includes/externalstore/ExternalStoreDB.php

   1 <?php
   2 /**
   3  * External storage in SQL database.
   4  *
   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.
   9  *
  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.
  14  *
  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
  19  *
  20  * @file
  21  */
  22
  23 use Wikimedia\Rdbms\LBFactory;
  24 use Wikimedia\Rdbms\ILoadBalancer;
  25 use Wikimedia\Rdbms\IDatabase;
  26 use Wikimedia\Rdbms\DBConnRef;
  27 use Wikimedia\Rdbms\MaintainableDBConnRef;
  28 use Wikimedia\Rdbms\DatabaseDomain;
  29
  30 /**
  31  * DB accessible external objects.
  32  *
  33  * In this system, each store "location" maps to a database "cluster".
  34  * The clusters must be defined in the normal LBFactory configuration.
  35  *
  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          * The provided URL is in the form of DB://cluster/id
  57          * or DB://cluster/id/itemid for concatened storage.
  58          *
  59          * @param string $url
  60          * @return string|bool False if missing
  61          * @see ExternalStoreMedium::fetchFromURL()
  62          */
  63         public function fetchFromURL( $url ) {
  64                 list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  65                 $ret = $this->fetchBlob( $cluster, $id, $itemID );
  66
  67                 if ( $itemID !== false && $ret !== false ) {
  68                         return $ret->getItem( $itemID );
  69                 }
  70
  71                 return $ret;
  72         }
  73
  74         /**
  75          * Fetch data from given external store URLs.
  76          * The provided URLs are in the form of DB://cluster/id
  77          * or DB://cluster/id/itemid for concatened storage.
  78          *
  79          * @param array $urls An array of external store URLs
  80          * @return array A map from url to stored content. Failed results
  81          *     are not represented.
  82          */
  83         public function batchFetchFromURLs( array $urls ) {
  84                 $batched = $inverseUrlMap = [];
  85                 foreach ( $urls as $url ) {
  86                         list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  87                         $batched[$cluster][$id][] = $itemID;
  88                         // false $itemID gets cast to int, but should be ok
  89                         // since we do === from the $itemID in $batched
  90                         $inverseUrlMap[$cluster][$id][$itemID] = $url;
  91                 }
  92                 $ret = [];
  93                 foreach ( $batched as $cluster => $batchByCluster ) {
  94                         $res = $this->batchFetchBlobs( $cluster, $batchByCluster );
  95                         /** @var HistoryBlob $blob */
  96                         foreach ( $res as $id => $blob ) {
  97                                 foreach ( $batchByCluster[$id] as $itemID ) {
  98                                         $url = $inverseUrlMap[$cluster][$id][$itemID];
  99                                         if ( $itemID === false ) {
 100                                                 $ret[$url] = $blob;
 101                                         } else {
 102                                                 $ret[$url] = $blob->getItem( $itemID );
 103                                         }
 104                                 }
 105                         }
 106                 }
 107
 108                 return $ret;
 109         }
 110
 111         /**
 112          * @inheritDoc
 113          */
 114         public function store( $location, $data ) {
 115                 $dbw = $this->getMaster( $location );
 116                 $dbw->insert( $this->getTable( $dbw ), [ 'blob_text' => $data ], __METHOD__ );
 117                 $id = $dbw->insertId();
 118                 if ( !$id ) {
 119                         throw new MWException( __METHOD__ . ': no insert ID' );
 120                 }
 121
 122                 return "DB://$location/$id";
 123         }
 124
 125         /**
 126          * @inheritDoc
 127          */
 128         public function isReadOnly( $location ) {
 129                 if ( parent::isReadOnly( $location ) ) {
 130                         return true;
 131                 }
 132
 133                 $lb = $this->getLoadBalancer( $location );
 134                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 135
 136                 return ( $lb->getReadOnlyReason( $domainId ) !== false );
 137         }
 138
 139         /**
 140          * Get a LoadBalancer for the specified cluster
 141          *
 142          * @param string $cluster Cluster name
 143          * @return ILoadBalancer
 144          */
 145         private function getLoadBalancer( $cluster ) {
 146                 return $this->lbFactory->getExternalLB( $cluster );
 147         }
 148
 149         /**
 150          * Get a replica DB connection for the specified cluster
 151          *
 152          * @param string $cluster Cluster name
 153          * @return DBConnRef
 154          */
 155         public function getSlave( $cluster ) {
 156                 $lb = $this->getLoadBalancer( $cluster );
 157                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 158
 159                 $db = $lb->getConnectionRef( DB_REPLICA, [], $domainId );
 160                 $db->clearFlag( DBO_TRX ); // sanity
 161
 162                 return $db;
 163         }
 164
 165         /**
 166          * Get a master database connection for the specified cluster
 167          *
 168          * @param string $cluster Cluster name
 169          * @return MaintainableDBConnRef
 170          */
 171         public function getMaster( $cluster ) {
 172                 $lb = $this->getLoadBalancer( $cluster );
 173                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 174
 175                 $db = $lb->getMaintenanceConnectionRef( DB_MASTER, [], $domainId );
 176                 $db->clearFlag( DBO_TRX ); // sanity
 177
 178                 return $db;
 179         }
 180
 181         /**
 182          * @param array $server Master DB server configuration array for LoadBalancer
 183          * @return string|bool Database domain ID or false
 184          */
 185         private function getDomainId( array $server ) {
 186                 if ( $this->isDbDomainExplicit ) {
 187                         return $this->dbDomain; // explicit foreign domain
 188                 }
 189
 190                 if ( isset( $server['dbname'] ) ) {
 191                         // T200471: for b/c, treat any "dbname" field as forcing which database to use.
 192                         // MediaWiki/LoadBalancer previously did not enforce any concept of a local DB
 193                         // domain, but rather assumed that the LB server configuration matched $wgDBname.
 194                         // This check is useful when the external storage DB for this cluster does not use
 195                         // the same name as the corresponding "main" DB(s) for wikis.
 196                         $domain = new DatabaseDomain(
 197                                 $server['dbname'],
 198                                 $server['schema'] ?? null,
 199                                 $server['tablePrefix'] ?? ''
 200                         );
 201
 202                         return $domain->getId();
 203                 }
 204
 205                 return false; // local LB domain
 206         }
 207
 208         /**
 209          * Get the 'blobs' table name for this database
 210          *
 211          * @param IDatabase $db
 212          * @return string Table name ('blobs' by default)
 213          */
 214         public function getTable( $db ) {
 215                 $table = $db->getLBInfo( 'blobs table' );
 216                 if ( is_null( $table ) ) {
 217                         $table = 'blobs';
 218                 }
 219
 220                 return $table;
 221         }
 222
 223         /**
 224          * Fetch a blob item out of the database; a cache of the last-loaded
 225          * blob will be kept so that multiple loads out of a multi-item blob
 226          * can avoid redundant database access and decompression.
 227          * @param string $cluster
 228          * @param string $id
 229          * @param string $itemID
 230          * @return HistoryBlob|bool Returns false if missing
 231          */
 232         private function fetchBlob( $cluster, $id, $itemID ) {
 233                 /**
 234                  * One-step cache variable to hold base blobs; operations that
 235                  * pull multiple revisions may often pull multiple times from
 236                  * the same blob. By keeping the last-used one open, we avoid
 237                  * redundant unserialization and decompression overhead.
 238                  */
 239                 static $externalBlobCache = [];
 240
 241                 $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
 242                 $cacheID = "$cacheID@{$this->dbDomain}";
 243
 244                 if ( isset( $externalBlobCache[$cacheID] ) ) {
 245                         $this->logger->debug( "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
 246
 247                         return $externalBlobCache[$cacheID];
 248                 }
 249
 250                 $this->logger->debug( "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
 251
 252                 $dbr = $this->getSlave( $cluster );
 253                 $ret = $dbr->selectField( $this->getTable( $dbr ),
 254                         'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 255                 if ( $ret === false ) {
 256                         $this->logger->info( "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
 257                         // Try the master
 258                         $dbw = $this->getMaster( $cluster );
 259                         $ret = $dbw->selectField( $this->getTable( $dbw ),
 260                                 'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 261                         if ( $ret === false ) {
 262                                 $this->logger->error( "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
 263                         }
 264                 }
 265                 if ( $itemID !== false && $ret !== false ) {
 266                         // Unserialise object; caller extracts item
 267                         $ret = unserialize( $ret );
 268                 }
 269
 270                 $externalBlobCache = [ $cacheID => $ret ];
 271
 272                 return $ret;
 273         }
 274
 275         /**
 276          * Fetch multiple blob items out of the database
 277          *
 278          * @param string $cluster A cluster name valid for use with LBFactory
 279          * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs
 280          * @return array A map from the blob_id's requested to their content.
 281          *   Unlocated ids are not represented
 282          */
 283         private function batchFetchBlobs( $cluster, array $ids ) {
 284                 $dbr = $this->getSlave( $cluster );
 285                 $res = $dbr->select(
 286                         $this->getTable( $dbr ),
 287                         [ 'blob_id', 'blob_text' ],
 288                         [ 'blob_id' => array_keys( $ids ) ],
 289                         __METHOD__
 290                 );
 291
 292                 $ret = [];
 293                 if ( $res !== false ) {
 294                         $this->mergeBatchResult( $ret, $ids, $res );
 295                 }
 296                 if ( $ids ) {
 297                         $this->logger->info(
 298                                 __METHOD__ . ": master fallback on '$cluster' for: " .
 299                                 implode( ',', array_keys( $ids ) )
 300                         );
 301                         // Try the master
 302                         $dbw = $this->getMaster( $cluster );
 303                         $res = $dbw->select( $this->getTable( $dbr ),
 304                                 [ 'blob_id', 'blob_text' ],
 305                                 [ 'blob_id' => array_keys( $ids ) ],
 306                                 __METHOD__ );
 307                         if ( $res === false ) {
 308                                 $this->logger->error( __METHOD__ . ": master failed on '$cluster'" );
 309                         } else {
 310                                 $this->mergeBatchResult( $ret, $ids, $res );
 311                         }
 312                 }
 313                 if ( $ids ) {
 314                         $this->logger->error(
 315                                 __METHOD__ . ": master on '$cluster' failed locating items: " .
 316                                 implode( ',', array_keys( $ids ) )
 317                         );
 318                 }
 319
 320                 return $ret;
 321         }
 322
 323         /**
 324          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 325          * @param array &$ret Current self::batchFetchBlobs return value
 326          * @param array &$ids Map from blob_id to requested itemIDs
 327          * @param mixed $res DB result from Database::select
 328          */
 329         private function mergeBatchResult( array &$ret, array &$ids, $res ) {
 330                 foreach ( $res as $row ) {
 331                         $id = $row->blob_id;
 332                         $itemIDs = $ids[$id];
 333                         unset( $ids[$id] ); // to track if everything is found
 334                         if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
 335                                 // single result stored per blob
 336                                 $ret[$id] = $row->blob_text;
 337                         } else {
 338                                 // multi result stored per blob
 339                                 $ret[$id] = unserialize( $row->blob_text );
 340                         }
 341                 }
 342         }
 343
 344         /**
 345          * @param string $url
 346          * @return array
 347          */
 348         protected function parseURL( $url ) {
 349                 $path = explode( '/', $url );
 350
 351                 return [
 352                         $path[2], // cluster
 353                         $path[3], // id
 354                         $path[4] ?? false // itemID
 355                 ];
 356         }
 357 }