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