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)', false, __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 IN( ' . implode( ',', $blobIds ) . ')' ], __METHOD__
);
212 foreach ( $res as $row ) {
213 unset( $xBlobIds[$row->blob_id
] );
215 $extDb->freeResult( $res );
216 // Print errors for missing blobs rows
217 foreach ( $xBlobIds as $blobId => $oldId ) {
220 "Error: missing target $blobId for one-part ES URL",
226 // Check local objects
230 if ( count( $objectRevs ) ) {
234 [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
235 [ 'old_id' => $objectRevs ],
238 foreach ( $res as $row ) {
239 $oldId = $row->old_id
;
241 if ( !preg_match( '/^O:(\d+):"(\w+)"/', $row->header
, $matches ) ) {
242 $this->addError( 'restore text', "Error: invalid object header", $oldId );
246 $className = strtolower( $matches[2] );
247 if ( strlen( $className ) != $matches[1] ) {
250 "Error: invalid object header, wrong class name length",
256 $objectStats = $objectStats +
[ $className => 0 ];
257 $objectStats[$className]++
;
259 switch ( $className ) {
260 case 'concatenatedgziphistoryblob':
263 case 'historyblobstub':
264 case 'historyblobcurstub':
265 if ( strlen( $row->header
) == $headerLength ) {
266 $this->addError( 'unfixable', "Error: overlong stub header", $oldId );
269 $stubObj = unserialize( $row->header
);
270 if ( !is_object( $stubObj ) ) {
271 $this->addError( 'restore text', "Error: unable to unserialize stub object", $oldId );
274 if ( $className == 'historyblobstub' ) {
275 $concatBlobs[$stubObj->mOldId
][] = $oldId;
277 $curIds[$stubObj->mCurId
][] = $oldId;
281 $this->addError( 'unfixable', "Error: unrecognised object class \"$className\"", $oldId );
284 $dbr->freeResult( $res );
287 // Check local concat blob validity
288 $externalConcatBlobs = [];
289 if ( count( $concatBlobs ) ) {
293 [ 'old_id', 'old_flags', "LEFT(old_text, $headerLength) AS header" ],
294 [ 'old_id' => array_keys( $concatBlobs ) ],
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 ( $urlParts[0] != 'DB:' ) {
306 "Error: unrecognised external storage type \"{$urlParts[0]}",
310 $cluster = $urlParts[2];
312 if ( !isset( $externalConcatBlobs[$cluster][$id] ) ) {
313 $externalConcatBlobs[$cluster][$id] = [];
315 $externalConcatBlobs[$cluster][$id] = array_merge(
316 $externalConcatBlobs[$cluster][$id], $concatBlobs[$row->old_id
]
322 "Error: invalid flags \"{$row->old_flags}\" on concat bulk row {$row->old_id}",
323 $concatBlobs[$row->old_id
] );
325 } elseif ( strcasecmp(
326 substr( $row->header
, 0, strlen( self
::CONCAT_HEADER
) ),
331 "Error: Incorrect object header for concat bulk row {$row->old_id}",
332 $concatBlobs[$row->old_id
]
336 unset( $concatBlobs[$row->old_id
] );
338 $dbr->freeResult( $res );
341 // Check targets of unresolved stubs
342 $this->checkExternalConcatBlobs( $externalConcatBlobs );
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";
354 if ( count( $this->errors
['restore text'] ) && $fix ) {
355 if ( (string)$xml !== '' ) {
356 $this->restoreText( array_keys( $this->errors
['restore text'] ), $xml );
358 echo "Can't fix text, no XML backup specified\n";
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 );
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 );
374 function addError( $type, $msg, $ids ) {
375 if ( is_array( $ids ) && count( $ids ) == 1 ) {
376 $ids = reset( $ids );
378 if ( is_array( $ids ) ) {
380 foreach ( $ids as $id ) {
381 $revIds = array_merge( $revIds, array_keys( $this->oldIdMap
, $id ) );
383 print "$msg in text rows " . implode( ', ', $ids ) .
384 ", revisions " . implode( ', ', $revIds ) . "\n";
387 $revIds = array_keys( $this->oldIdMap
, $id );
388 if ( count( $revIds ) == 1 ) {
389 print "$msg in old_id $id, rev_id {$revIds[0]}\n";
391 print "$msg in old_id $id, revisions " . implode( ', ', $revIds ) . "\n";
394 $this->errors
[$type] = $this->errors
[$type] +
array_flip( $revIds );
397 function checkExternalConcatBlobs( $externalConcatBlobs ) {
398 if ( !count( $externalConcatBlobs ) ) {
402 if ( is_null( $this->dbStore
) ) {
403 $this->dbStore
= new ExternalStoreDB
;
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
) ) {
418 "Error: invalid header on target $cluster/{$row->blob_id} of two-part ES URL",
419 $oldIds[$row->blob_id
]
422 unset( $oldIds[$row->blob_id
] );
424 $extDb->freeResult( $res );
426 // Print errors for missing blobs rows
427 foreach ( $oldIds as $blobId => $oldIds2 ) {
430 "Error: missing target $cluster/$blobId for two-part ES URL",
437 function restoreText( $revIds, $xml ) {
439 $tmpDir = wfTempDir();
441 if ( !count( $revIds ) ) {
445 print "Restoring text from XML backup...\n";
447 $revFileName = "$tmpDir/broken-revlist-$wgDBname";
448 $filteredXmlFileName = "$tmpDir/filtered-$wgDBname.xml";
450 // Write revision list
451 if ( !file_put_contents( $revFileName, implode( "\n", $revIds ) ) ) {
452 echo "Error writing revision list, can't restore text\n";
458 echo "Filtering XML dump...\n";
460 passthru( 'mwdumper ' .
462 "--output=file:$filteredXmlFileName",
463 "--filter=revlist:$revFileName",
469 echo "mwdumper died with exit status $exitStatus\n";
474 $file = fopen( $filteredXmlFileName, 'r' );
476 echo "Unable to open filtered XML file\n";
481 $dbr = wfGetDB( DB_REPLICA
);
482 $dbw = wfGetDB( DB_MASTER
);
486 $source = new ImportStreamSource( $file );
487 $importer = new WikiImporter(
489 MediaWikiServices
::getInstance()->getMainConfig()
491 $importer->setRevisionCallback( [ $this, 'importRevision' ] );
492 $importer->doImport();
495 function importRevision( &$revision, &$importer ) {
496 $id = $revision->getID();
497 $content = $revision->getContent( Revision
::RAW
);
498 $id = $id ?
$id : '';
500 if ( $content === null ) {
501 echo "Revision $id is broken, we have no content available\n";
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
513 echo "Revision $id is blank in the dump, may have been broken before export\n";
519 // No ID, can't import
520 echo "No id tag in revision, can't import\n";
525 // Find text row again
526 $dbr = wfGetDB( DB_REPLICA
);
527 $oldId = $dbr->selectField( 'revision', 'rev_text_id', [ 'rev_id' => $id ], __METHOD__
);
529 echo "Missing revision row for rev_id $id\n";
535 $flags = Revision
::compressRevisionText( $text );
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 ]
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;