3 * Adds blobs from a given external storage cluster to the blob_tracking table.
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
22 * @see wfWaitForSlaves()
25 use MediaWiki\MediaWikiServices
;
26 use Wikimedia\Rdbms\DBConnectionError
;
28 require __DIR__
. '/../commandLine.inc';
30 if ( count( $args ) < 1 ) {
31 echo "Usage: php trackBlobs.php <cluster> [... <cluster>]\n";
32 echo "Adds blobs from a given ES cluster to the blob_tracking table\n";
33 echo "Automatically deletes the tracking table and starts from the start again when restarted.\n";
37 $tracker = new TrackBlobs( $args );
42 public $clusters, $textClause;
43 public $doBlobOrphans;
44 public $trackedBlobs = [];
46 public $batchSize = 1000;
47 public $reportingInterval = 10;
49 function __construct( $clusters ) {
50 $this->clusters
= $clusters;
51 if ( extension_loaded( 'gmp' ) ) {
52 $this->doBlobOrphans
= true;
53 foreach ( $clusters as $cluster ) {
54 $this->trackedBlobs
[$cluster] = gmp_init( 0 );
57 echo "Warning: the gmp extension is needed to find orphan blobs\n";
62 $this->checkIntegrity();
63 $this->initTrackingTable();
64 $this->trackRevisions();
65 $this->trackOrphanText();
66 if ( $this->doBlobOrphans
) {
67 $this->findOrphanBlobs();
71 function checkIntegrity() {
72 echo "Doing integrity check...\n";
73 $dbr = wfGetDB( DB_REPLICA
);
75 // Scan for HistoryBlobStub objects in the text table (T22757)
77 $exists = $dbr->selectField( 'text', 1,
78 'old_flags LIKE \'%object%\' AND old_flags NOT LIKE \'%external%\' ' .
79 'AND LOWER(CONVERT(LEFT(old_text,22) USING latin1)) = \'o:15:"historyblobstub"\'',
84 echo "Integrity check failed: found HistoryBlobStub objects in your text table.\n" .
85 "This script could destroy these objects if it continued. Run resolveStubs.php\n" .
90 // Scan the archive table for HistoryBlobStub objects or external flags (T24624)
91 $flags = $dbr->selectField( 'archive', 'ar_flags',
92 'ar_flags LIKE \'%external%\' OR (' .
93 'ar_flags LIKE \'%object%\' ' .
94 'AND LOWER(CONVERT(LEFT(ar_text,22) USING latin1)) = \'o:15:"historyblobstub"\' )',
98 if ( strpos( $flags, 'external' ) !== false ) {
99 echo "Integrity check failed: found external storage pointers in your archive table.\n" .
100 "Run normaliseArchiveTable.php to fix this.\n";
102 } elseif ( $flags ) {
103 echo "Integrity check failed: found HistoryBlobStub objects in your archive table.\n" .
104 "These objects are probably already broken, continuing would make them\n" .
105 "unrecoverable. Run \"normaliseArchiveTable.php --fix-cgz-bug\" to fix this.\n";
109 echo "Integrity check OK\n";
112 function initTrackingTable() {
113 $dbw = wfGetDB( DB_MASTER
);
114 if ( $dbw->tableExists( 'blob_tracking' ) ) {
115 $dbw->query( 'DROP TABLE ' . $dbw->tableName( 'blob_tracking' ) );
116 $dbw->query( 'DROP TABLE ' . $dbw->tableName( 'blob_orphans' ) );
118 $dbw->sourceFile( __DIR__
. '/blob_tracking.sql' );
121 function getTextClause() {
122 if ( !$this->textClause
) {
123 $dbr = wfGetDB( DB_REPLICA
);
124 $this->textClause
= '';
125 foreach ( $this->clusters
as $cluster ) {
126 if ( $this->textClause
!= '' ) {
127 $this->textClause
.= ' OR ';
129 $this->textClause
.= 'old_text' . $dbr->buildLike( "DB://$cluster/", $dbr->anyString() );
133 return $this->textClause
;
136 function interpretPointer( $text ) {
137 if ( !preg_match( '!^DB://(\w+)/(\d+)(?:/([0-9a-fA-F]+)|)$!', $text, $m ) ) {
143 'id' => intval( $m[2] ),
144 'hash' => isset( $m[3] ) ?
$m[3] : null
149 * Scan the revision table for rows stored in the specified clusters
151 function trackRevisions() {
152 $dbw = wfGetDB( DB_MASTER
);
153 $dbr = wfGetDB( DB_REPLICA
);
155 $textClause = $this->getTextClause();
157 $endId = $dbr->selectField( 'revision', 'MAX(rev_id)', '', __METHOD__
);
161 echo "Finding revisions...\n";
164 $res = $dbr->select( [ 'revision', 'text' ],
165 [ 'rev_id', 'rev_page', 'old_id', 'old_flags', 'old_text' ],
167 'rev_id > ' . $dbr->addQuotes( $startId ),
168 'rev_text_id=old_id',
170 'old_flags ' . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ),
174 'ORDER BY' => 'rev_id',
175 'LIMIT' => $this->batchSize
178 if ( !$res->numRows() ) {
183 foreach ( $res as $row ) {
184 $startId = $row->rev_id
;
185 $info = $this->interpretPointer( $row->old_text
);
187 echo "Invalid DB:// URL in rev_id {$row->rev_id}\n";
190 if ( !in_array( $info['cluster'], $this->clusters
) ) {
191 echo "Invalid cluster returned in SQL query: {$info['cluster']}\n";
195 'bt_page' => $row->rev_page
,
196 'bt_rev_id' => $row->rev_id
,
197 'bt_text_id' => $row->old_id
,
198 'bt_cluster' => $info['cluster'],
199 'bt_blob_id' => $info['id'],
200 'bt_cgz_hash' => $info['hash']
202 if ( $this->doBlobOrphans
) {
203 gmp_setbit( $this->trackedBlobs
[$info['cluster']], $info['id'] );
206 $dbw->insert( 'blob_tracking', $insertBatch, __METHOD__
);
207 $rowsInserted +
= count( $insertBatch );
210 if ( $batchesDone >= $this->reportingInterval
) {
212 echo "$startId / $endId\n";
216 echo "Found $rowsInserted revisions\n";
220 * Scan the text table for orphan text
221 * Orphan text here does not imply DB corruption -- deleted text tracked by the
222 * archive table counts as orphan for our purposes.
224 function trackOrphanText() {
225 # Wait until the blob_tracking table is available in the replica DB
226 $dbw = wfGetDB( DB_MASTER
);
227 $dbr = wfGetDB( DB_REPLICA
);
228 $pos = $dbw->getMasterPos();
229 $dbr->masterPosWait( $pos, 100000 );
231 $textClause = $this->getTextClause( $this->clusters
);
233 $endId = $dbr->selectField( 'text', 'MAX(old_id)', '', __METHOD__
);
237 echo "Finding orphan text...\n";
239 # Scan the text table for orphan text
241 $res = $dbr->select( [ 'text', 'blob_tracking' ],
242 [ 'old_id', 'old_flags', 'old_text' ],
244 'old_id>' . $dbr->addQuotes( $startId ),
246 'old_flags ' . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ),
251 'ORDER BY' => 'old_id',
252 'LIMIT' => $this->batchSize
254 [ 'blob_tracking' => [ 'LEFT JOIN', 'bt_text_id=old_id' ] ]
257 foreach ( $res as $row ) {
258 $ids[] = $row->old_id
;
261 if ( !$res->numRows() ) {
266 foreach ( $res as $row ) {
267 $startId = $row->old_id
;
268 $info = $this->interpretPointer( $row->old_text
);
270 echo "Invalid DB:// URL in old_id {$row->old_id}\n";
273 if ( !in_array( $info['cluster'], $this->clusters
) ) {
274 echo "Invalid cluster returned in SQL query\n";
281 'bt_text_id' => $row->old_id
,
282 'bt_cluster' => $info['cluster'],
283 'bt_blob_id' => $info['id'],
284 'bt_cgz_hash' => $info['hash']
286 if ( $this->doBlobOrphans
) {
287 gmp_setbit( $this->trackedBlobs
[$info['cluster']], $info['id'] );
290 $dbw->insert( 'blob_tracking', $insertBatch, __METHOD__
);
292 $rowsInserted +
= count( $insertBatch );
294 if ( $batchesDone >= $this->reportingInterval
) {
296 echo "$startId / $endId\n";
300 echo "Found $rowsInserted orphan text rows\n";
304 * Scan the blobs table for rows not registered in blob_tracking (and thus not
305 * registered in the text table).
307 * Orphan blobs are indicative of DB corruption. They are inaccessible and
308 * should probably be deleted.
310 function findOrphanBlobs() {
311 if ( !extension_loaded( 'gmp' ) ) {
312 echo "Can't find orphan blobs, need bitfield support provided by GMP.\n";
317 $dbw = wfGetDB( DB_MASTER
);
319 foreach ( $this->clusters
as $cluster ) {
320 echo "Searching for orphan blobs in $cluster...\n";
321 $lbFactory = MediaWikiServices
::getInstance()->getDBLoadBalancerFactory();
322 $lb = $lbFactory->getExternalLB( $cluster );
324 $extDB = $lb->getConnection( DB_REPLICA
);
325 } catch ( DBConnectionError
$e ) {
326 if ( strpos( $e->error
, 'Unknown database' ) !== false ) {
327 echo "No database on $cluster\n";
329 echo "Error on $cluster: " . $e->getMessage() . "\n";
333 $table = $extDB->getLBInfo( 'blobs table' );
334 if ( is_null( $table ) ) {
337 if ( !$extDB->tableExists( $table ) ) {
338 echo "No blobs table on cluster $cluster\n";
343 $actualBlobs = gmp_init( 0 );
344 $endId = $extDB->selectField( $table, 'MAX(blob_id)', '', __METHOD__
);
346 // Build a bitmap of actual blob rows
348 $res = $extDB->select( $table,
350 [ 'blob_id > ' . $extDB->addQuotes( $startId ) ],
352 [ 'LIMIT' => $this->batchSize
, 'ORDER BY' => 'blob_id' ]
355 if ( !$res->numRows() ) {
359 foreach ( $res as $row ) {
360 gmp_setbit( $actualBlobs, $row->blob_id
);
362 $startId = $row->blob_id
;
365 if ( $batchesDone >= $this->reportingInterval
) {
367 echo "$startId / $endId\n";
371 // Find actual blobs that weren't tracked by the previous passes
372 // This is a set-theoretic difference A \ B, or in bitwise terms, A & ~B
373 $orphans = gmp_and( $actualBlobs, gmp_com( $this->trackedBlobs
[$cluster] ) );
375 // Traverse the orphan list
380 $id = gmp_scan1( $orphans, $id );
385 'bo_cluster' => $cluster,
388 if ( count( $insertBatch ) > $this->batchSize
) {
389 $dbw->insert( 'blob_orphans', $insertBatch, __METHOD__
);
396 if ( $insertBatch ) {
397 $dbw->insert( 'blob_orphans', $insertBatch, __METHOD__
);
399 echo "Found $numOrphans orphan(s) in $cluster\n";