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                 $lb = $this->getLoadBalancer( $location );
 110                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 111                 return ( $lb->getReadOnlyReason( $domainId ) !== false );
 112         }
 113
 114         /**
 115          * Get a LoadBalancer for the specified cluster
 116          *
 117          * @param string $cluster Cluster name
 118          * @return ILoadBalancer
 119          */
 120         private function getLoadBalancer( $cluster ) {
 121                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 122                 return $lbFactory->getExternalLB( $cluster );
 123         }
 124
 125         /**
 126          * Get a replica DB connection for the specified cluster
 127          *
 128          * @param string $cluster Cluster name
 129          * @return DBConnRef
 130          */
 131         public function getSlave( $cluster ) {
 132                 global $wgDefaultExternalStore;
 133
 134                 $lb = $this->getLoadBalancer( $cluster );
 135                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 136
 137                 if ( !in_array( "DB://" . $cluster, (array)$wgDefaultExternalStore ) ) {
 138                         wfDebug( "read only external store\n" );
 139                         $lb->allowLagged( true );
 140                 } else {
 141                         wfDebug( "writable external store\n" );
 142                 }
 143
 144                 $db = $lb->getConnectionRef( DB_REPLICA, [], $domainId );
 145                 $db->clearFlag( DBO_TRX ); // sanity
 146
 147                 return $db;
 148         }
 149
 150         /**
 151          * Get a master database connection for the specified cluster
 152          *
 153          * @param string $cluster Cluster name
 154          * @return MaintainableDBConnRef
 155          */
 156         public function getMaster( $cluster ) {
 157                 $lb = $this->getLoadBalancer( $cluster );
 158                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 159
 160                 $db = $lb->getMaintenanceConnectionRef( DB_MASTER, [], $domainId );
 161                 $db->clearFlag( DBO_TRX ); // sanity
 162
 163                 return $db;
 164         }
 165
 166         /**
 167          * @param array $server Master DB server configuration array for LoadBalancer
 168          * @return string|bool Database domain ID or false
 169          */
 170         private function getDomainId( array $server ) {
 171                 if ( isset( $this->params['wiki'] ) && $this->params['wiki'] !== false ) {
 172                         return $this->params['wiki']; // explicit domain
 173                 }
 174
 175                 if ( isset( $server['dbname'] ) ) {
 176                         // T200471: for b/c, treat any "dbname" field as forcing which database to use.
 177                         // MediaWiki/LoadBalancer previously did not enforce any concept of a local DB
 178                         // domain, but rather assumed that the LB server configuration matched $wgDBname.
 179                         // This check is useful when the external storage DB for this cluster does not use
 180                         // the same name as the corresponding "main" DB(s) for wikis.
 181                         $domain = new DatabaseDomain(
 182                                 $server['dbname'],
 183                                 $server['schema'] ?? null,
 184                                 $server['tablePrefix'] ?? ''
 185                         );
 186
 187                         return $domain->getId();
 188                 }
 189
 190                 return false; // local LB domain
 191         }
 192
 193         /**
 194          * Get the 'blobs' table name for this database
 195          *
 196          * @param IDatabase $db
 197          * @return string Table name ('blobs' by default)
 198          */
 199         public function getTable( $db ) {
 200                 $table = $db->getLBInfo( 'blobs table' );
 201                 if ( is_null( $table ) ) {
 202                         $table = 'blobs';
 203                 }
 204
 205                 return $table;
 206         }
 207
 208         /**
 209          * Fetch a blob item out of the database; a cache of the last-loaded
 210          * blob will be kept so that multiple loads out of a multi-item blob
 211          * can avoid redundant database access and decompression.
 212          * @param string $cluster
 213          * @param string $id
 214          * @param string $itemID
 215          * @return HistoryBlob|bool Returns false if missing
 216          */
 217         private function fetchBlob( $cluster, $id, $itemID ) {
 218                 /**
 219                  * One-step cache variable to hold base blobs; operations that
 220                  * pull multiple revisions may often pull multiple times from
 221                  * the same blob. By keeping the last-used one open, we avoid
 222                  * redundant unserialization and decompression overhead.
 223                  */
 224                 static $externalBlobCache = [];
 225
 226                 $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
 227
 228                 $wiki = $this->params['wiki'] ?? false;
 229                 $cacheID = ( $wiki === false ) ? $cacheID : "$cacheID@$wiki";
 230
 231                 if ( isset( $externalBlobCache[$cacheID] ) ) {
 232                         wfDebugLog( 'ExternalStoreDB-cache',
 233                                 "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
 234
 235                         return $externalBlobCache[$cacheID];
 236                 }
 237
 238                 wfDebugLog( 'ExternalStoreDB-cache',
 239                         "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
 240
 241                 $dbr = $this->getSlave( $cluster );
 242                 $ret = $dbr->selectField( $this->getTable( $dbr ),
 243                         'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 244                 if ( $ret === false ) {
 245                         wfDebugLog( 'ExternalStoreDB',
 246                                 "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
 247                         // Try the master
 248                         $dbw = $this->getMaster( $cluster );
 249                         $ret = $dbw->selectField( $this->getTable( $dbw ),
 250                                 'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 251                         if ( $ret === false ) {
 252                                 wfDebugLog( 'ExternalStoreDB',
 253                                         "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
 254                         }
 255                 }
 256                 if ( $itemID !== false && $ret !== false ) {
 257                         // Unserialise object; caller extracts item
 258                         $ret = unserialize( $ret );
 259                 }
 260
 261                 $externalBlobCache = [ $cacheID => $ret ];
 262
 263                 return $ret;
 264         }
 265
 266         /**
 267          * Fetch multiple blob items out of the database
 268          *
 269          * @param string $cluster A cluster name valid for use with LBFactory
 270          * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs
 271          * @return array A map from the blob_id's requested to their content.
 272          *   Unlocated ids are not represented
 273          */
 274         private function batchFetchBlobs( $cluster, array $ids ) {
 275                 $dbr = $this->getSlave( $cluster );
 276                 $res = $dbr->select( $this->getTable( $dbr ),
 277                         [ 'blob_id', 'blob_text' ], [ 'blob_id' => array_keys( $ids ) ], __METHOD__ );
 278                 $ret = [];
 279                 if ( $res !== false ) {
 280                         $this->mergeBatchResult( $ret, $ids, $res );
 281                 }
 282                 if ( $ids ) {
 283                         wfDebugLog( __CLASS__, __METHOD__ .
 284                                 " master fallback on '$cluster' for: " .
 285                                 implode( ',', array_keys( $ids ) ) );
 286                         // Try the master
 287                         $dbw = $this->getMaster( $cluster );
 288                         $res = $dbw->select( $this->getTable( $dbr ),
 289                                 [ 'blob_id', 'blob_text' ],
 290                                 [ 'blob_id' => array_keys( $ids ) ],
 291                                 __METHOD__ );
 292                         if ( $res === false ) {
 293                                 wfDebugLog( __CLASS__, __METHOD__ . " master failed on '$cluster'" );
 294                         } else {
 295                                 $this->mergeBatchResult( $ret, $ids, $res );
 296                         }
 297                 }
 298                 if ( $ids ) {
 299                         wfDebugLog( __CLASS__, __METHOD__ .
 300                                 " master on '$cluster' failed locating items: " .
 301                                 implode( ',', array_keys( $ids ) ) );
 302                 }
 303
 304                 return $ret;
 305         }
 306
 307         /**
 308          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 309          * @param array &$ret Current self::batchFetchBlobs return value
 310          * @param array &$ids Map from blob_id to requested itemIDs
 311          * @param mixed $res DB result from Database::select
 312          */
 313         private function mergeBatchResult( array &$ret, array &$ids, $res ) {
 314                 foreach ( $res as $row ) {
 315                         $id = $row->blob_id;
 316                         $itemIDs = $ids[$id];
 317                         unset( $ids[$id] ); // to track if everything is found
 318                         if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
 319                                 // single result stored per blob
 320                                 $ret[$id] = $row->blob_text;
 321                         } else {
 322                                 // multi result stored per blob
 323                                 $ret[$id] = unserialize( $row->blob_text );
 324                         }
 325                 }
 326         }
 327
 328         /**
 329          * @param string $url
 330          * @return array
 331          */
 332         protected function parseURL( $url ) {
 333                 $path = explode( '/', $url );
 334
 335                 return [
 336                         $path[2], // cluster
 337                         $path[3], // id
 338                         $path[4] ?? false // itemID
 339                 ];
 340         }
 341 }