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