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