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