Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 354 |
|
0.00% |
0 / 7 |
CRAP | |
0.00% |
0 / 1 |
| CheckStorage | |
0.00% |
0 / 354 |
|
0.00% |
0 / 7 |
6162 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| execute | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| check | |
0.00% |
0 / 228 |
|
0.00% |
0 / 1 |
2756 | |||
| addError | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
42 | |||
| checkExternalConcatBlobs | |
0.00% |
0 / 29 |
|
0.00% |
0 / 1 |
56 | |||
| restoreText | |
0.00% |
0 / 40 |
|
0.00% |
0 / 1 |
30 | |||
| importRevision | |
0.00% |
0 / 37 |
|
0.00% |
0 / 1 |
42 | |||
| 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\Maintenance\Maintenance; |
| 25 | use MediaWiki\Permissions\UltimateAuthority; |
| 26 | use MediaWiki\Shell\Shell; |
| 27 | use MediaWiki\User\User; |
| 28 | |
| 29 | // @codeCoverageIgnoreStart |
| 30 | require_once __DIR__ . '/../Maintenance.php'; |
| 31 | // @codeCoverageIgnoreEnd |
| 32 | |
| 33 | // ---------------------------------------------------------------------------------- |
| 34 | |
| 35 | /** |
| 36 | * Maintenance script to do various checks on external storage. |
| 37 | * |
| 38 | * @fixme this should extend the base Maintenance class |
| 39 | * @ingroup Maintenance ExternalStorage |
| 40 | */ |
| 41 | class CheckStorage extends Maintenance { |
| 42 | private const CONCAT_HEADER = 'O:27:"concatenatedgziphistoryblob"'; |
| 43 | |
| 44 | public array $oldIdMap; |
| 45 | public array $errors; |
| 46 | |
| 47 | /** @var ExternalStoreDB */ |
| 48 | public $dbStore = null; |
| 49 | |
| 50 | public function __construct() { |
| 51 | parent::__construct(); |
| 52 | |
| 53 | $this->addOption( 'fix', 'Fix errors if possible' ); |
| 54 | $this->addArg( 'xml', 'Path to an XML dump', false ); |
| 55 | } |
| 56 | |
| 57 | public function execute() { |
| 58 | $fix = $this->hasOption( 'fix' ); |
| 59 | $xml = $this->getArg( 'xml', false ); |
| 60 | $this->check( $fix, $xml ); |
| 61 | } |
| 62 | |
| 63 | /** @var string[] */ |
| 64 | public $errorDescriptions = [ |
| 65 | 'restore text' => 'Damaged text, need to be restored from a backup', |
| 66 | 'restore revision' => 'Damaged revision row, need to be restored from a backup', |
| 67 | 'unfixable' => 'Unexpected errors with no automated fixing method', |
| 68 | 'fixed' => 'Errors already fixed', |
| 69 | 'fixable' => 'Errors which would already be fixed if --fix was specified', |
| 70 | ]; |
| 71 | |
| 72 | public function check( $fix = false, $xml = '' ) { |
| 73 | $dbr = $this->getReplicaDB(); |
| 74 | if ( $fix ) { |
| 75 | print "Checking, will fix errors if possible...\n"; |
| 76 | } else { |
| 77 | print "Checking...\n"; |
| 78 | } |
| 79 | $maxRevId = $dbr->newSelectQueryBuilder() |
| 80 | ->select( 'MAX(rev_id)' ) |
| 81 | ->from( 'revision' ) |
| 82 | ->caller( __METHOD__ )->fetchField(); |
| 83 | $chunkSize = 1000; |
| 84 | $flagStats = []; |
| 85 | $objectStats = []; |
| 86 | $knownFlags = [ 'external', 'gzip', 'object', 'utf-8' ]; |
| 87 | $this->errors = [ |
| 88 | 'restore text' => [], |
| 89 | 'restore revision' => [], |
| 90 | 'unfixable' => [], |
| 91 | 'fixed' => [], |
| 92 | 'fixable' => [], |
| 93 | ]; |
| 94 | |
| 95 | for ( $chunkStart = 1; $chunkStart < $maxRevId; $chunkStart += $chunkSize ) { |
| 96 | $chunkEnd = $chunkStart + $chunkSize - 1; |
| 97 | // print "$chunkStart of $maxRevId\n"; |
| 98 | |
| 99 | $this->oldIdMap = []; |
| 100 | $dbr->ping(); |
| 101 | |
| 102 | // Fetch revision rows |
| 103 | $res = $dbr->newSelectQueryBuilder() |
| 104 | ->select( [ 'slot_revision_id', 'content_address' ] ) |
| 105 | ->from( 'slots' ) |
| 106 | ->join( 'content', null, 'content_id = slot_content_id' ) |
| 107 | ->where( [ |
| 108 | $dbr->expr( 'slot_revision_id', '>=', $chunkStart ), |
| 109 | $dbr->expr( 'slot_revision_id', '<=', $chunkEnd ), |
| 110 | ] ) |
| 111 | ->caller( __METHOD__ )->fetchResultSet(); |
| 112 | /** @var \MediaWiki\Storage\SqlBlobStore $blobStore */ |
| 113 | $blobStore = $this->getServiceContainer()->getBlobStore(); |
| 114 | '@phan-var \MediaWiki\Storage\SqlBlobStore $blobStore'; |
| 115 | foreach ( $res as $row ) { |
| 116 | $textId = $blobStore->getTextIdFromAddress( $row->content_address ); |
| 117 | if ( $textId ) { |
| 118 | if ( !isset( $this->oldIdMap[$textId] ) ) { |
| 119 | $this->oldIdMap[ $textId ] = [ $row->slot_revision_id ]; |
| 120 | } elseif ( !in_array( $row->slot_revision_id, $this->oldIdMap[$textId] ) ) { |
| 121 | $this->oldIdMap[ $textId ][] = $row->slot_revision_id; |
| 122 | } |
| 123 | } |
| 124 | } |
| 125 | |
| 126 | if ( !count( $this->oldIdMap ) ) { |
| 127 | continue; |
| 128 | } |
| 129 | |
| 130 | // Fetch old_flags |
| 131 | $missingTextRows = $this->oldIdMap; |
| 132 | $externalRevs = []; |
| 133 | $objectRevs = []; |
| 134 | $res = $dbr->newSelectQueryBuilder() |
| 135 | ->select( [ 'old_id', 'old_flags' ] ) |
| 136 | ->from( 'text' ) |
| 137 | ->where( [ 'old_id' => array_keys( $this->oldIdMap ) ] ) |
| 138 | ->caller( __METHOD__ )->fetchResultSet(); |
| 139 | foreach ( $res as $row ) { |
| 140 | /** |
| 141 | * @var int $flags |
| 142 | */ |
| 143 | $flags = $row->old_flags; |
| 144 | $id = $row->old_id; |
| 145 | |
| 146 | // Create flagStats row if it doesn't exist |
| 147 | $flagStats += [ $flags => 0 ]; |
| 148 | // Increment counter |
| 149 | $flagStats[$flags]++; |
| 150 | |
| 151 | // Not missing |
| 152 | unset( $missingTextRows[$row->old_id] ); |
| 153 | |
| 154 | // Check for external or object |
| 155 | if ( $flags == '' ) { |
| 156 | $flagArray = []; |
| 157 | } else { |
| 158 | $flagArray = explode( ',', $flags ); |
| 159 | } |
| 160 | if ( in_array( 'external', $flagArray ) ) { |
| 161 | $externalRevs[] = $id; |
| 162 | } elseif ( in_array( 'object', $flagArray ) ) { |
| 163 | $objectRevs[] = $id; |
| 164 | } |
| 165 | |
| 166 | // Check for unrecognised flags |
| 167 | if ( $flags == '0' ) { |
| 168 | // This is a known bug from 2004 |
| 169 | // It's safe to just erase the old_flags field |
| 170 | if ( $fix ) { |
| 171 | $this->addError( 'fixed', "Warning: old_flags set to 0", $id ); |
| 172 | $dbw = $this->getPrimaryDB(); |
| 173 | $dbw->ping(); |
| 174 | $dbw->newUpdateQueryBuilder() |
| 175 | ->update( 'text' ) |
| 176 | ->set( [ 'old_flags' => '' ] ) |
| 177 | ->where( [ 'old_id' => $id ] ) |
| 178 | ->caller( __METHOD__ ) |
| 179 | ->execute(); |
| 180 | echo "Fixed\n"; |
| 181 | } else { |
| 182 | $this->addError( 'fixable', "Warning: old_flags set to 0", $id ); |
| 183 | } |
| 184 | } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) { |
| 185 | $this->addError( 'unfixable', "Error: invalid flags field \"$flags\"", $id ); |
| 186 | } |
| 187 | } |
| 188 | |
| 189 | // Output errors for any missing text rows |
| 190 | foreach ( $missingTextRows as $oldId => $revIds ) { |
| 191 | $this->addError( 'restore revision', "Error: missing text row", $oldId ); |
| 192 | } |
| 193 | |
| 194 | // Verify external revisions |
| 195 | $externalConcatBlobs = []; |
| 196 | $externalNormalBlobs = []; |
| 197 | if ( count( $externalRevs ) ) { |
| 198 | $res = $dbr->newSelectQueryBuilder() |
| 199 | ->select( [ 'old_id', 'old_flags', 'old_text' ] ) |
| 200 | ->from( 'text' ) |
| 201 | ->where( [ 'old_id' => $externalRevs ] ) |
| 202 | ->caller( __METHOD__ )->fetchResultSet(); |
| 203 | foreach ( $res as $row ) { |
| 204 | $urlParts = explode( '://', $row->old_text, 2 ); |
| 205 | if ( count( $urlParts ) !== 2 || $urlParts[1] == '' ) { |
| 206 | $this->addError( 'restore text', "Error: invalid URL \"{$row->old_text}\"", $row->old_id ); |
| 207 | continue; |
| 208 | } |
| 209 | [ $proto, ] = $urlParts; |
| 210 | if ( $proto != 'DB' ) { |
| 211 | $this->addError( |
| 212 | 'restore text', |
| 213 | "Error: invalid external protocol \"$proto\"", |
| 214 | $row->old_id ); |
| 215 | continue; |
| 216 | } |
| 217 | $path = explode( '/', $row->old_text ); |
| 218 | $cluster = $path[2]; |
| 219 | $id = $path[3]; |
| 220 | if ( isset( $path[4] ) ) { |
| 221 | $externalConcatBlobs[$cluster][$id][] = $row->old_id; |
| 222 | } else { |
| 223 | $externalNormalBlobs[$cluster][$id][] = $row->old_id; |
| 224 | } |
| 225 | } |
| 226 | } |
| 227 | |
| 228 | // Check external concat blobs for the right header |
| 229 | $this->checkExternalConcatBlobs( $externalConcatBlobs ); |
| 230 | |
| 231 | // Check external normal blobs for existence |
| 232 | if ( count( $externalNormalBlobs ) ) { |
| 233 | if ( $this->dbStore === null ) { |
| 234 | $esFactory = $this->getServiceContainer()->getExternalStoreFactory(); |
| 235 | $this->dbStore = $esFactory->getStore( 'DB' ); |
| 236 | } |
| 237 | foreach ( $externalConcatBlobs as $cluster => $xBlobIds ) { |
| 238 | $blobIds = array_keys( $xBlobIds ); |
| 239 | $extDb = $this->dbStore->getReplica( $cluster ); |
| 240 | $blobsTable = $this->dbStore->getTable( $cluster ); |
| 241 | $res = $extDb->newSelectQueryBuilder() |
| 242 | ->select( [ 'blob_id' ] ) |
| 243 | ->from( $blobsTable ) |
| 244 | ->where( [ 'blob_id' => $blobIds ] ) |
| 245 | ->caller( __METHOD__ )->fetchResultSet(); |
| 246 | foreach ( $res as $row ) { |
| 247 | unset( $xBlobIds[$row->blob_id] ); |
| 248 | } |
| 249 | // Print errors for missing blobs rows |
| 250 | foreach ( $xBlobIds as $blobId => $oldId ) { |
| 251 | $this->addError( |
| 252 | 'restore text', |
| 253 | "Error: missing target $blobId for one-part ES URL", |
| 254 | $oldId ); |
| 255 | } |
| 256 | } |
| 257 | } |
| 258 | |
| 259 | // Check local objects |
| 260 | $dbr->ping(); |
| 261 | $concatBlobs = []; |
| 262 | $curIds = []; |
| 263 | if ( count( $objectRevs ) ) { |
| 264 | $headerLength = 300; |
| 265 | $res = $dbr->newSelectQueryBuilder() |
| 266 | ->select( [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ] ) |
| 267 | ->from( 'text' ) |
| 268 | ->where( [ 'old_id' => $objectRevs ] ) |
| 269 | ->caller( __METHOD__ )->fetchResultSet(); |
| 270 | foreach ( $res as $row ) { |
| 271 | $oldId = $row->old_id; |
| 272 | $matches = []; |
| 273 | if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header, $matches ) ) { |
| 274 | $this->addError( 'restore text', "Error: invalid object header", $oldId ); |
| 275 | continue; |
| 276 | } |
| 277 | |
| 278 | $className = strtolower( $matches[2] ); |
| 279 | if ( strlen( $className ) != $matches[1] ) { |
| 280 | $this->addError( |
| 281 | 'restore text', |
| 282 | "Error: invalid object header, wrong class name length", |
| 283 | $oldId |
| 284 | ); |
| 285 | continue; |
| 286 | } |
| 287 | |
| 288 | $objectStats += [ $className => 0 ]; |
| 289 | $objectStats[$className]++; |
| 290 | |
| 291 | switch ( $className ) { |
| 292 | case 'concatenatedgziphistoryblob': |
| 293 | // Good |
| 294 | break; |
| 295 | case 'historyblobstub': |
| 296 | case 'historyblobcurstub': |
| 297 | if ( strlen( $row->header ) == $headerLength ) { |
| 298 | $this->addError( 'unfixable', "Error: overlong stub header", $oldId ); |
| 299 | break; |
| 300 | } |
| 301 | $stubObj = unserialize( $row->header ); |
| 302 | if ( !is_object( $stubObj ) ) { |
| 303 | $this->addError( 'restore text', "Error: unable to unserialize stub object", $oldId ); |
| 304 | break; |
| 305 | } |
| 306 | if ( $className == 'historyblobstub' ) { |
| 307 | $concatBlobs[$stubObj->getLocation()][] = $oldId; |
| 308 | } else { |
| 309 | $curIds[$stubObj->mCurId][] = $oldId; |
| 310 | } |
| 311 | break; |
| 312 | default: |
| 313 | $this->addError( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId ); |
| 314 | } |
| 315 | } |
| 316 | } |
| 317 | |
| 318 | // Check local concat blob validity |
| 319 | $externalConcatBlobs = []; |
| 320 | if ( count( $concatBlobs ) ) { |
| 321 | $headerLength = 300; |
| 322 | $res = $dbr->newSelectQueryBuilder() |
| 323 | ->select( [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ] ) |
| 324 | ->from( 'text' ) |
| 325 | ->where( [ 'old_id' => array_keys( $concatBlobs ) ] ) |
| 326 | ->caller( __METHOD__ )->fetchResultSet(); |
| 327 | foreach ( $res as $row ) { |
| 328 | $flags = explode( ',', $row->old_flags ); |
| 329 | if ( in_array( 'external', $flags ) ) { |
| 330 | // Concat blob is in external storage? |
| 331 | if ( in_array( 'object', $flags ) ) { |
| 332 | $urlParts = explode( '/', $row->header ); |
| 333 | if ( $urlParts[0] != 'DB:' ) { |
| 334 | $this->addError( |
| 335 | 'unfixable', |
| 336 | "Error: unrecognised external storage type \"{$urlParts[0]}", |
| 337 | $row->old_id |
| 338 | ); |
| 339 | } else { |
| 340 | $cluster = $urlParts[2]; |
| 341 | $id = $urlParts[3]; |
| 342 | if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) { |
| 343 | $externalConcatBlobs[$cluster][$id] = []; |
| 344 | } |
| 345 | $externalConcatBlobs[$cluster][$id] = array_merge( |
| 346 | $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id] |
| 347 | ); |
| 348 | } |
| 349 | } else { |
| 350 | $this->addError( |
| 351 | 'unfixable', |
| 352 | "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}", |
| 353 | $concatBlobs[$row->old_id] ); |
| 354 | } |
| 355 | } elseif ( strcasecmp( |
| 356 | substr( $row->header, 0, strlen( self::CONCAT_HEADER ) ), |
| 357 | self::CONCAT_HEADER |
| 358 | ) ) { |
| 359 | $this->addError( |
| 360 | 'restore text', |
| 361 | "Error: Incorrect object header for concat bulk row {$row->old_id}", |
| 362 | $concatBlobs[$row->old_id] |
| 363 | ); |
| 364 | } |
| 365 | |
| 366 | unset( $concatBlobs[$row->old_id] ); |
| 367 | } |
| 368 | } |
| 369 | |
| 370 | // Check targets of unresolved stubs |
| 371 | $this->checkExternalConcatBlobs( $externalConcatBlobs ); |
| 372 | // next chunk |
| 373 | } |
| 374 | |
| 375 | print "\n\nErrors:\n"; |
| 376 | foreach ( $this->errors as $name => $errors ) { |
| 377 | if ( count( $errors ) ) { |
| 378 | $description = $this->errorDescriptions[$name]; |
| 379 | echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n"; |
| 380 | } |
| 381 | } |
| 382 | |
| 383 | if ( count( $this->errors['restore text'] ) && $fix ) { |
| 384 | if ( (string)$xml !== '' ) { |
| 385 | $this->restoreText( array_keys( $this->errors['restore text'] ), $xml ); |
| 386 | } else { |
| 387 | echo "Can't fix text, no XML backup specified\n"; |
| 388 | } |
| 389 | } |
| 390 | |
| 391 | print "\nFlag statistics:\n"; |
| 392 | $total = array_sum( $flagStats ); |
| 393 | foreach ( $flagStats as $flag => $count ) { |
| 394 | printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 ); |
| 395 | } |
| 396 | print "\nLocal object statistics:\n"; |
| 397 | $total = array_sum( $objectStats ); |
| 398 | foreach ( $objectStats as $className => $count ) { |
| 399 | printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 ); |
| 400 | } |
| 401 | } |
| 402 | |
| 403 | /** |
| 404 | * @param string $type |
| 405 | * @param string $msg |
| 406 | * @param int|int[] $ids |
| 407 | */ |
| 408 | private function addError( string $type, string $msg, $ids ) { |
| 409 | if ( is_array( $ids ) && count( $ids ) == 1 ) { |
| 410 | $ids = reset( $ids ); |
| 411 | } |
| 412 | if ( is_array( $ids ) ) { |
| 413 | $revIds = []; |
| 414 | foreach ( $ids as $id ) { |
| 415 | $revIds = array_unique( array_merge( $revIds, $this->oldIdMap[$id] ) ); |
| 416 | } |
| 417 | print "$msg in text rows " . implode( ', ', $ids ) . |
| 418 | ", revisions " . implode( ', ', $revIds ) . "\n"; |
| 419 | } else { |
| 420 | $id = $ids; |
| 421 | $revIds = $this->oldIdMap[$id]; |
| 422 | if ( count( $revIds ) == 1 ) { |
| 423 | print "$msg in old_id $id, rev_id {$revIds[0]}\n"; |
| 424 | } else { |
| 425 | print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n"; |
| 426 | } |
| 427 | } |
| 428 | $this->errors[$type] += array_fill_keys( $revIds, true ); |
| 429 | } |
| 430 | |
| 431 | private function checkExternalConcatBlobs( array $externalConcatBlobs ) { |
| 432 | if ( !count( $externalConcatBlobs ) ) { |
| 433 | return; |
| 434 | } |
| 435 | |
| 436 | if ( $this->dbStore === null ) { |
| 437 | $esFactory = $this->getServiceContainer()->getExternalStoreFactory(); |
| 438 | $this->dbStore = $esFactory->getStore( 'DB' ); |
| 439 | } |
| 440 | |
| 441 | foreach ( $externalConcatBlobs as $cluster => $oldIds ) { |
| 442 | $blobIds = array_keys( $oldIds ); |
| 443 | $extDb = $this->dbStore->getReplica( $cluster ); |
| 444 | $blobsTable = $this->dbStore->getTable( $cluster ); |
| 445 | $headerLength = strlen( self::CONCAT_HEADER ); |
| 446 | $res = $extDb->newSelectQueryBuilder() |
| 447 | ->select( [ 'blob_id', "LEFT(blob_text, $headerLength) AS header" ] ) |
| 448 | ->from( $blobsTable ) |
| 449 | ->where( [ 'blob_id' => $blobIds ] ) |
| 450 | ->caller( __METHOD__ )->fetchResultSet(); |
| 451 | foreach ( $res as $row ) { |
| 452 | if ( strcasecmp( $row->header, self::CONCAT_HEADER ) ) { |
| 453 | $this->addError( |
| 454 | 'restore text', |
| 455 | "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL", |
| 456 | $oldIds[$row->blob_id] |
| 457 | ); |
| 458 | } |
| 459 | unset( $oldIds[$row->blob_id] ); |
| 460 | } |
| 461 | |
| 462 | // Print errors for missing blobs rows |
| 463 | foreach ( $oldIds as $blobId => $oldIds2 ) { |
| 464 | $this->addError( |
| 465 | 'restore text', |
| 466 | "Error: missing target $cluster/$blobId for two-part ES URL", |
| 467 | $oldIds2 |
| 468 | ); |
| 469 | } |
| 470 | } |
| 471 | } |
| 472 | |
| 473 | private function restoreText( array $revIds, string $xml ) { |
| 474 | global $wgDBname; |
| 475 | $tmpDir = wfTempDir(); |
| 476 | |
| 477 | if ( !count( $revIds ) ) { |
| 478 | return; |
| 479 | } |
| 480 | |
| 481 | print "Restoring text from XML backup...\n"; |
| 482 | |
| 483 | $revFileName = "$tmpDir/broken-revlist-$wgDBname"; |
| 484 | $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml"; |
| 485 | |
| 486 | // Write revision list |
| 487 | if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) { |
| 488 | echo "Error writing revision list, can't restore text\n"; |
| 489 | |
| 490 | return; |
| 491 | } |
| 492 | |
| 493 | // Run mwdumper |
| 494 | echo "Filtering XML dump...\n"; |
| 495 | $exitStatus = 0; |
| 496 | // phpcs:ignore MediaWiki.Usage.ForbiddenFunctions.passthru |
| 497 | passthru( 'mwdumper ' . |
| 498 | Shell::escape( |
| 499 | "--output=file:$filteredXmlFileName", |
| 500 | "--filter=revlist:$revFileName", |
| 501 | $xml |
| 502 | ), $exitStatus |
| 503 | ); |
| 504 | |
| 505 | if ( $exitStatus ) { |
| 506 | echo "mwdumper died with exit status $exitStatus\n"; |
| 507 | |
| 508 | return; |
| 509 | } |
| 510 | |
| 511 | $file = fopen( $filteredXmlFileName, 'r' ); |
| 512 | if ( !$file ) { |
| 513 | echo "Unable to open filtered XML file\n"; |
| 514 | |
| 515 | return; |
| 516 | } |
| 517 | |
| 518 | $dbr = $this->getReplicaDB(); |
| 519 | $dbw = $this->getPrimaryDB(); |
| 520 | $dbr->ping(); |
| 521 | $dbw->ping(); |
| 522 | |
| 523 | $source = new ImportStreamSource( $file ); |
| 524 | $user = User::newSystemUser( User::MAINTENANCE_SCRIPT_USER, [ 'steal' => true ] ); |
| 525 | $importer = $this->getServiceContainer() |
| 526 | ->getWikiImporterFactory() |
| 527 | ->getWikiImporter( $source, new UltimateAuthority( $user ) ); |
| 528 | $importer->setRevisionCallback( [ $this, 'importRevision' ] ); |
| 529 | $importer->setNoticeCallback( static function ( $msg, $params ) { |
| 530 | echo wfMessage( $msg, $params )->text() . "\n"; |
| 531 | } ); |
| 532 | $importer->doImport(); |
| 533 | } |
| 534 | |
| 535 | /** |
| 536 | * @param WikiRevision $revision |
| 537 | */ |
| 538 | public function importRevision( $revision ) { |
| 539 | $id = $revision->getID(); |
| 540 | $content = $revision->getContent(); |
| 541 | $id = $id ?: ''; |
| 542 | |
| 543 | if ( $content === null ) { |
| 544 | echo "Revision $id is broken, we have no content available\n"; |
| 545 | |
| 546 | return; |
| 547 | } |
| 548 | |
| 549 | $text = $content->serialize(); |
| 550 | if ( $text === '' ) { |
| 551 | // This is what happens if the revision was broken at the time the |
| 552 | // dump was made. Unfortunately, it also happens if the revision was |
| 553 | // legitimately blank, so there's no way to tell the difference. To |
| 554 | // be safe, we'll skip it and leave it broken |
| 555 | |
| 556 | echo "Revision $id is blank in the dump, may have been broken before export\n"; |
| 557 | |
| 558 | return; |
| 559 | } |
| 560 | |
| 561 | if ( !$id ) { |
| 562 | // No ID, can't import |
| 563 | echo "No id tag in revision, can't import\n"; |
| 564 | |
| 565 | return; |
| 566 | } |
| 567 | |
| 568 | // Find text row again |
| 569 | $dbr = $this->getReplicaDB(); |
| 570 | $res = $dbr->newSelectQueryBuilder() |
| 571 | ->select( [ 'content_address' ] ) |
| 572 | ->from( 'slots' ) |
| 573 | ->join( 'content', null, 'content_id = slot_content_id' ) |
| 574 | ->where( [ 'slot_revision_id' => $id ] ) |
| 575 | ->caller( __METHOD__ )->fetchRow(); |
| 576 | |
| 577 | $blobStore = $this->getServiceContainer() |
| 578 | ->getBlobStoreFactory() |
| 579 | ->newSqlBlobStore(); |
| 580 | $oldId = $blobStore->getTextIdFromAddress( $res->content_address ); |
| 581 | |
| 582 | if ( !$oldId ) { |
| 583 | echo "Missing revision row for rev_id $id\n"; |
| 584 | return; |
| 585 | } |
| 586 | |
| 587 | // Compress the text |
| 588 | $flags = $blobStore->compressData( $text ); |
| 589 | |
| 590 | // Update the text row |
| 591 | $dbw = $this->getPrimaryDB(); |
| 592 | $dbw->newUpdateQueryBuilder() |
| 593 | ->update( 'text' ) |
| 594 | ->set( [ 'old_flags' => $flags, 'old_text' => $text ] ) |
| 595 | ->where( [ 'old_id' => $oldId ] ) |
| 596 | ->caller( __METHOD__ ) |
| 597 | ->execute(); |
| 598 | |
| 599 | // Remove it from the unfixed list and add it to the fixed list |
| 600 | unset( $this->errors['restore text'][$id] ); |
| 601 | $this->errors['fixed'][$id] = true; |
| 602 | } |
| 603 | |
| 604 | } |
| 605 | |
| 606 | // @codeCoverageIgnoreStart |
| 607 | $maintClass = CheckStorage::class; |
| 608 | require_once RUN_MAINTENANCE_IF_MAIN; |
| 609 | // @codeCoverageIgnoreEnd |