         'restore text',
 242                                                         "Error: invalid object header, wrong class name length",
 243                                                         $oldId
 244                                                 );
 245                                                 continue;
 246                                         }
 247
 248                                         $objectStats = $objectStats + [ $className => 0 ];
 249                                         $objectStats[$className]++;
 250
 251                                         switch ( $className ) {
 252                                                 case 'concatenatedgziphistoryblob':
 253                                                         // Good
 254                                                         break;
 255                                                 case 'historyblobstub':
 256                                                 case 'historyblobcurstub':
 257                                                         if ( strlen( $row->header ) == $headerLength ) {
 258                                                                 $this->addError( 'unfixable', "Error: overlong stub header", $oldId );
 259                                                                 continue;
 260                                                         }
 261                                                         $stubObj = unserialize( $row->header );
 262                                                         if ( !is_object( $stubObj ) ) {
 263                                                                 $this->addError( 'restore text', "Error: unable to unserialize stub object", $oldId );
 264                                                                 continue;
 265                                                         }
 266                                                         if ( $className == 'historyblobstub' ) {
 267                                                                 $concatBlobs[$stubObj->mOldId][] = $oldId;
 268                                                         } else {
 269                                                                 $curIds[$stubObj->mCurId][] = $oldId;
 270                                                         }
 271                                                         break;
 272                                                 default:
 273                                                         $this->addError( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
 274                                         }
 275                                 }
 276                                 $dbr->freeResult( $res );
 277                         }
 278
 279                         // Check local concat blob validity
 280                         $externalConcatBlobs = [];
 281                         if ( count( $concatBlobs ) ) {
 282                                 $headerLength = 300;
 283                                 $res = $dbr->select(
 284                                         'text',
 285                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 286                                         [ 'old_id IN (' . implode( ',', array_keys( $concatBlobs ) ) . ')' ],
 287                                         __METHOD__
 288                                 );
 289                                 foreach ( $res as $row ) {
 290                                         $flags = explode( ',', $row->old_flags );
 291                                         if ( in_array( 'external', $flags ) ) {
 292                                                 // Concat blob is in external storage?
 293                                                 if ( in_array( 'object', $flags ) ) {
 294                                                         $urlParts = explode( '/', $row->header );
 295                                                         if ( $urlParts[0] != 'DB:' ) {
 296                                                                 $this->addError(
 297                                                                         'unfixable',
 298                                                                         "Error: unrecognised external storage type \"{$urlParts[0]}",
 299                                                                         $row->old_id
 300                                                                 );
 301                                                         } else {
 302                                                                 $cluster = $urlParts[2];
 303                                                                 $id = $urlParts[3];
 304                                                                 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
 305                                                                         $externalConcatBlobs[$cluster][$id] = [];
 306                                                                 }
 307                                                                 $externalConcatBlobs[$cluster][$id] = array_merge(
 308                                                                         $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id]
 309                                                                 );
 310                                                         }
 311                                                 } else {
 312                                                         $this->addError(
 313                                                                 'unfixable',
 314                                                                 "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
 315                                                                 $concatBlobs[$row->old_id] );
 316                                                 }
 317                                         } elseif ( strcasecmp(
 318                                                 substr( $row->header, 0, strlen( self::CONCAT_HEADER ) ),
 319                                                 self::CONCAT_HEADER
 320                                         ) ) {
 321                                                 $this->addError(
 322                                                         'restore text',
 323                                                         "Error: Incorrect object header for concat bulk row {$row->old_id}",
 324                                                         $concatBlobs[$row->old_id]
 325                                                 );
 326                                         } # else good
 327
 328                                         unset( $concatBlobs[$row->old_id] );
 329                                 }
 330                                 $dbr->freeResult( $res );
 331                         }
 332
 333                         // Check targets of unresolved stubs
 334                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 335                         // next chunk
 336                 }
 337
 338                 print "\n\nErrors:\n";
 339                 foreach ( $this->errors as $name => $errors ) {
 340                         if ( count( $errors ) ) {
 341                                 $description = $this->errorDescriptions[$name];
 342                                 echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n";
 343                         }
 344                 }
 345
 346                 if ( count( $this->errors['restore text'] ) && $fix ) {
 347                         if ( (string)$xml !== '' ) {
 348                                 $this->restoreText( array_keys( $this->errors['restore text'] ), $xml );
 349                         } else {
 350                                 echo "Can't fix text, no XML backup specified\n";
 351                         }
 352                 }
 353
 354                 print "\nFlag statistics:\n";
 355                 $total = array_sum( $flagStats );
 356                 foreach ( $flagStats as $flag => $count ) {
 357                         printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 );
 358                 }
 359                 print "\nLocal object statistics:\n";
 360                 $total = array_sum( $objectStats );
 361                 foreach ( $objectStats as $className => $count ) {
 362                         printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 );
 363                 }
 364         }
 365
 366         function addError( $type, $msg, $ids ) {
 367                 if ( is_array( $ids ) && count( $ids ) == 1 ) {
 368                         $ids = reset( $ids );
 369                 }
 370                 if ( is_array( $ids ) ) {
 371                         $revIds = [];
 372                         foreach ( $ids as $id ) {
 373                                 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap, $id ) );
 374                         }
 375                         print "$msg in text rows " . implode( ', ', $ids ) .
 376                                 ", revisions " . implode( ', ', $revIds ) . "\n";
 377                 } else {
 378                         $id = $ids;
 379                         $revIds = array_keys( $this->oldIdMap, $id );
 380                         if ( count( $revIds ) == 1 ) {
 381                                 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
 382                         } else {
 383                                 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
 384                         }
 385                 }
 386                 $this->errors[$type] = $this->errors[$type] + array_flip( $revIds );
 387         }
 388
 389         function checkExternalConcatBlobs( $externalConcatBlobs ) {
 390                 if ( !count( $externalConcatBlobs ) ) {
 391                         return;
 392                 }
 393
 394                 if ( is_null( $this->dbStore ) ) {
 395                         $this->dbStore = new ExternalStoreDB;
 396                 }
 397
 398                 foreach ( $externalConcatBlobs as $cluster => $oldIds ) {
 399                         $blobIds = array_keys( $oldIds );
 400                         $extDb =& $this->dbStore->getSlave( $cluster );
 401                         $blobsTable = $this->dbStore->getTable( $extDb );
 402                         $headerLength = strlen( self::CONCAT_HEADER );
 403                         $res = $extDb->select( $blobsTable,
 404                                 [ 'blob_id', "LEFT(blob_text, $headerLength) AS header" ],
 405                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 406                         foreach ( $res as $row ) {
 407                                 if ( strcasecmp( $row->header, self::CONCAT_HEADER ) ) {
 408                                         $this->addError(
 409                                                 'restore text',
 410                                                 "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
 411                                                 $oldIds[$row->blob_id]
 412                                         );
 413                                 }
 414                                 unset( $oldIds[$row->blob_id] );
 415                         }
 416                         $extDb->freeResult( $res );
 417
 418                         // Print errors for missing blobs rows
 419                         foreach ( $oldIds as $blobId => $oldIds2 ) {
 420                                 $this->addError(
 421                                         'restore text',
 422                                         "Error: missing target $cluster/$blobId for two-part ES URL",
 423                                         $oldIds2
 424                                 );
 425                         }
 426                 }
 427         }
 428
 429         function restoreText( $revIds, $xml ) {
 430                 global $wgDBname;
 431                 $tmpDir = wfTempDir();
 432
 433                 if ( !count( $revIds ) ) {
 434                         return;
 435                 }
 436
 437                 print "Restoring text from XML backup...\n";
 438
 439                 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
 440                 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
 441
 442                 // Write revision list
 443                 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
 444                         echo "Error writing revision list, can't restore text\n";
 445
 446                         return;
 447                 }
 448
 449                 // Run mwdumper
 450                 echo "Filtering XML dump...\n";
 451                 $exitStatus = 0;
 452                 passthru( 'mwdumper ' .
 453                         wfEscapeShellArg(
 454                                 "--output=file:$filteredXmlFileName",
 455                                 "--filter=revlist:$revFileName",
 456                                 $xml
 457                         ), $exitStatus
 458                 );
 459
 460                 if ( $exitStatus ) {
 461                         echo "mwdumper died with exit status $exitStatus\n";
 462
 463                         return;
 464                 }
 465
 466                 $file = fopen( $filteredXmlFileName, 'r' );
 467                 if ( !$file ) {
 468                         echo "Unable to open filtered XML file\n";
 469
 470                         return;
 471                 }
 472
 473                 $dbr = wfGetDB( DB_REPLICA );
 474                 $dbw = wfGetDB( DB_MASTER );
 475                 $dbr->ping();
 476                 $dbw->ping();
 477
 478                 $source = new ImportStreamSource( $file );
 479                 $importer = new WikiImporter(
 480                         $source,
 481                         MediaWikiServices::getInstance()->getMainConfig()
 482                 );
 483                 $importer->setRevisionCallback( [ $this, 'importRevision' ] );
 484                 $importer->doImport();
 485         }
 486
 487         function importRevision( &$revision, &$importer ) {
 488                 $id = $revision->getID();
 489                 $content = $revision->getContent( Revision::RAW );
 490                 $id = $id ? $id : '';
 491
 492                 if ( $content === null ) {
 493                         echo "Revision $id is broken, we have no content available\n";
 494
 495                         return;
 496                 }
 497
 498                 $text = $content->serialize();
 499                 if ( $text === '' ) {
 500                         // This is what happens if the revision was broken at the time the
 501                         // dump was made. Unfortunately, it also happens if the revision was
 502                         // legitimately blank, so there's no way to tell the difference. To
 503                         // be safe, we'll skip it and leave it broken
 504
 505                         echo "Revision $id is blank in the dump, may have been broken before export\n";
 506
 507                         return;
 508                 }
 509
 510                 if ( !$id ) {
 511                         // No ID, can't import
 512                         echo "No id tag in revision, can't import\n";
 513
 514                         return;
 515                 }
 516
 517                 // Find text row again
 518                 $dbr = wfGetDB( DB_REPLICA );
 519                 $oldId = $dbr->selectField( 'revision', 'rev_text_id', [ 'rev_id' => $id ], __METHOD__ );
 520                 if ( !$oldId ) {
 521                         echo "Missing revision row for rev_id $id\n";
 522
 523                         return;
 524                 }
 525
 526                 // Compress the text
 527                 $flags = Revision::compressRevisionText( $text );
 528
 529                 // Update the text row
 530                 $dbw = wfGetDB( DB_MASTER );
 531                 $dbw->update( 'text',
 532                         [ 'old_flags' => $flags, 'old_text' => $text ],
 533                         [ 'old_id' => $oldId ],
 534                         __METHOD__, [ 'LIMIT' => 1 ]
 535                 );
 536
 537                 // Remove it from the unfixed list and add it to the fixed list
 538                 unset( $this->errors['restore text'][$id] );
 539                 $this->errors['fixed'][$id] = true;
 540         }
 541 }