[Summary view] [Print] [Text view]

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