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         /**
 107          * Get a LoadBalancer for the specified cluster
 108          *
 109          * @param string $cluster Cluster name
 110          * @return LoadBalancer
 111          */
 112         private function getLoadBalancer( $cluster ) {
 113                 return wfGetLBFactory()->getExternalLB( $cluster );
 114         }
 115
 116         /**
 117          * Get a replica DB connection for the specified cluster
 118          *
 119          * @param string $cluster Cluster name
 120          * @return DBConnRef
 121          */
 122         public function getSlave( $cluster ) {
 123                 global $wgDefaultExternalStore;
 124
 125                 $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
 126                 $lb = $this->getLoadBalancer( $cluster );
 127
 128                 if ( !in_array( "DB://" . $cluster, (array)$wgDefaultExternalStore ) ) {
 129                         wfDebug( "read only external store\n" );
 130                         $lb->allowLagged( true );
 131                 } else {
 132                         wfDebug( "writable external store\n" );
 133                 }
 134
 135                 $db = $lb->getConnectionRef( DB_REPLICA, [], $wiki );
 136                 $db->clearFlag( DBO_TRX ); // sanity
 137
 138                 return $db;
 139         }
 140
 141         /**
 142          * Get a master database connection for the specified cluster
 143          *
 144          * @param string $cluster Cluster name
 145          * @return MaintainableDBConnRef
 146          */
 147         public function getMaster( $cluster ) {
 148                 $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
 149                 $lb = $this->getLoadBalancer( $cluster );
 150
 151                 $db = $lb->getMaintenanceConnectionRef( DB_MASTER, [], $wiki );
 152                 $db->clearFlag( DBO_TRX ); // sanity
 153
 154                 return $db;
 155         }
 156
 157         /**
 158          * Get the 'blobs' table name for this database
 159          *
 160          * @param IDatabase $db
 161          * @return string Table name ('blobs' by default)
 162          */
 163         public function getTable( $db ) {
 164                 $table = $db->getLBInfo( 'blobs table' );
 165                 if ( is_null( $table ) ) {
 166                         $table = 'blobs';
 167                 }
 168
 169                 return $table;
 170         }
 171
 172         /**
 173          * Fetch a blob item out of the database; a cache of the last-loaded
 174          * blob will be kept so that multiple loads out of a multi-item blob
 175          * can avoid redundant database access and decompression.
 176          * @param string $cluster
 177          * @param string $id
 178          * @param string $itemID
 179          * @return HistoryBlob|bool Returns false if missing
 180          */
 181         private 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         private 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 $ret;
 265         }
 266
 267         /**
 268          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 269          * @param array &$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 }