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(
 104                                 'text',
 105                                 [ 'old_id', 'old_flags' ],
 106                                 [ 'old_id' => $this->oldIdMap ],
 107                                 __METHOD__
 108                         );
 109                         foreach ( $res as $row ) {
 110                                 /**
 111                                  * @var $flags int
 112                                  */
 113                                 $flags = $row->old_flags;
 114                                 $id = $row->old_id;
 115
 116                                 // Create flagStats row if it doesn't exist
 117                                 $flagStats = $flagStats + [ $flags => 0 ];
 118                                 // Increment counter
 119                                 $flagStats[$flags]++;
 120
 121                                 // Not missing
 122                                 unset( $missingTextRows[$row->old_id] );
 123
 124                                 // Check for external or object
 125                                 if ( $flags == '' ) {
 126                                         $flagArray = [];
 127                                 } else {
 128                                         $flagArray = explode( ',', $flags );
 129                                 }
 130                                 if ( in_array( 'external', $flagArray ) ) {
 131                                         $externalRevs[] = $id;
 132                                 } elseif ( in_array( 'object', $flagArray ) ) {
 133                                         $objectRevs[] = $id;
 134                                 }
 135
 136                                 // Check for unrecognised flags
 137                                 if ( $flags == '0' ) {
 138                                         // This is a known bug from 2004
 139                                         // It's safe to just erase the old_flags field
 140                                         if ( $fix ) {
 141                                                 $this->addError( 'fixed', "Warning: old_flags set to 0", $id );
 142                                                 $dbw = wfGetDB( DB_MASTER );
 143                                                 $dbw->ping();
 144                                                 $dbw->update( 'text', [ 'old_flags' => '' ],
 145                                                         [ 'old_id' => $id ], __METHOD__ );
 146                                                 echo "Fixed\n";
 147                                         } else {
 148                                                 $this->addError( 'fixable', "Warning: old_flags set to 0", $id );
 149                                         }
 150                                 } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) {
 151                                         $this->addError( 'unfixable', "Error: invalid flags field \"$flags\"", $id );
 152                                 }
 153                         }
 154                         $dbr->freeResult( $res );
 155
 156                         // Output errors for any missing text rows
 157                         foreach ( $missingTextRows as $oldId => $revId ) {
 158                                 $this->addError( 'restore revision', "Error: missing text row", $oldId );
 159                         }
 160
 161                         // Verify external revisions
 162                         $externalConcatBlobs = [];
 163                         $externalNormalBlobs = [];
 164                         if ( count( $externalRevs ) ) {
 165                                 $res = $dbr->select(
 166                                         'text',
 167                                         [ 'old_id', 'old_flags', 'old_text' ],
 168                                         [ 'old_id' => $externalRevs ],
 169                                         __METHOD__
 170                                 );
 171                                 foreach ( $res as $row ) {
 172                                         $urlParts = explode( '://', $row->old_text, 2 );
 173                                         if ( count( $urlParts ) !== 2 || $urlParts[1] == '' ) {
 174                                                 $this->addError( 'restore text', "Error: invalid URL \"{$row->old_text}\"", $row->old_id );
 175                                                 continue;
 176                                         }
 177                                         list( $proto, ) = $urlParts;
 178                                         if ( $proto != 'DB' ) {
 179                                                 $this->addError(
 180                                                         'restore text',
 181                                                         "Error: invalid external protocol \"$proto\"",
 182                                                         $row->old_id );
 183                                                 continue;
 184                                         }
 185                                         $path = explode( '/', $row->old_text );
 186                                         $cluster = $path[2];
 187                                         $id = $path[3];
 188                                         if ( isset( $path[4] ) ) {
 189                                                 $externalConcatBlobs[$cluster][$id][] = $row->old_id;
 190                                         } else {
 191                                                 $externalNormalBlobs[$cluster][$id][] = $row->old_id;
 192                                         }
 193                                 }
 194                                 $dbr->freeResult( $res );
 195                         }
 196
 197                         // Check external concat blobs for the right header
 198                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 199
 200                         // Check external normal blobs for existence
 201                         if ( count( $externalNormalBlobs ) ) {
 202                                 if ( is_null( $this->dbStore ) ) {
 203                                         $this->dbStore = new ExternalStoreDB;
 204                                 }
 205                                 foreach ( $externalConcatBlobs as $cluster => $xBlobIds ) {
 206                                         $blobIds = array_keys( $xBlobIds );
 207                                         $extDb =& $this->dbStore->getSlave( $cluster );
 208                                         $blobsTable = $this->dbStore->getTable( $extDb );
 209                                         $res = $extDb->select( $blobsTable,
 210                                                 [ 'blob_id' ],
 211                                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 212                                         foreach ( $res as $row ) {
 213                                                 unset( $xBlobIds[$row->blob_id] );
 214                                         }
 215                                         $extDb->freeResult( $res );
 216                                         // Print errors for missing blobs rows
 217                                         foreach ( $xBlobIds as $blobId => $oldId ) {
 218                                                 $this->addError(
 219                                                         'restore text',
 220                                                         "Error: missing target $blobId for one-part ES URL",
 221                                                         $oldId );
 222                                         }
 223                                 }
 224                         }
 225
 226                         // Check local objects
 227                         $dbr->ping();
 228                         $concatBlobs = [];
 229                         $curIds = [];
 230                         if ( count( $objectRevs ) ) {
 231                                 $headerLength = 300;
 232                                 $res = $dbr->select(
 233                                         'text',
 234                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 235                                         [ 'old_id' => $objectRevs ],
 236                                         __METHOD__
 237                                 );
 238                                 foreach ( $res as $row ) {
 239                                         $oldId = $row->old_id;
 240                                         $matches = [];
 241                                         if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header, $matches ) ) {
 242                                                 $this->addError( 'restore text', "Error: invalid object header", $oldId );
 243                                                 continue;
 244                                         }
 245
 246                                         $className = strtolower( $matches[2] );
 247                                         if ( strlen( $className ) != $matches[1] ) {
 248                                                 $this->addError(
 249                                                         'restore text',
 250                                                         "Error: invalid object header, wrong class name length",
 251                                                         $oldId
 252                                                 );
 253                                                 continue;
 254                                         }
 255
 256                                         $objectStats = $objectStats + [ $className => 0 ];
 257                                         $objectStats[$className]++;
 258
 259                                         switch ( $className ) {
 260                                                 case 'concatenatedgziphistoryblob':
 261                                                         // Good
 262                                                         break;
 263                                                 case 'historyblobstub':
 264                                                 case 'historyblobcurstub':
 265                                                         if ( strlen( $row->header ) == $headerLength ) {
 266                                                                 $this->addError( 'unfixable', "Error: overlong stub header", $oldId );
 267                                                                 continue;
 268                                                         }
 269                                                         $stubObj = unserialize( $row->header );
 270                                                         if ( !is_object( $stubObj ) ) {
 271                                                                 $this->addError( 'restore text', "Error: unable to unserialize stub object", $oldId );
 272                                                                 continue;
 273                                                         }
 274                                                         if ( $className == 'historyblobstub' ) {
 275                                                                 $concatBlobs[$stubObj->mOldId][] = $oldId;
 276                                                         } else {
 277                                                                 $curIds[$stubObj->mCurId][] = $oldId;
 278                                                         }
 279                                                         break;
 280                                                 default:
 281                                                         $this->addError( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
 282                                         }
 283                                 }
 284                                 $dbr->freeResult( $res );
 285                         }
 286
 287                         // Check local concat blob validity
 288                         $externalConcatBlobs = [];
 289                         if ( count( $concatBlobs ) ) {
 290                                 $headerLength = 300;
 291                                 $res = $dbr->select(
 292                                         'text',
 293                                         [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
 294                                         [ 'old_id' => array_keys( $concatBlobs ) ],
 295                                         __METHOD__
 296                                 );
 297                                 foreach ( $res as $row ) {
 298                                         $flags = explode( ',', $row->old_flags );
 299                                         if ( in_array( 'external', $flags ) ) {
 300                                                 // Concat blob is in external storage?
 301                                                 if ( in_array( 'object', $flags ) ) {
 302                                                         $urlParts = explode( '/', $row->header );
 303                                                         if ( $urlParts[0] != 'DB:' ) {
 304                                                                 $this->addError(
 305                                                                         'unfixable',
 306                                                                         "Error: unrecognised external storage type \"{$urlParts[0]}",
 307                                                                         $row->old_id
 308                                                                 );
 309                                                         } else {
 310                                                                 $cluster = $urlParts[2];
 311                                                                 $id = $urlParts[3];
 312                                                                 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
 313                                                                         $externalConcatBlobs[$cluster][$id] = [];
 314                                                                 }
 315                                                                 $externalConcatBlobs[$cluster][$id] = array_merge(
 316                                                                         $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id]
 317                                                                 );
 318                                                         }
 319                                                 } else {
 320                                                         $this->addError(
 321                                                                 'unfixable',
 322                                                                 "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
 323                                                                 $concatBlobs[$row->old_id] );
 324                                                 }
 325                                         } elseif ( strcasecmp(
 326                                                 substr( $row->header, 0, strlen( self::CONCAT_HEADER ) ),
 327                                                 self::CONCAT_HEADER
 328                                         ) ) {
 329                                                 $this->addError(
 330                                                         'restore text',
 331                                                         "Error: Incorrect object header for concat bulk row {$row->old_id}",
 332                                                         $concatBlobs[$row->old_id]
 333                                                 );
 334                                         } # else good
 335
 336                                         unset( $concatBlobs[$row->old_id] );
 337                                 }
 338                                 $dbr->freeResult( $res );
 339                         }
 340
 341                         // Check targets of unresolved stubs
 342                         $this->checkExternalConcatBlobs( $externalConcatBlobs );
 343                         // next chunk
 344                 }
 345
 346                 print "\n\nErrors:\n";
 347                 foreach ( $this->errors as $name => $errors ) {
 348                         if ( count( $errors ) ) {
 349                                 $description = $this->errorDescriptions[$name];
 350                                 echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n";
 351                         }
 352                 }
 353
 354                 if ( count( $this->errors['restore text'] ) && $fix ) {
 355                         if ( (string)$xml !== '' ) {
 356                                 $this->restoreText( array_keys( $this->errors['restore text'] ), $xml );
 357                         } else {
 358                                 echo "Can't fix text, no XML backup specified\n";
 359                         }
 360                 }
 361
 362                 print "\nFlag statistics:\n";
 363                 $total = array_sum( $flagStats );
 364                 foreach ( $flagStats as $flag => $count ) {
 365                         printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 );
 366                 }
 367                 print "\nLocal object statistics:\n";
 368                 $total = array_sum( $objectStats );
 369                 foreach ( $objectStats as $className => $count ) {
 370                         printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 );
 371                 }
 372         }
 373
 374         function addError( $type, $msg, $ids ) {
 375                 if ( is_array( $ids ) && count( $ids ) == 1 ) {
 376                         $ids = reset( $ids );
 377                 }
 378                 if ( is_array( $ids ) ) {
 379                         $revIds = [];
 380                         foreach ( $ids as $id ) {
 381                                 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap, $id ) );
 382                         }
 383                         print "$msg in text rows " . implode( ', ', $ids ) .
 384                                 ", revisions " . implode( ', ', $revIds ) . "\n";
 385                 } else {
 386                         $id = $ids;
 387                         $revIds = array_keys( $this->oldIdMap, $id );
 388                         if ( count( $revIds ) == 1 ) {
 389                                 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
 390                         } else {
 391                                 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
 392                         }
 393                 }
 394                 $this->errors[$type] = $this->errors[$type] + array_flip( $revIds );
 395         }
 396
 397         function checkExternalConcatBlobs( $externalConcatBlobs ) {
 398                 if ( !count( $externalConcatBlobs ) ) {
 399                         return;
 400                 }
 401
 402                 if ( is_null( $this->dbStore ) ) {
 403                         $this->dbStore = new ExternalStoreDB;
 404                 }
 405
 406                 foreach ( $externalConcatBlobs as $cluster => $oldIds ) {
 407                         $blobIds = array_keys( $oldIds );
 408                         $extDb =& $this->dbStore->getSlave( $cluster );
 409                         $blobsTable = $this->dbStore->getTable( $extDb );
 410                         $headerLength = strlen( self::CONCAT_HEADER );
 411                         $res = $extDb->select( $blobsTable,
 412                                 [ 'blob_id', "LEFT(blob_text, $headerLength) AS header" ],
 413                                 [ 'blob_id IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__ );
 414                         foreach ( $res as $row ) {
 415                                 if ( strcasecmp( $row->header, self::CONCAT_HEADER ) ) {
 416                                         $this->addError(
 417                                                 'restore text',
 418                                                 "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
 419                                                 $oldIds[$row->blob_id]
 420                                         );
 421                                 }
 422                                 unset( $oldIds[$row->blob_id] );
 423                         }
 424                         $extDb->freeResult( $res );
 425
 426                         // Print errors for missing blobs rows
 427                         foreach ( $oldIds as $blobId => $oldIds2 ) {
 428                                 $this->addError(
 429                                         'restore text',
 430                                         "Error: missing target $cluster/$blobId for two-part ES URL",
 431                                         $oldIds2
 432                                 );
 433                         }
 434                 }
 435         }
 436
 437         function restoreText( $revIds, $xml ) {
 438                 global $wgDBname;
 439                 $tmpDir = wfTempDir();
 440
 441                 if ( !count( $revIds ) ) {
 442                         return;
 443                 }
 444
 445                 print "Restoring text from XML backup...\n";
 446
 447                 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
 448                 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
 449
 450                 // Write revision list
 451                 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
 452                         echo "Error writing revision list, can't restore text\n";
 453
 454                         return;
 455                 }
 456
 457                 // Run mwdumper
 458                 echo "Filtering XML dump...\n";
 459                 $exitStatus = 0;
 460                 passthru( 'mwdumper ' .
 461                         wfEscapeShellArg(
 462                                 "--output=file:$filteredXmlFileName",
 463                                 "--filter=revlist:$revFileName",
 464                                 $xml
 465                         ), $exitStatus
 466                 );
 467
 468                 if ( $exitStatus ) {
 469                         echo "mwdumper died with exit status $exitStatus\n";
 470
 471                         return;
 472                 }
 473
 474                 $file = fopen( $filteredXmlFileName, 'r' );
 475                 if ( !$file ) {
 476                         echo "Unable to open filtered XML file\n";
 477
 478                         return;
 479                 }
 480
 481                 $dbr = wfGetDB( DB_REPLICA );
 482                 $dbw = wfGetDB( DB_MASTER );
 483                 $dbr->ping();
 484                 $dbw->ping();
 485
 486                 $source = new ImportStreamSource( $file );
 487                 $importer = new WikiImporter(
 488                         $source,
 489                         MediaWikiServices::getInstance()->getMainConfig()
 490                 );
 491                 $importer->setRevisionCallback( [ $this, 'importRevision' ] );
 492                 $importer->doImport();
 493         }
 494
 495         function importRevision( &$revision, &$importer ) {
 496                 $id = $revision->getID();
 497                 $content = $revision->getContent( Revision::RAW );
 498                 $id = $id ? $id : '';
 499
 500                 if ( $content === null ) {
 501                         echo "Revision $id is broken, we have no content available\n";
 502
 503                         return;
 504                 }
 505
 506                 $text = $content->serialize();
 507                 if ( $text === '' ) {
 508                         // This is what happens if the revision was broken at the time the
 509                         // dump was made. Unfortunately, it also happens if the revision was
 510                         // legitimately blank, so there's no way to tell the difference. To
 511                         // be safe, we'll skip it and leave it broken
 512
 513                         echo "Revision $id is blank in the dump, may have been broken before export\n";
 514
 515                         return;
 516                 }
 517
 518                 if ( !$id ) {
 519                         // No ID, can't import
 520                         echo "No id tag in revision, can't import\n";
 521
 522                         return;
 523                 }
 524
 525                 // Find text row again
 526                 $dbr = wfGetDB( DB_REPLICA );
 527                 $oldId = $dbr->selectField( 'revision', 'rev_text_id', [ 'rev_id' => $id ], __METHOD__ );
 528                 if ( !$oldId ) {
 529                         echo "Missing revision row for rev_id $id\n";
 530
 531                         return;
 532                 }
 533
 534                 // Compress the text
 535                 $flags = Revision::compressRevisionText( $text );
 536
 537                 // Update the text row
 538                 $dbw = wfGetDB( DB_MASTER );
 539                 $dbw->update( 'text',
 540                         [ 'old_flags' => $flags, 'old_text' => $text ],
 541                         [ 'old_id' => $oldId ],
 542                         __METHOD__, [ 'LIMIT' => 1 ]
 543                 );
 544
 545                 // Remove it from the unfixed list and add it to the fixed list
 546                 unset( $this->errors['restore text'][$id] );
 547                 $this->errors['fixed'][$id] = true;
 548         }
 549 }