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