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