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.
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.
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
21 * @ingroup Maintenance ExternalStorage
24 use MediaWiki\MediaWikiServices
;
26 if ( !defined( 'MEDIAWIKI' ) ) {
27 $optionsWithoutArgs = [ 'fix' ];
28 require_once __DIR__
. '/../commandLine.inc';
30 $cs = new CheckStorage
;
31 $fix = isset( $options['fix'] );
32 if ( isset( $args[0] ) ) {
37 $cs->check( $fix, $xml );
40 // ----------------------------------------------------------------------------------
43 * Maintenance script to do various checks on external storage.
45 * @fixme this should extend the base Maintenance class
46 * @ingroup Maintenance ExternalStorage
49 const CONCAT_HEADER
= 'O:27:"concatenatedgziphistoryblob"';
50 public $oldIdMap, $errors;
51 public $dbStore = null;
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',
61 function check( $fix = false, $xml = '' ) {
62 $dbr = wfGetDB( DB_REPLICA
);
64 print "Checking, will fix errors if possible...\n";
66 print "Checking...\n";
68 $maxRevId = $dbr->selectField( 'revision', 'MAX(rev_id)', '', __METHOD__
);
72 $knownFlags = [ 'external', 'gzip', 'object', 'utf-8' ];
75 'restore revision' => [],
81 for ( $chunkStart = 1; $chunkStart < $maxRevId; $chunkStart +
= $chunkSize ) {
82 $chunkEnd = $chunkStart +
$chunkSize - 1;
83 // print "$chunkStart of $maxRevId\n";
85 // Fetch revision rows
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
;
93 $dbr->freeResult( $res );
95 if ( !count( $this->oldIdMap
) ) {
100 $missingTextRows = array_flip( $this->oldIdMap
);
105 [ 'old_id', 'old_flags' ],
106 [ 'old_id' => $this->oldIdMap
],
109 foreach ( $res as $row ) {
113 $flags = $row->old_flags
;
116 // Create flagStats row if it doesn't exist
117 $flagStats = $flagStats +
[ $flags => 0 ];
119 $flagStats[$flags]++
;
122 unset( $missingTextRows[$row->old_id
] );
124 // Check for external or object
125 if ( $flags == '' ) {
128 $flagArray = explode( ',', $flags );
130 if ( in_array( 'external', $flagArray ) ) {
131 $externalRevs[] = $id;
132 } elseif ( in_array( 'object', $flagArray ) ) {
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
141 $this->addError( 'fixed', "Warning: old_flags set to 0", $id );
142 $dbw = wfGetDB( DB_MASTER
);
144 $dbw->update( 'text', [ 'old_flags' => '' ],
145 [ 'old_id' => $id ], __METHOD__
);
148 $this->addError( 'fixable', "Warning: old_flags set to 0", $id );
150 } elseif ( count( array_diff( $flagArray, $knownFlags ) ) ) {
151 $this->addError( 'unfixable', "Error: invalid flags field \"$flags\"", $id );
154 $dbr->freeResult( $res );
156 // Output errors for any missing text rows
157 foreach ( $missingTextRows as $oldId => $revId ) {
158 $this->addError( 'restore revision', "Error: missing text row", $oldId );
161 // Verify external revisions
162 $externalConcatBlobs = [];
163 $externalNormalBlobs = [];
164 if ( count( $externalRevs ) ) {
167 [ 'old_id', 'old_flags', 'old_text' ],
168 [ 'old_id' => $externalRevs ],
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
);
177 list( $proto, ) = $urlParts;
178 if ( $proto != 'DB' ) {
181 "Error: invalid external protocol \"$proto\"",
185 $path = explode( '/', $row->old_text
);
188 if ( isset( $path[4] ) ) {
189 $externalConcatBlobs[$cluster][$id][] = $row->old_id
;
191 $externalNormalBlobs[$cluster][$id][] = $row->old_id
;
194 $dbr->freeResult( $res );
197 // Check external concat blobs for the right header
198 $this->checkExternalConcatBlobs( $externalConcatBlobs );
200 // Check external normal blobs for existence
201 if ( count( $externalNormalBlobs ) ) {
202 if ( is_null( $this->dbStore
) ) {
203 $this->dbStore
= new ExternalStoreDB
;
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,
211 [ 'blob_id' => $blobIds ],
214 foreach ( $res as $row ) {
215 unset( $xBlobIds[$row->blob_id
] );
217 $extDb->freeResult( $res );
218 // Print errors for missing blobs rows
219 foreach ( $xBlobIds as $blobId => $oldId ) {
222 "Error: missing target $blobId for one-part ES URL",
228 // Check local objects
232 if ( count( $objectRevs ) ) {
236 [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
237 [ 'old_id' => $objectRevs ],
240 foreach ( $res as $row ) {
241 $oldId = $row->old_id
;
243 if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header
, $matches ) ) {
244 $this->addError( 'restore text', "Error: invalid object header", $oldId );
248 $className = strtolower( $matches[2] );
249 if ( strlen( $className ) != $matches[1] ) {
252 "Error: invalid object header, wrong class name length",
258 $objectStats = $objectStats +
[ $className => 0 ];
259 $objectStats[$className]++
;
261 switch ( $className ) {
262 case 'concatenatedgziphistoryblob':
265 case 'historyblobstub':
266 case 'historyblobcurstub':
267 if ( strlen( $row->header
) == $headerLength ) {
268 $this->addError( 'unfixable', "Error: overlong stub header", $oldId );
271 $stubObj = unserialize( $row->header
);
272 if ( !is_object( $stubObj ) ) {
273 $this->addError( 'restore text', "Error: unable to unserialize stub object", $oldId );
276 if ( $className == 'historyblobstub' ) {
277 $concatBlobs[$stubObj->mOldId
][] = $oldId;
279 $curIds[$stubObj->mCurId
][] = $oldId;
283 $this->addError( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
286 $dbr->freeResult( $res );
289 // Check local concat blob validity
290 $externalConcatBlobs = [];
291 if ( count( $concatBlobs ) ) {
295 [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
296 [ 'old_id' => array_keys( $concatBlobs ) ],
299 foreach ( $res as $row ) {
300 $flags = explode( ',', $row->old_flags
);
301 if ( in_array( 'external', $flags ) ) {
302 // Concat blob is in external storage?
303 if ( in_array( 'object', $flags ) ) {
304 $urlParts = explode( '/', $row->header
);
305 if ( $urlParts[0] != 'DB:' ) {
308 "Error: unrecognised external storage type \"{$urlParts[0]}",
312 $cluster = $urlParts[2];
314 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
315 $externalConcatBlobs[$cluster][$id] = [];
317 $externalConcatBlobs[$cluster][$id] = array_merge(
318 $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id
]
324 "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
325 $concatBlobs[$row->old_id
] );
327 } elseif ( strcasecmp(
328 substr( $row->header
, 0, strlen( self
::CONCAT_HEADER
) ),
333 "Error: Incorrect object header for concat bulk row {$row->old_id}",
334 $concatBlobs[$row->old_id
]
338 unset( $concatBlobs[$row->old_id
] );
340 $dbr->freeResult( $res );
343 // Check targets of unresolved stubs
344 $this->checkExternalConcatBlobs( $externalConcatBlobs );
348 print "\n\nErrors:\n";
349 foreach ( $this->errors
as $name => $errors ) {
350 if ( count( $errors ) ) {
351 $description = $this->errorDescriptions
[$name];
352 echo "$description: " . implode( ',', array_keys( $errors ) ) . "\n";
356 if ( count( $this->errors
['restore text'] ) && $fix ) {
357 if ( (string)$xml !== '' ) {
358 $this->restoreText( array_keys( $this->errors
['restore text'] ), $xml );
360 echo "Can't fix text, no XML backup specified\n";
364 print "\nFlag statistics:\n";
365 $total = array_sum( $flagStats );
366 foreach ( $flagStats as $flag => $count ) {
367 printf( "%-30s %10d %5.2f%%\n", $flag, $count, $count / $total * 100 );
369 print "\nLocal object statistics:\n";
370 $total = array_sum( $objectStats );
371 foreach ( $objectStats as $className => $count ) {
372 printf( "%-30s %10d %5.2f%%\n", $className, $count, $count / $total * 100 );
376 function addError( $type, $msg, $ids ) {
377 if ( is_array( $ids ) && count( $ids ) == 1 ) {
378 $ids = reset( $ids );
380 if ( is_array( $ids ) ) {
382 foreach ( $ids as $id ) {
383 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap
, $id ) );
385 print "$msg in text rows " . implode( ', ', $ids ) .
386 ", revisions " . implode( ', ', $revIds ) . "\n";
389 $revIds = array_keys( $this->oldIdMap
, $id );
390 if ( count( $revIds ) == 1 ) {
391 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
393 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
396 $this->errors
[$type] = $this->errors
[$type] +
array_flip( $revIds );
399 function checkExternalConcatBlobs( $externalConcatBlobs ) {
400 if ( !count( $externalConcatBlobs ) ) {
404 if ( is_null( $this->dbStore
) ) {
405 $this->dbStore
= new ExternalStoreDB
;
408 foreach ( $externalConcatBlobs as $cluster => $oldIds ) {
409 $blobIds = array_keys( $oldIds );
410 $extDb =& $this->dbStore
->getSlave( $cluster );
411 $blobsTable = $this->dbStore
->getTable( $extDb );
412 $headerLength = strlen( self
::CONCAT_HEADER
);
413 $res = $extDb->select( $blobsTable,
414 [ 'blob_id', "LEFT(blob_text, $headerLength) AS header" ],
415 [ 'blob_id' => $blobIds ],
418 foreach ( $res as $row ) {
419 if ( strcasecmp( $row->header
, self
::CONCAT_HEADER
) ) {
422 "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
423 $oldIds[$row->blob_id
]
426 unset( $oldIds[$row->blob_id
] );
428 $extDb->freeResult( $res );
430 // Print errors for missing blobs rows
431 foreach ( $oldIds as $blobId => $oldIds2 ) {
434 "Error: missing target $cluster/$blobId for two-part ES URL",
441 function restoreText( $revIds, $xml ) {
443 $tmpDir = wfTempDir();
445 if ( !count( $revIds ) ) {
449 print "Restoring text from XML backup...\n";
451 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
452 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
454 // Write revision list
455 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
456 echo "Error writing revision list, can't restore text\n";
462 echo "Filtering XML dump...\n";
464 passthru( 'mwdumper ' .
466 "--output=file:$filteredXmlFileName",
467 "--filter=revlist:$revFileName",
473 echo "mwdumper died with exit status $exitStatus\n";
478 $file = fopen( $filteredXmlFileName, 'r' );
480 echo "Unable to open filtered XML file\n";
485 $dbr = wfGetDB( DB_REPLICA
);
486 $dbw = wfGetDB( DB_MASTER
);
490 $source = new ImportStreamSource( $file );
491 $importer = new WikiImporter(
493 MediaWikiServices
::getInstance()->getMainConfig()
495 $importer->setRevisionCallback( [ $this, 'importRevision' ] );
496 $importer->setNoticeCallback( function ( $msg, $params ) {
497 echo wfMessage( $msg, $params )->text() . "\n";
499 $importer->doImport();
502 function importRevision( &$revision, &$importer ) {
503 $id = $revision->getID();
504 $content = $revision->getContent( Revision
::RAW
);
507 if ( $content === null ) {
508 echo "Revision $id is broken, we have no content available\n";
513 $text = $content->serialize();
514 if ( $text === '' ) {
515 // This is what happens if the revision was broken at the time the
516 // dump was made. Unfortunately, it also happens if the revision was
517 // legitimately blank, so there's no way to tell the difference. To
518 // be safe, we'll skip it and leave it broken
520 echo "Revision $id is blank in the dump, may have been broken before export\n";
526 // No ID, can't import
527 echo "No id tag in revision, can't import\n";
532 // Find text row again
533 $dbr = wfGetDB( DB_REPLICA
);
534 $oldId = $dbr->selectField( 'revision', 'rev_text_id', [ 'rev_id' => $id ], __METHOD__
);
536 echo "Missing revision row for rev_id $id\n";
542 $flags = Revision
::compressRevisionText( $text );
544 // Update the text row
545 $dbw = wfGetDB( DB_MASTER
);
546 $dbw->update( 'text',
547 [ 'old_flags' => $flags, 'old_text' => $text ],
548 [ 'old_id' => $oldId ],
549 __METHOD__
, [ 'LIMIT' => 1 ]
552 // Remove it from the unfixed list and add it to the fixed list
553 unset( $this->errors
['restore text'][$id] );
554 $this->errors
['fixed'][$id] = true;