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\LoadBalancer;
  24 use Wikimedia\Rdbms\IDatabase;
  25 use Wikimedia\Rdbms\DBConnRef;
  26 use Wikimedia\Rdbms\MaintainableDBConnRef;
  27
  28 /**
  29  * DB accessable external objects.
  30  *
  31  * In this system, each store "location" maps to a database "cluster".
  32  * The clusters must be defined in the normal LBFactory configuration.
  33  *
  34  * @ingroup ExternalStorage
  35  */
  36 class ExternalStoreDB extends ExternalStoreMedium {
  37         /**
  38          * The provided URL is in the form of DB://cluster/id
  39          * or DB://cluster/id/itemid for concatened storage.
  40          *
  41          * @param string $url
  42          * @return string|bool False if missing
  43          * @see ExternalStoreMedium::fetchFromURL()
  44          */
  45         public function fetchFromURL( $url ) {
  46                 list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  47                 $ret = $this->fetchBlob( $cluster, $id, $itemID );
  48
  49                 if ( $itemID !== false && $ret !== false ) {
  50                         return $ret->getItem( $itemID );
  51                 }
  52
  53                 return $ret;
  54         }
  55
  56         /**
  57          * Fetch data from given external store URLs.
  58          * The provided URLs are in the form of DB://cluster/id
  59          * or DB://cluster/id/itemid for concatened storage.
  60          *
  61          * @param array $urls An array of external store URLs
  62          * @return array A map from url to stored content. Failed results
  63          *     are not represented.
  64          */
  65         public function batchFetchFromURLs( array $urls ) {
  66                 $batched = $inverseUrlMap = [];
  67                 foreach ( $urls as $url ) {
  68                         list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  69                         $batched[$cluster][$id][] = $itemID;
  70                         // false $itemID gets cast to int, but should be ok
  71                         // since we do === from the $itemID in $batched
  72                         $inverseUrlMap[$cluster][$id][$itemID] = $url;
  73                 }
  74                 $ret = [];
  75                 foreach ( $batched as $cluster => $batchByCluster ) {
  76                         $res = $this->batchFetchBlobs( $cluster, $batchByCluster );
  77                         /** @var HistoryBlob $blob */
  78                         foreach ( $res as $id => $blob ) {
  79                                 foreach ( $batchByCluster[$id] as $itemID ) {
  80                                         $url = $inverseUrlMap[$cluster][$id][$itemID];
  81                                         if ( $itemID === false ) {
  82                                                 $ret[$url] = $blob;
  83                                         } else {
  84                                                 $ret[$url] = $blob->getItem( $itemID );
  85                                         }
  86                                 }
  87                         }
  88                 }
  89
  90                 return $ret;
  91         }
  92
  93         public function store( $location, $data ) {
  94                 $dbw = $this->getMaster( $location );
  95                 $dbw->insert( $this->getTable( $dbw ),
  96                         [ 'blob_text' => $data ],
  97                         __METHOD__ );
  98                 $id = $dbw->insertId();
  99                 if ( !$id ) {
 100                         throw new MWException( __METHOD__ . ': no insert ID' );
 101                 }
 102
 103                 return "DB://$location/$id";
 104         }
 105
 106         public function isReadOnly( $location ) {
 107                 return ( $this->getLoadBalancer( $location )->getReadOnlyReason() !== false );
 108         }
 109
 110         /**
 111          * Get a LoadBalancer for the specified cluster
 112          *
 113          * @param string $cluster Cluster name
 114          * @return LoadBalancer
 115          */
 116         private function getLoadBalancer( $cluster ) {
 117                 return wfGetLBFactory()->getExternalLB( $cluster );
 118         }
 119
 120         /**
 121          * Get a replica DB connection for the specified cluster
 122          *
 123          * @param string $cluster Cluster name
 124          * @return DBConnRef
 125          */
 126         public function getSlave( $cluster ) {
 127                 global $wgDefaultExternalStore;
 128
 129                 $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
 130                 $lb = $this->getLoadBalancer( $cluster );
 131
 132                 if ( !in_array( "DB://" . $cluster, (array)$wgDefaultExternalStore ) ) {
 133                         wfDebug( "read only external store\n" );
 134                         $lb->allowLagged( true );
 135                 } else {
 136                         wfDebug( "writable external store\n" );
 137                 }
 138
 139                 $db = $lb->getConnectionRef( DB_REPLICA, [], $wiki );
 140                 $db->clearFlag( DBO_TRX ); // sanity
 141
 142                 return $db;
 143         }
 144
 145         /**
 146          * Get a master database connection for the specified cluster
 147          *
 148          * @param string $cluster Cluster name
 149          * @return MaintainableDBConnRef
 150          */
 151         public function getMaster( $cluster ) {
 152                 $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
 153                 $lb = $this->getLoadBalancer( $cluster );
 154
 155                 $db = $lb->getMaintenanceConnectionRef( DB_MASTER, [], $wiki );
 156                 $db->clearFlag( DBO_TRX ); // sanity
 157
 158                 return $db;
 159         }
 160
 161         /**
 162          * Get the 'blobs' table name for this database
 163          *
 164          * @param IDatabase $db
 165          * @return string Table name ('blobs' by default)
 166          */
 167         public function getTable( $db ) {
 168                 $table = $db->getLBInfo( 'blobs table' );
 169                 if ( is_null( $table ) ) {
 170                         $table = 'blobs';
 171                 }
 172
 173                 return $table;
 174         }
 175
 176         /**
 177          * Fetch a blob item out of the database; a cache of the last-loaded
 178          * blob will be kept so that multiple loads out of a multi-item blob
 179          * can avoid redundant database access and decompression.
 180          * @param string $cluster
 181          * @param string $id
 182          * @param string $itemID
 183          * @return HistoryBlob|bool Returns false if missing
 184          */
 185         private function fetchBlob( $cluster, $id, $itemID ) {
 186                 /**
 187                  * One-step cache variable to hold base blobs; operations that
 188                  * pull multiple revisions may often pull multiple times from
 189                  * the same blob. By keeping the last-used one open, we avoid
 190                  * redundant unserialization and decompression overhead.
 191                  */
 192                 static $externalBlobCache = [];
 193
 194                 $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
 195                 if ( isset( $externalBlobCache[$cacheID] ) ) {
 196                         wfDebugLog( 'ExternalStoreDB-cache',
 197                                 "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
 198
 199                         return $externalBlobCache[$cacheID];
 200                 }
 201
 202                 wfDebugLog( 'ExternalStoreDB-cache',
 203                         "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
 204
 205                 $dbr = $this->getSlave( $cluster );
 206                 $ret = $dbr->selectField( $this->getTable( $dbr ),
 207                         'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 208                 if ( $ret === false ) {
 209                         wfDebugLog( 'ExternalStoreDB',
 210                                 "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
 211                         // Try the master
 212                         $dbw = $this->getMaster( $cluster );
 213                         $ret = $dbw->selectField( $this->getTable( $dbw ),
 214                                 'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 215                         if ( $ret === false ) {
 216                                 wfDebugLog( 'ExternalStoreDB',
 217                                         "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
 218                         }
 219                 }
 220                 if ( $itemID !== false && $ret !== false ) {
 221                         // Unserialise object; caller extracts item
 222                         $ret = unserialize( $ret );
 223                 }
 224
 225                 $externalBlobCache = [ $cacheID => $ret ];
 226
 227                 return $ret;
 228         }
 229
 230         /**
 231          * Fetch multiple blob items out of the database
 232          *
 233          * @param string $cluster A cluster name valid for use with LBFactory
 234          * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs
 235          * @return array A map from the blob_id's requested to their content.
 236          *   Unlocated ids are not represented
 237          */
 238         private function batchFetchBlobs( $cluster, array $ids ) {
 239                 $dbr = $this->getSlave( $cluster );
 240                 $res = $dbr->select( $this->getTable( $dbr ),
 241                         [ 'blob_id', 'blob_text' ], [ 'blob_id' => array_keys( $ids ) ], __METHOD__ );
 242                 $ret = [];
 243                 if ( $res !== false ) {
 244                         $this->mergeBatchResult( $ret, $ids, $res );
 245                 }
 246                 if ( $ids ) {
 247                         wfDebugLog( __CLASS__, __METHOD__ .
 248                                 " master fallback on '$cluster' for: " .
 249                                 implode( ',', array_keys( $ids ) ) );
 250                         // Try the master
 251                         $dbw = $this->getMaster( $cluster );
 252                         $res = $dbw->select( $this->getTable( $dbr ),
 253                                 [ 'blob_id', 'blob_text' ],
 254                                 [ 'blob_id' => array_keys( $ids ) ],
 255                                 __METHOD__ );
 256                         if ( $res === false ) {
 257                                 wfDebugLog( __CLASS__, __METHOD__ . " master failed on '$cluster'" );
 258                         } else {
 259                                 $this->mergeBatchResult( $ret, $ids, $res );
 260                         }
 261                 }
 262                 if ( $ids ) {
 263                         wfDebugLog( __CLASS__, __METHOD__ .
 264                                 " master on '$cluster' failed locating items: " .
 265                                 implode( ',', array_keys( $ids ) ) );
 266                 }
 267
 268                 return $ret;
 269         }
 270
 271         /**
 272          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 273          * @param array &$ret Current self::batchFetchBlobs return value
 274          * @param array &$ids Map from blob_id to requested itemIDs
 275          * @param mixed $res DB result from Database::select
 276          */
 277         private function mergeBatchResult( array &$ret, array &$ids, $res ) {
 278                 foreach ( $res as $row ) {
 279                         $id = $row->blob_id;
 280                         $itemIDs = $ids[$id];
 281                         unset( $ids[$id] ); // to track if everything is found
 282                         if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
 283                                 // single result stored per blob
 284                                 $ret[$id] = $row->blob_text;
 285                         } else {
 286                                 // multi result stored per blob
 287                                 $ret[$id] = unserialize( $row->blob_text );
 288                         }
 289                 }
 290         }
 291
 292         /**
 293          * @param string $url
 294          * @return array
 295          */
 296         protected function parseURL( $url ) {
 297                 $path = explode( '/', $url );
 298
 299                 return [
 300                         $path[2], // cluster
 301                         $path[3], // id
 302                         isset( $path[4] ) ? $path[4] : false // itemID
 303                 ];
 304         }
 305 }