Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 171 |
|
0.00% |
0 / 17 |
CRAP | |
0.00% |
0 / 1 |
| ExternalStoreDB | |
0.00% |
0 / 171 |
|
0.00% |
0 / 17 |
1892 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
12 | |||
| fetchFromURL | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
12 | |||
| batchFetchFromURLs | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
42 | |||
| store | |
0.00% |
0 / 10 |
|
0.00% |
0 / 1 |
6 | |||
| isReadOnly | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
| getLoadBalancer | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getReplica | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
2 | |||
| getPrimary | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
2 | |||
| getDomainId | |
0.00% |
0 / 10 |
|
0.00% |
0 / 1 |
12 | |||
| getTable | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| initializeTable | |
0.00% |
0 / 26 |
|
0.00% |
0 / 1 |
12 | |||
| fetchBlob | |
0.00% |
0 / 31 |
|
0.00% |
0 / 1 |
56 | |||
| batchFetchBlobs | |
0.00% |
0 / 32 |
|
0.00% |
0 / 1 |
12 | |||
| mergeBatchResult | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
20 | |||
| parseURL | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
| getClusterForUrl | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| getDomainIdForCluster | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * This program is free software; you can redistribute it and/or modify |
| 4 | * it under the terms of the GNU General Public License as published by |
| 5 | * the Free Software Foundation; either version 2 of the License, or |
| 6 | * (at your option) any later version. |
| 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 11 | * GNU General Public License for more details. |
| 12 | * |
| 13 | * You should have received a copy of the GNU General Public License along |
| 14 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 15 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 16 | * http://www.gnu.org/copyleft/gpl.html |
| 17 | * |
| 18 | * @file |
| 19 | */ |
| 20 | |
| 21 | use Wikimedia\Rdbms\DatabaseDomain; |
| 22 | use Wikimedia\Rdbms\DBUnexpectedError; |
| 23 | use Wikimedia\Rdbms\ILoadBalancer; |
| 24 | use Wikimedia\Rdbms\LBFactory; |
| 25 | use Wikimedia\Rdbms\Query; |
| 26 | use Wikimedia\Rdbms\ServerInfo; |
| 27 | use Wikimedia\ScopedCallback; |
| 28 | |
| 29 | /** |
| 30 | * External storage in a SQL database. |
| 31 | * |
| 32 | * In this system, each store "location" maps to a database "cluster". |
| 33 | * The clusters must be defined in the normal LBFactory configuration. |
| 34 | * |
| 35 | * @see ExternalStoreAccess |
| 36 | * @ingroup ExternalStorage |
| 37 | */ |
| 38 | class ExternalStoreDB extends ExternalStoreMedium { |
| 39 | /** @var LBFactory */ |
| 40 | private $lbFactory; |
| 41 | |
| 42 | /** |
| 43 | * @see ExternalStoreMedium::__construct() |
| 44 | * @param array $params Additional parameters include: |
| 45 | * - lbFactory: an LBFactory instance |
| 46 | */ |
| 47 | public function __construct( array $params ) { |
| 48 | parent::__construct( $params ); |
| 49 | if ( !isset( $params['lbFactory'] ) || !( $params['lbFactory'] instanceof LBFactory ) ) { |
| 50 | throw new InvalidArgumentException( "LBFactory required in 'lbFactory' field." ); |
| 51 | } |
| 52 | $this->lbFactory = $params['lbFactory']; |
| 53 | } |
| 54 | |
| 55 | /** |
| 56 | * Fetch data from given external store URL. |
| 57 | * |
| 58 | * The provided URL is in the form of `DB://cluster/id` or `DB://cluster/id/itemid` |
| 59 | * for concatenated storage if ConcatenatedGzipHistoryBlob was used. |
| 60 | * |
| 61 | * @param string $url |
| 62 | * @return string|false False if missing |
| 63 | * @see ExternalStoreMedium::fetchFromURL() |
| 64 | */ |
| 65 | public function fetchFromURL( $url ) { |
| 66 | [ $cluster, $id, $itemID ] = $this->parseURL( $url ); |
| 67 | $ret = $this->fetchBlob( $cluster, $id, $itemID ); |
| 68 | |
| 69 | if ( $itemID !== false && $ret !== false ) { |
| 70 | return $ret->getItem( $itemID ); |
| 71 | } |
| 72 | |
| 73 | return $ret; |
| 74 | } |
| 75 | |
| 76 | /** |
| 77 | * Fetch multiple URLs from given external store. |
| 78 | * |
| 79 | * The provided URLs are in the form of `DB://cluster/id`, or `DB://cluster/id/itemid` |
| 80 | * for concatenated storage if ConcatenatedGzipHistoryBlob was used. |
| 81 | * |
| 82 | * @param array $urls An array of external store URLs |
| 83 | * @return array A map from url to stored content. Failed results |
| 84 | * are not represented. |
| 85 | */ |
| 86 | public function batchFetchFromURLs( array $urls ) { |
| 87 | $batched = $inverseUrlMap = []; |
| 88 | foreach ( $urls as $url ) { |
| 89 | [ $cluster, $id, $itemID ] = $this->parseURL( $url ); |
| 90 | $batched[$cluster][$id][] = $itemID; |
| 91 | // false $itemID gets cast to int, but should be ok |
| 92 | // since we do === from the $itemID in $batched |
| 93 | $inverseUrlMap[$cluster][$id][$itemID] = $url; |
| 94 | } |
| 95 | $ret = []; |
| 96 | foreach ( $batched as $cluster => $batchByCluster ) { |
| 97 | $res = $this->batchFetchBlobs( $cluster, $batchByCluster ); |
| 98 | /** @var HistoryBlob $blob */ |
| 99 | foreach ( $res as $id => $blob ) { |
| 100 | foreach ( $batchByCluster[$id] as $itemID ) { |
| 101 | $url = $inverseUrlMap[$cluster][$id][$itemID]; |
| 102 | if ( $itemID === false ) { |
| 103 | $ret[$url] = $blob; |
| 104 | } else { |
| 105 | $ret[$url] = $blob->getItem( $itemID ); |
| 106 | } |
| 107 | } |
| 108 | } |
| 109 | } |
| 110 | |
| 111 | return $ret; |
| 112 | } |
| 113 | |
| 114 | /** |
| 115 | * @inheritDoc |
| 116 | */ |
| 117 | public function store( $location, $data ) { |
| 118 | $blobsTable = $this->getTable( $location ); |
| 119 | |
| 120 | $dbw = $this->getPrimary( $location ); |
| 121 | $dbw->newInsertQueryBuilder() |
| 122 | ->insertInto( $blobsTable ) |
| 123 | ->row( [ 'blob_text' => $data ] ) |
| 124 | ->caller( __METHOD__ )->execute(); |
| 125 | |
| 126 | $id = $dbw->insertId(); |
| 127 | if ( !$id ) { |
| 128 | throw new ExternalStoreException( __METHOD__ . ': no insert ID' ); |
| 129 | } |
| 130 | |
| 131 | return "DB://$location/$id"; |
| 132 | } |
| 133 | |
| 134 | /** |
| 135 | * @inheritDoc |
| 136 | */ |
| 137 | public function isReadOnly( $location ) { |
| 138 | if ( parent::isReadOnly( $location ) ) { |
| 139 | return true; |
| 140 | } |
| 141 | |
| 142 | return ( $this->getLoadBalancer( $location )->getReadOnlyReason() !== false ); |
| 143 | } |
| 144 | |
| 145 | /** |
| 146 | * Get a LoadBalancer for the specified cluster |
| 147 | * |
| 148 | * @param string $cluster Cluster name |
| 149 | * @return ILoadBalancer |
| 150 | */ |
| 151 | private function getLoadBalancer( $cluster ) { |
| 152 | return $this->lbFactory->getExternalLB( $cluster ); |
| 153 | } |
| 154 | |
| 155 | /** |
| 156 | * Get a replica DB connection for the specified cluster |
| 157 | * |
| 158 | * @since 1.34 |
| 159 | * @param string $cluster Cluster name |
| 160 | * @return \Wikimedia\Rdbms\IReadableDatabase |
| 161 | */ |
| 162 | public function getReplica( $cluster ) { |
| 163 | $lb = $this->getLoadBalancer( $cluster ); |
| 164 | |
| 165 | return $lb->getConnection( |
| 166 | DB_REPLICA, |
| 167 | [], |
| 168 | $this->getDomainId( $lb->getServerInfo( ServerInfo::WRITER_INDEX ) ), |
| 169 | $lb::CONN_TRX_AUTOCOMMIT |
| 170 | ); |
| 171 | } |
| 172 | |
| 173 | /** |
| 174 | * Get a primary database connection for the specified cluster |
| 175 | * |
| 176 | * @param string $cluster Cluster name |
| 177 | * @return \Wikimedia\Rdbms\IMaintainableDatabase |
| 178 | * @since 1.37 |
| 179 | */ |
| 180 | public function getPrimary( $cluster ) { |
| 181 | $lb = $this->getLoadBalancer( $cluster ); |
| 182 | |
| 183 | return $lb->getMaintenanceConnectionRef( |
| 184 | DB_PRIMARY, |
| 185 | [], |
| 186 | $this->getDomainId( $lb->getServerInfo( ServerInfo::WRITER_INDEX ) ), |
| 187 | $lb::CONN_TRX_AUTOCOMMIT |
| 188 | ); |
| 189 | } |
| 190 | |
| 191 | /** |
| 192 | * @param array $server Primary DB server configuration array for LoadBalancer |
| 193 | * @return string|false Database domain ID or false |
| 194 | */ |
| 195 | private function getDomainId( array $server ) { |
| 196 | if ( $this->isDbDomainExplicit ) { |
| 197 | return $this->dbDomain; // explicit foreign domain |
| 198 | } |
| 199 | |
| 200 | if ( isset( $server['dbname'] ) ) { |
| 201 | // T200471: for b/c, treat any "dbname" field as forcing which database to use. |
| 202 | // MediaWiki/LoadBalancer previously did not enforce any concept of a local DB |
| 203 | // domain, but rather assumed that the LB server configuration matched $wgDBname. |
| 204 | // This check is useful when the external storage DB for this cluster does not use |
| 205 | // the same name as the corresponding "main" DB(s) for wikis. |
| 206 | $domain = new DatabaseDomain( |
| 207 | $server['dbname'], |
| 208 | $server['schema'] ?? null, |
| 209 | $server['tablePrefix'] ?? '' |
| 210 | ); |
| 211 | |
| 212 | return $domain->getId(); |
| 213 | } |
| 214 | |
| 215 | return false; // local LB domain |
| 216 | } |
| 217 | |
| 218 | /** |
| 219 | * Get the configured blobs table name for this database |
| 220 | * |
| 221 | * Typically, a suffix like "_clusterX" can be used to facilitate clean merging of |
| 222 | * read-only storage clusters by simply cloning tables to the new cluster servers. |
| 223 | * |
| 224 | * @param string $cluster Cluster name |
| 225 | * @return string Unqualified table name (e.g. "blobs_cluster32" or default "blobs") |
| 226 | * @internal Only for use within ExternalStoreDB and its core maintenance scripts |
| 227 | */ |
| 228 | public function getTable( string $cluster ) { |
| 229 | $lb = $this->getLoadBalancer( $cluster ); |
| 230 | $info = $lb->getServerInfo( ServerInfo::WRITER_INDEX ); |
| 231 | |
| 232 | return $info['blobs table'] ?? 'blobs'; |
| 233 | } |
| 234 | |
| 235 | /** |
| 236 | * Create the appropriate blobs table on this cluster |
| 237 | * |
| 238 | * @since 1.34 |
| 239 | * @param string $cluster |
| 240 | */ |
| 241 | public function initializeTable( $cluster ) { |
| 242 | global $IP; |
| 243 | |
| 244 | static $supportedTypes = [ 'mysql', 'sqlite' ]; |
| 245 | |
| 246 | $dbw = $this->getPrimary( $cluster ); |
| 247 | if ( !in_array( $dbw->getType(), $supportedTypes, true ) ) { |
| 248 | throw new DBUnexpectedError( $dbw, "RDBMS type '{$dbw->getType()}' not supported." ); |
| 249 | } |
| 250 | |
| 251 | $sqlFilePath = "$IP/maintenance/storage/blobs.sql"; |
| 252 | $sql = file_get_contents( $sqlFilePath ); |
| 253 | if ( $sql === false ) { |
| 254 | throw new RuntimeException( "Failed to read '$sqlFilePath'." ); |
| 255 | } |
| 256 | |
| 257 | $blobsTable = $this->getTable( $cluster ); |
| 258 | $encTable = $dbw->tableName( $blobsTable ); |
| 259 | $sqlWithReplacedVars = str_replace( |
| 260 | [ '/*$wgDBprefix*/blobs', '/*_*/blobs' ], |
| 261 | [ $encTable, $encTable ], |
| 262 | $sql |
| 263 | ); |
| 264 | |
| 265 | $dbw->query( |
| 266 | new Query( |
| 267 | $sqlWithReplacedVars, |
| 268 | $dbw::QUERY_CHANGE_SCHEMA, |
| 269 | 'CREATE', |
| 270 | $blobsTable, |
| 271 | $sqlWithReplacedVars |
| 272 | ), |
| 273 | __METHOD__ |
| 274 | ); |
| 275 | } |
| 276 | |
| 277 | /** |
| 278 | * Fetch a blob item out of the database; a cache of the last-loaded |
| 279 | * blob will be kept so that multiple loads out of a multi-item blob |
| 280 | * can avoid redundant database access and decompression. |
| 281 | * @param string $cluster |
| 282 | * @param string $id |
| 283 | * @param string $itemID |
| 284 | * @return HistoryBlob|false Returns false if missing |
| 285 | */ |
| 286 | private function fetchBlob( $cluster, $id, $itemID ) { |
| 287 | /** |
| 288 | * One-step cache variable to hold base blobs; operations that |
| 289 | * pull multiple revisions may often pull multiple times from |
| 290 | * the same blob. By keeping the last-used one open, we avoid |
| 291 | * redundant unserialization and decompression overhead. |
| 292 | */ |
| 293 | static $externalBlobCache = []; |
| 294 | |
| 295 | $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/"; |
| 296 | $cacheID = "$cacheID@{$this->dbDomain}"; |
| 297 | |
| 298 | if ( isset( $externalBlobCache[$cacheID] ) ) { |
| 299 | $this->logger->debug( __METHOD__ . ": cache hit on $cacheID" ); |
| 300 | |
| 301 | return $externalBlobCache[$cacheID]; |
| 302 | } |
| 303 | |
| 304 | $this->logger->debug( __METHOD__ . ": cache miss on $cacheID" ); |
| 305 | |
| 306 | $blobsTable = $this->getTable( $cluster ); |
| 307 | |
| 308 | $dbr = $this->getReplica( $cluster ); |
| 309 | $ret = $dbr->newSelectQueryBuilder() |
| 310 | ->select( 'blob_text' ) |
| 311 | ->from( $blobsTable ) |
| 312 | ->where( [ 'blob_id' => $id ] ) |
| 313 | ->caller( __METHOD__ )->fetchField(); |
| 314 | |
| 315 | if ( $ret === false ) { |
| 316 | // Try the primary DB |
| 317 | $this->logger->warning( __METHOD__ . ": primary DB fallback on $cacheID" ); |
| 318 | $trxProfiler = $this->lbFactory->getTransactionProfiler(); |
| 319 | $scope = $trxProfiler->silenceForScope( $trxProfiler::EXPECTATION_REPLICAS_ONLY ); |
| 320 | $dbw = $this->getPrimary( $cluster ); |
| 321 | $ret = $dbw->newSelectQueryBuilder() |
| 322 | ->select( 'blob_text' ) |
| 323 | ->from( $blobsTable ) |
| 324 | ->where( [ 'blob_id' => $id ] ) |
| 325 | ->caller( __METHOD__ )->fetchField(); |
| 326 | ScopedCallback::consume( $scope ); |
| 327 | if ( $ret === false ) { |
| 328 | $this->logger->warning( __METHOD__ . ": primary DB failed to find $cacheID" ); |
| 329 | } |
| 330 | } |
| 331 | if ( $itemID !== false && $ret !== false ) { |
| 332 | // Unserialise object; caller extracts item |
| 333 | $ret = HistoryBlobUtils::unserialize( $ret ); |
| 334 | } |
| 335 | |
| 336 | $externalBlobCache = [ $cacheID => $ret ]; |
| 337 | |
| 338 | return $ret; |
| 339 | } |
| 340 | |
| 341 | /** |
| 342 | * Fetch multiple blob items out of the database |
| 343 | * |
| 344 | * @param string $cluster A cluster name valid for use with LBFactory |
| 345 | * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs |
| 346 | * @return array A map from the blob_id's requested to their content. |
| 347 | * Unlocated ids are not represented |
| 348 | */ |
| 349 | private function batchFetchBlobs( $cluster, array $ids ) { |
| 350 | $blobsTable = $this->getTable( $cluster ); |
| 351 | |
| 352 | $dbr = $this->getReplica( $cluster ); |
| 353 | $res = $dbr->newSelectQueryBuilder() |
| 354 | ->select( [ 'blob_id', 'blob_text' ] ) |
| 355 | ->from( $blobsTable ) |
| 356 | ->where( [ 'blob_id' => array_keys( $ids ) ] ) |
| 357 | ->caller( __METHOD__ ) |
| 358 | ->fetchResultSet(); |
| 359 | |
| 360 | $ret = []; |
| 361 | $this->mergeBatchResult( $ret, $ids, $res ); |
| 362 | if ( $ids ) { |
| 363 | // Try the primary |
| 364 | $this->logger->info( |
| 365 | __METHOD__ . ": primary fallback on '$cluster' for: " . |
| 366 | implode( ',', array_keys( $ids ) ) |
| 367 | ); |
| 368 | $trxProfiler = $this->lbFactory->getTransactionProfiler(); |
| 369 | $scope = $trxProfiler->silenceForScope( $trxProfiler::EXPECTATION_REPLICAS_ONLY ); |
| 370 | $dbw = $this->getPrimary( $cluster ); |
| 371 | $res = $dbw->newSelectQueryBuilder() |
| 372 | ->select( [ 'blob_id', 'blob_text' ] ) |
| 373 | ->from( $blobsTable ) |
| 374 | ->where( [ 'blob_id' => array_keys( $ids ) ] ) |
| 375 | ->caller( __METHOD__ ) |
| 376 | ->fetchResultSet(); |
| 377 | ScopedCallback::consume( $scope ); |
| 378 | $this->mergeBatchResult( $ret, $ids, $res ); |
| 379 | } |
| 380 | if ( $ids ) { |
| 381 | $this->logger->error( |
| 382 | __METHOD__ . ": primary on '$cluster' failed locating items: " . |
| 383 | implode( ',', array_keys( $ids ) ) |
| 384 | ); |
| 385 | } |
| 386 | |
| 387 | return $ret; |
| 388 | } |
| 389 | |
| 390 | /** |
| 391 | * Helper function for self::batchFetchBlobs for merging primary/replica DB results |
| 392 | * @param array &$ret Current self::batchFetchBlobs return value |
| 393 | * @param array &$ids Map from blob_id to requested itemIDs |
| 394 | * @param mixed $res DB result from Database::select |
| 395 | */ |
| 396 | private function mergeBatchResult( array &$ret, array &$ids, $res ) { |
| 397 | foreach ( $res as $row ) { |
| 398 | $id = $row->blob_id; |
| 399 | $itemIDs = $ids[$id]; |
| 400 | unset( $ids[$id] ); // to track if everything is found |
| 401 | if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) { |
| 402 | // single result stored per blob |
| 403 | $ret[$id] = $row->blob_text; |
| 404 | } else { |
| 405 | // multi result stored per blob |
| 406 | $ret[$id] = HistoryBlobUtils::unserialize( $row->blob_text ); |
| 407 | } |
| 408 | } |
| 409 | } |
| 410 | |
| 411 | /** |
| 412 | * @param string $url |
| 413 | * @return array |
| 414 | */ |
| 415 | protected function parseURL( $url ) { |
| 416 | $path = explode( '/', $url ); |
| 417 | |
| 418 | return [ |
| 419 | $path[2], // cluster |
| 420 | $path[3], // id |
| 421 | $path[4] ?? false // itemID |
| 422 | ]; |
| 423 | } |
| 424 | |
| 425 | /** |
| 426 | * Get the cluster part of a URL |
| 427 | * |
| 428 | * @internal for installer |
| 429 | * @param string $url |
| 430 | * @return string|null |
| 431 | */ |
| 432 | public function getClusterForUrl( $url ) { |
| 433 | $parts = explode( '/', $url ); |
| 434 | return $parts[2] ?? null; |
| 435 | } |
| 436 | |
| 437 | /** |
| 438 | * Get the domain ID for a given cluster, which is false for the local wiki ID |
| 439 | * |
| 440 | * @internal for installer |
| 441 | * @param string $cluster |
| 442 | * @return string|false |
| 443 | */ |
| 444 | public function getDomainIdForCluster( $cluster ) { |
| 445 | $lb = $this->getLoadBalancer( $cluster ); |
| 446 | return $this->getDomainId( $lb->getServerInfo( ServerInfo::WRITER_INDEX ) ); |
| 447 | } |
| 448 | } |