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 echo "Integrity check OK\n";
93 function initTrackingTable() {
94 $dbw = wfGetDB( DB_MASTER
);
95 if ( $dbw->tableExists( 'blob_tracking' ) ) {
96 $dbw->query( 'DROP TABLE ' . $dbw->tableName( 'blob_tracking' ) );
97 $dbw->query( 'DROP TABLE ' . $dbw->tableName( 'blob_orphans' ) );
99 $dbw->sourceFile( __DIR__
. '/blob_tracking.sql' );
102 function getTextClause() {
103 if ( !$this->textClause
) {
104 $dbr = wfGetDB( DB_REPLICA
);
105 $this->textClause
= '';
106 foreach ( $this->clusters
as $cluster ) {
107 if ( $this->textClause
!= '' ) {
108 $this->textClause
.= ' OR ';
110 $this->textClause
.= 'old_text' . $dbr->buildLike( "DB://$cluster/", $dbr->anyString() );
114 return $this->textClause
;
117 function interpretPointer( $text ) {
118 if ( !preg_match( '!^DB://(\w+)/(\d+)(?:/([0-9a-fA-F]+)|)$!', $text, $m ) ) {
124 'id' => intval( $m[2] ),
125 'hash' => $m[3] ??
null
130 * Scan the revision table for rows stored in the specified clusters
132 function trackRevisions() {
133 $dbw = wfGetDB( DB_MASTER
);
134 $dbr = wfGetDB( DB_REPLICA
);
136 $textClause = $this->getTextClause();
138 $endId = $dbr->selectField( 'revision', 'MAX(rev_id)', '', __METHOD__
);
142 echo "Finding revisions...\n";
145 $res = $dbr->select( [ 'revision', 'text' ],
146 [ 'rev_id', 'rev_page', 'old_id', 'old_flags', 'old_text' ],
148 'rev_id > ' . $dbr->addQuotes( $startId ),
149 'rev_text_id=old_id',
151 'old_flags ' . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ),
155 'ORDER BY' => 'rev_id',
156 'LIMIT' => $this->batchSize
159 if ( !$res->numRows() ) {
164 foreach ( $res as $row ) {
165 $startId = $row->rev_id
;
166 $info = $this->interpretPointer( $row->old_text
);
168 echo "Invalid DB:// URL in rev_id {$row->rev_id}\n";
171 if ( !in_array( $info['cluster'], $this->clusters
) ) {
172 echo "Invalid cluster returned in SQL query: {$info['cluster']}\n";
176 'bt_page' => $row->rev_page
,
177 'bt_rev_id' => $row->rev_id
,
178 'bt_text_id' => $row->old_id
,
179 'bt_cluster' => $info['cluster'],
180 'bt_blob_id' => $info['id'],
181 'bt_cgz_hash' => $info['hash']
183 if ( $this->doBlobOrphans
) {
184 gmp_setbit( $this->trackedBlobs
[$info['cluster']], $info['id'] );
187 $dbw->insert( 'blob_tracking', $insertBatch, __METHOD__
);
188 $rowsInserted +
= count( $insertBatch );
191 if ( $batchesDone >= $this->reportingInterval
) {
193 echo "$startId / $endId\n";
197 echo "Found $rowsInserted revisions\n";
201 * Scan the text table for orphan text
202 * Orphan text here does not imply DB corruption -- deleted text tracked by the
203 * archive table counts as orphan for our purposes.
205 function trackOrphanText() {
206 # Wait until the blob_tracking table is available in the replica DB
207 $dbw = wfGetDB( DB_MASTER
);
208 $dbr = wfGetDB( DB_REPLICA
);
209 $pos = $dbw->getMasterPos();
210 $dbr->masterPosWait( $pos, 100000 );
212 $textClause = $this->getTextClause( $this->clusters
);
214 $endId = $dbr->selectField( 'text', 'MAX(old_id)', '', __METHOD__
);
218 echo "Finding orphan text...\n";
220 # Scan the text table for orphan text
222 $res = $dbr->select( [ 'text', 'blob_tracking' ],
223 [ 'old_id', 'old_flags', 'old_text' ],
225 'old_id>' . $dbr->addQuotes( $startId ),
227 'old_flags ' . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ),
232 'ORDER BY' => 'old_id',
233 'LIMIT' => $this->batchSize
235 [ 'blob_tracking' => [ 'LEFT JOIN', 'bt_text_id=old_id' ] ]
238 foreach ( $res as $row ) {
239 $ids[] = $row->old_id
;
242 if ( !$res->numRows() ) {
247 foreach ( $res as $row ) {
248 $startId = $row->old_id
;
249 $info = $this->interpretPointer( $row->old_text
);
251 echo "Invalid DB:// URL in old_id {$row->old_id}\n";
254 if ( !in_array( $info['cluster'], $this->clusters
) ) {
255 echo "Invalid cluster returned in SQL query\n";
262 'bt_text_id' => $row->old_id
,
263 'bt_cluster' => $info['cluster'],
264 'bt_blob_id' => $info['id'],
265 'bt_cgz_hash' => $info['hash']
267 if ( $this->doBlobOrphans
) {
268 gmp_setbit( $this->trackedBlobs
[$info['cluster']], $info['id'] );
271 $dbw->insert( 'blob_tracking', $insertBatch, __METHOD__
);
273 $rowsInserted +
= count( $insertBatch );
275 if ( $batchesDone >= $this->reportingInterval
) {
277 echo "$startId / $endId\n";
281 echo "Found $rowsInserted orphan text rows\n";
285 * Scan the blobs table for rows not registered in blob_tracking (and thus not
286 * registered in the text table).
288 * Orphan blobs are indicative of DB corruption. They are inaccessible and
289 * should probably be deleted.
291 function findOrphanBlobs() {
292 if ( !extension_loaded( 'gmp' ) ) {
293 echo "Can't find orphan blobs, need bitfield support provided by GMP.\n";
298 $dbw = wfGetDB( DB_MASTER
);
300 foreach ( $this->clusters
as $cluster ) {
301 echo "Searching for orphan blobs in $cluster...\n";
302 $lbFactory = MediaWikiServices
::getInstance()->getDBLoadBalancerFactory();
303 $lb = $lbFactory->getExternalLB( $cluster );
305 $extDB = $lb->getConnection( DB_REPLICA
);
306 } catch ( DBConnectionError
$e ) {
307 if ( strpos( $e->error
, 'Unknown database' ) !== false ) {
308 echo "No database on $cluster\n";
310 echo "Error on $cluster: " . $e->getMessage() . "\n";
314 $table = $extDB->getLBInfo( 'blobs table' );
315 if ( is_null( $table ) ) {
318 if ( !$extDB->tableExists( $table ) ) {
319 echo "No blobs table on cluster $cluster\n";
324 $actualBlobs = gmp_init( 0 );
325 $endId = $extDB->selectField( $table, 'MAX(blob_id)', '', __METHOD__
);
327 // Build a bitmap of actual blob rows
329 $res = $extDB->select( $table,
331 [ 'blob_id > ' . $extDB->addQuotes( $startId ) ],
333 [ 'LIMIT' => $this->batchSize
, 'ORDER BY' => 'blob_id' ]
336 if ( !$res->numRows() ) {
340 foreach ( $res as $row ) {
341 gmp_setbit( $actualBlobs, $row->blob_id
);
343 $startId = $row->blob_id
;
346 if ( $batchesDone >= $this->reportingInterval
) {
348 echo "$startId / $endId\n";
352 // Find actual blobs that weren't tracked by the previous passes
353 // This is a set-theoretic difference A \ B, or in bitwise terms, A & ~B
354 $orphans = gmp_and( $actualBlobs, gmp_com( $this->trackedBlobs
[$cluster] ) );
356 // Traverse the orphan list
361 $id = gmp_scan1( $orphans, $id );
366 'bo_cluster' => $cluster,
369 if ( count( $insertBatch ) > $this->batchSize
) {
370 $dbw->insert( 'blob_orphans', $insertBatch, __METHOD__
);
377 if ( $insertBatch ) {
378 $dbw->insert( 'blob_orphans', $insertBatch, __METHOD__
);
380 echo "Found $numOrphans orphan(s) in $cluster\n";