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 MediaWiki\MediaWikiServices;
  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         /**
  40          * The provided URL is in the form of DB://cluster/id
  41          * or DB://cluster/id/itemid for concatened storage.
  42          *
  43          * @param string $url
  44          * @return string|bool False if missing
  45          * @see ExternalStoreMedium::fetchFromURL()
  46          */
  47         public function fetchFromURL( $url ) {
  48                 list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  49                 $ret = $this->fetchBlob( $cluster, $id, $itemID );
  50
  51                 if ( $itemID !== false && $ret !== false ) {
  52                         return $ret->getItem( $itemID );
  53                 }
  54
  55                 return $ret;
  56         }
  57
  58         /**
  59          * Fetch data from given external store URLs.
  60          * The provided URLs are in the form of DB://cluster/id
  61          * or DB://cluster/id/itemid for concatened storage.
  62          *
  63          * @param array $urls An array of external store URLs
  64          * @return array A map from url to stored content. Failed results
  65          *     are not represented.
  66          */
  67         public function batchFetchFromURLs( array $urls ) {
  68                 $batched = $inverseUrlMap = [];
  69                 foreach ( $urls as $url ) {
  70                         list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  71                         $batched[$cluster][$id][] = $itemID;
  72                         // false $itemID gets cast to int, but should be ok
  73                         // since we do === from the $itemID in $batched
  74                         $inverseUrlMap[$cluster][$id][$itemID] = $url;
  75                 }
  76                 $ret = [];
  77                 foreach ( $batched as $cluster => $batchByCluster ) {
  78                         $res = $this->batchFetchBlobs( $cluster, $batchByCluster );
  79                         /** @var HistoryBlob $blob */
  80                         foreach ( $res as $id => $blob ) {
  81                                 foreach ( $batchByCluster[$id] as $itemID ) {
  82                                         $url = $inverseUrlMap[$cluster][$id][$itemID];
  83                                         if ( $itemID === false ) {
  84                                                 $ret[$url] = $blob;
  85                                         } else {
  86                                                 $ret[$url] = $blob->getItem( $itemID );
  87                                         }
  88                                 }
  89                         }
  90                 }
  91
  92                 return $ret;
  93         }
  94
  95         public function store( $location, $data ) {
  96                 $dbw = $this->getMaster( $location );
  97                 $dbw->insert( $this->getTable( $dbw ),
  98                         [ 'blob_text' => $data ],
  99                         __METHOD__ );
 100                 $id = $dbw->insertId();
 101                 if ( !$id ) {
 102                         throw new MWException( __METHOD__ . ': no insert ID' );
 103                 }
 104
 105                 return "DB://$location/$id";
 106         }
 107
 108         public function isReadOnly( $location ) {
 109                 return ( $this->getLoadBalancer( $location )->getReadOnlyReason() !== false );
 110         }
 111
 112         /**
 113          * Get a LoadBalancer for the specified cluster
 114          *
 115          * @param string $cluster Cluster name
 116          * @return ILoadBalancer
 117          */
 118         private function getLoadBalancer( $cluster ) {
 119                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 120                 return $lbFactory->getExternalLB( $cluster );
 121         }
 122
 123         /**
 124          * Get a replica DB connection for the specified cluster
 125          *
 126          * @param string $cluster Cluster name
 127          * @return DBConnRef
 128          */
 129         public function getSlave( $cluster ) {
 130                 global $wgDefaultExternalStore;
 131
 132                 $lb = $this->getLoadBalancer( $cluster );
 133                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 134
 135                 if ( !in_array( "DB://" . $cluster, (array)$wgDefaultExternalStore ) ) {
 136                         wfDebug( "read only external store\n" );
 137                         $lb->allowLagged( true );
 138                 } else {
 139                         wfDebug( "writable external store\n" );
 140                 }
 141
 142                 $db = $lb->getConnectionRef( DB_REPLICA, [], $domainId );
 143                 $db->clearFlag( DBO_TRX ); // sanity
 144
 145                 return $db;
 146         }
 147
 148         /**
 149          * Get a master database connection for the specified cluster
 150          *
 151          * @param string $cluster Cluster name
 152          * @return MaintainableDBConnRef
 153          */
 154         public function getMaster( $cluster ) {
 155                 $lb = $this->getLoadBalancer( $cluster );
 156                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 157
 158                 $db = $lb->getMaintenanceConnectionRef( DB_MASTER, [], $domainId );
 159                 $db->clearFlag( DBO_TRX ); // sanity
 160
 161                 return $db;
 162         }
 163
 164         /**
 165          * @param array $server Master DB server configuration array for LoadBalancer
 166          * @return string|bool Database domain ID or false
 167          */
 168         private function getDomainId( array $server ) {
 169                 if ( isset( $this->params['wiki'] ) ) {
 170                         return $this->params['wiki']; // explicit domain
 171                 }
 172
 173                 if ( isset( $server['dbname'] ) ) {
 174                         // T200471: for b/c, treat any "dbname" field as forcing which database to use.
 175                         // MediaWiki/LoadBalancer previously did not enforce any concept of a local DB
 176                         // domain, but rather assumed that the LB server configuration matched $wgDBname.
 177                         // This check is useful when the external storage DB for this cluster does not use
 178                         // the same name as the corresponding "main" DB(s) for wikis.
 179                         $domain = new DatabaseDomain(
 180                                 $server['dbname'],
 181                                 $server['schema'] ?? null,
 182                                 $server['tablePrefix'] ?? ''
 183                         );
 184
 185                         return $domain->getId();
 186                 }
 187
 188                 return false; // local LB domain
 189         }
 190
 191         /**
 192          * Get the 'blobs' table name for this database
 193          *
 194          * @param IDatabase $db
 195          * @return string Table name ('blobs' by default)
 196          */
 197         public function getTable( $db ) {
 198                 $table = $db->getLBInfo( 'blobs table' );
 199                 if ( is_null( $table ) ) {
 200                         $table = 'blobs';
 201                 }
 202
 203                 return $table;
 204         }
 205
 206         /**
 207          * Fetch a blob item out of the database; a cache of the last-loaded
 208          * blob will be kept so that multiple loads out of a multi-item blob
 209          * can avoid redundant database access and decompression.
 210          * @param string $cluster
 211          * @param string $id
 212          * @param string $itemID
 213          * @return HistoryBlob|bool Returns false if missing
 214          */
 215         private function fetchBlob( $cluster, $id, $itemID ) {
 216                 /**
 217                  * One-step cache variable to hold base blobs; operations that
 218                  * pull multiple revisions may often pull multiple times from
 219                  * the same blob. By keeping the last-used one open, we avoid
 220                  * redundant unserialization and decompression overhead.
 221                  */
 222                 static $externalBlobCache = [];
 223
 224                 $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
 225
 226                 $wiki = $this->params['wiki'] ?? false;
 227                 $cacheID = ( $wiki === false ) ? $cacheID : "$cacheID@$wiki";
 228
 229                 if ( isset( $externalBlobCache[$cacheID] ) ) {
 230                         wfDebugLog( 'ExternalStoreDB-cache',
 231                                 "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
 232
 233                         return $externalBlobCache[$cacheID];
 234                 }
 235
 236                 wfDebugLog( 'ExternalStoreDB-cache',
 237                         "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
 238
 239                 $dbr = $this->getSlave( $cluster );
 240                 $ret = $dbr->selectField( $this->getTable( $dbr ),
 241                         'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 242                 if ( $ret === false ) {
 243                         wfDebugLog( 'ExternalStoreDB',
 244                                 "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
 245                         // Try the master
 246                         $dbw = $this->getMaster( $cluster );
 247                         $ret = $dbw->selectField( $this->getTable( $dbw ),
 248                                 'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 249                         if ( $ret === false ) {
 250                                 wfDebugLog( 'ExternalStoreDB',
 251                                         "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
 252                         }
 253                 }
 254                 if ( $itemID !== false && $ret !== false ) {
 255                         // Unserialise object; caller extracts item
 256                         $ret = unserialize( $ret );
 257                 }
 258
 259                 $externalBlobCache = [ $cacheID => $ret ];
 260
 261                 return $ret;
 262         }
 263
 264         /**
 265          * Fetch multiple blob items out of the database
 266          *
 267          * @param string $cluster A cluster name valid for use with LBFactory
 268          * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs
 269          * @return array A map from the blob_id's requested to their content.
 270          *   Unlocated ids are not represented
 271          */
 272         private function batchFetchBlobs( $cluster, array $ids ) {
 273                 $dbr = $this->getSlave( $cluster );
 274                 $res = $dbr->select( $this->getTable( $dbr ),
 275                         [ 'blob_id', 'blob_text' ], [ 'blob_id' => array_keys( $ids ) ], __METHOD__ );
 276                 $ret = [];
 277                 if ( $res !== false ) {
 278                         $this->mergeBatchResult( $ret, $ids, $res );
 279                 }
 280                 if ( $ids ) {
 281                         wfDebugLog( __CLASS__, __METHOD__ .
 282                                 " master fallback on '$cluster' for: " .
 283                                 implode( ',', array_keys( $ids ) ) );
 284                         // Try the master
 285                         $dbw = $this->getMaster( $cluster );
 286                         $res = $dbw->select( $this->getTable( $dbr ),
 287                                 [ 'blob_id', 'blob_text' ],
 288                                 [ 'blob_id' => array_keys( $ids ) ],
 289                                 __METHOD__ );
 290                         if ( $res === false ) {
 291                                 wfDebugLog( __CLASS__, __METHOD__ . " master failed on '$cluster'" );
 292                         } else {
 293                                 $this->mergeBatchResult( $ret, $ids, $res );
 294                         }
 295                 }
 296                 if ( $ids ) {
 297                         wfDebugLog( __CLASS__, __METHOD__ .
 298                                 " master on '$cluster' failed locating items: " .
 299                                 implode( ',', array_keys( $ids ) ) );
 300                 }
 301
 302                 return $ret;
 303         }
 304
 305         /**
 306          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 307          * @param array &$ret Current self::batchFetchBlobs return value
 308          * @param array &$ids Map from blob_id to requested itemIDs
 309          * @param mixed $res DB result from Database::select
 310          */
 311         private function mergeBatchResult( array &$ret, array &$ids, $res ) {
 312                 foreach ( $res as $row ) {
 313                         $id = $row->blob_id;
 314                         $itemIDs = $ids[$id];
 315                         unset( $ids[$id] ); // to track if everything is found
 316                         if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
 317                                 // single result stored per blob
 318                                 $ret[$id] = $row->blob_text;
 319                         } else {
 320                                 // multi result stored per blob
 321                                 $ret[$id] = unserialize( $row->blob_text );
 322                         }
 323                 }
 324         }
 325
 326         /**
 327          * @param string $url
 328          * @return array
 329          */
 330         protected function parseURL( $url ) {
 331                 $path = explode( '/', $url );
 332
 333                 return [
 334                         $path[2], // cluster
 335                         $path[3], // id
 336                         $path[4] ?? false // itemID
 337                 ];
 338         }
 339 }