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\LBFactory;
  24 use Wikimedia\Rdbms\ILoadBalancer;
  25 use Wikimedia\Rdbms\IDatabase;
  26 use Wikimedia\Rdbms\DBConnRef;
  27 use Wikimedia\Rdbms\MaintainableDBConnRef;
  28 use Wikimedia\Rdbms\DatabaseDomain;
  29
  30 /**
  31  * DB accessible external objects.
  32  *
  33  * In this system, each store "location" maps to a database "cluster".
  34  * The clusters must be defined in the normal LBFactory configuration.
  35  *
  36  * @ingroup ExternalStorage
  37  */
  38 class ExternalStoreDB extends ExternalStoreMedium {
  39         /** @var LBFactory */
  40         private $lbFactory;
  41
  42         /**
  43          * @see ExternalStoreMedium::__construct()
  44          * @param array $params Additional parameters include:
  45          *   - lbFactory: an LBFactory instance
  46          */
  47         public function __construct( array $params ) {
  48                 parent::__construct( $params );
  49                 if ( !isset( $params['lbFactory'] ) || !( $params['lbFactory'] instanceof LBFactory ) ) {
  50                         throw new InvalidArgumentException( "LBFactory required in 'lbFactory' field." );
  51                 }
  52                 $this->lbFactory = $params['lbFactory'];
  53         }
  54
  55         /**
  56          * The provided URL is in the form of DB://cluster/id
  57          * or DB://cluster/id/itemid for concatened storage.
  58          *
  59          * @param string $url
  60          * @return string|bool False if missing
  61          * @see ExternalStoreMedium::fetchFromURL()
  62          */
  63         public function fetchFromURL( $url ) {
  64                 list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  65                 $ret = $this->fetchBlob( $cluster, $id, $itemID );
  66
  67                 if ( $itemID !== false && $ret !== false ) {
  68                         return $ret->getItem( $itemID );
  69                 }
  70
  71                 return $ret;
  72         }
  73
  74         /**
  75          * Fetch data from given external store URLs.
  76          * The provided URLs are in the form of DB://cluster/id
  77          * or DB://cluster/id/itemid for concatened storage.
  78          *
  79          * @param array $urls An array of external store URLs
  80          * @return array A map from url to stored content. Failed results
  81          *     are not represented.
  82          */
  83         public function batchFetchFromURLs( array $urls ) {
  84                 $batched = $inverseUrlMap = [];
  85                 foreach ( $urls as $url ) {
  86                         list( $cluster, $id, $itemID ) = $this->parseURL( $url );
  87                         $batched[$cluster][$id][] = $itemID;
  88                         // false $itemID gets cast to int, but should be ok
  89                         // since we do === from the $itemID in $batched
  90                         $inverseUrlMap[$cluster][$id][$itemID] = $url;
  91                 }
  92                 $ret = [];
  93                 foreach ( $batched as $cluster => $batchByCluster ) {
  94                         $res = $this->batchFetchBlobs( $cluster, $batchByCluster );
  95                         /** @var HistoryBlob $blob */
  96                         foreach ( $res as $id => $blob ) {
  97                                 foreach ( $batchByCluster[$id] as $itemID ) {
  98                                         $url = $inverseUrlMap[$cluster][$id][$itemID];
  99                                         if ( $itemID === false ) {
 100                                                 $ret[$url] = $blob;
 101                                         } else {
 102                                                 $ret[$url] = $blob->getItem( $itemID );
 103                                         }
 104                                 }
 105                         }
 106                 }
 107
 108                 return $ret;
 109         }
 110
 111         /**
 112          * @inheritDoc
 113          */
 114         public function store( $location, $data ) {
 115                 $dbw = $this->getMaster( $location );
 116                 $dbw->insert( $this->getTable( $dbw ), [ 'blob_text' => $data ], __METHOD__ );
 117                 $id = $dbw->insertId();
 118                 if ( !$id ) {
 119                         throw new MWException( __METHOD__ . ': no insert ID' );
 120                 }
 121
 122                 return "DB://$location/$id";
 123         }
 124
 125         /**
 126          * @inheritDoc
 127          */
 128         public function isReadOnly( $location ) {
 129                 if ( parent::isReadOnly( $location ) ) {
 130                         return true;
 131                 }
 132
 133                 $lb = $this->getLoadBalancer( $location );
 134                 $domainId = $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) );
 135
 136                 return ( $lb->getReadOnlyReason( $domainId ) !== false );
 137         }
 138
 139         /**
 140          * Get a LoadBalancer for the specified cluster
 141          *
 142          * @param string $cluster Cluster name
 143          * @return ILoadBalancer
 144          */
 145         private function getLoadBalancer( $cluster ) {
 146                 return $this->lbFactory->getExternalLB( $cluster );
 147         }
 148
 149         /**
 150          * Get a replica DB connection for the specified cluster
 151          *
 152          * @param string $cluster Cluster name
 153          * @return DBConnRef
 154          */
 155         public function getSlave( $cluster ) {
 156                 $lb = $this->getLoadBalancer( $cluster );
 157
 158                 return $lb->getConnectionRef(
 159                         DB_REPLICA,
 160                         [],
 161                         $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) ),
 162                         $lb::CONN_TRX_AUTOCOMMIT
 163                 );
 164         }
 165
 166         /**
 167          * Get a master database connection for the specified cluster
 168          *
 169          * @param string $cluster Cluster name
 170          * @return MaintainableDBConnRef
 171          */
 172         public function getMaster( $cluster ) {
 173                 $lb = $this->getLoadBalancer( $cluster );
 174
 175                 return $lb->getMaintenanceConnectionRef(
 176                         DB_MASTER,
 177                         [],
 178                         $this->getDomainId( $lb->getServerInfo( $lb->getWriterIndex() ) ),
 179                         $lb::CONN_TRX_AUTOCOMMIT
 180                 );
 181         }
 182
 183         /**
 184          * @param array $server Master DB server configuration array for LoadBalancer
 185          * @return string|bool Database domain ID or false
 186          */
 187         private function getDomainId( array $server ) {
 188                 if ( $this->isDbDomainExplicit ) {
 189                         return $this->dbDomain; // explicit foreign domain
 190                 }
 191
 192                 if ( isset( $server['dbname'] ) ) {
 193                         // T200471: for b/c, treat any "dbname" field as forcing which database to use.
 194                         // MediaWiki/LoadBalancer previously did not enforce any concept of a local DB
 195                         // domain, but rather assumed that the LB server configuration matched $wgDBname.
 196                         // This check is useful when the external storage DB for this cluster does not use
 197                         // the same name as the corresponding "main" DB(s) for wikis.
 198                         $domain = new DatabaseDomain(
 199                                 $server['dbname'],
 200                                 $server['schema'] ?? null,
 201                                 $server['tablePrefix'] ?? ''
 202                         );
 203
 204                         return $domain->getId();
 205                 }
 206
 207                 return false; // local LB domain
 208         }
 209
 210         /**
 211          * Get the 'blobs' table name for this database
 212          *
 213          * @param IDatabase $db
 214          * @return string Table name ('blobs' by default)
 215          */
 216         public function getTable( $db ) {
 217                 $table = $db->getLBInfo( 'blobs table' );
 218                 if ( is_null( $table ) ) {
 219                         $table = 'blobs';
 220                 }
 221
 222                 return $table;
 223         }
 224
 225         /**
 226          * Fetch a blob item out of the database; a cache of the last-loaded
 227          * blob will be kept so that multiple loads out of a multi-item blob
 228          * can avoid redundant database access and decompression.
 229          * @param string $cluster
 230          * @param string $id
 231          * @param string $itemID
 232          * @return HistoryBlob|bool Returns false if missing
 233          */
 234         private function fetchBlob( $cluster, $id, $itemID ) {
 235                 /**
 236                  * One-step cache variable to hold base blobs; operations that
 237                  * pull multiple revisions may often pull multiple times from
 238                  * the same blob. By keeping the last-used one open, we avoid
 239                  * redundant unserialization and decompression overhead.
 240                  */
 241                 static $externalBlobCache = [];
 242
 243                 $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
 244                 $cacheID = "$cacheID@{$this->dbDomain}";
 245
 246                 if ( isset( $externalBlobCache[$cacheID] ) ) {
 247                         $this->logger->debug( "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
 248
 249                         return $externalBlobCache[$cacheID];
 250                 }
 251
 252                 $this->logger->debug( "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
 253
 254                 $dbr = $this->getSlave( $cluster );
 255                 $ret = $dbr->selectField( $this->getTable( $dbr ),
 256                         'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 257                 if ( $ret === false ) {
 258                         $this->logger->info( "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
 259                         // Try the master
 260                         $dbw = $this->getMaster( $cluster );
 261                         $ret = $dbw->selectField( $this->getTable( $dbw ),
 262                                 'blob_text', [ 'blob_id' => $id ], __METHOD__ );
 263                         if ( $ret === false ) {
 264                                 $this->logger->error( "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
 265                         }
 266                 }
 267                 if ( $itemID !== false && $ret !== false ) {
 268                         // Unserialise object; caller extracts item
 269                         $ret = unserialize( $ret );
 270                 }
 271
 272                 $externalBlobCache = [ $cacheID => $ret ];
 273
 274                 return $ret;
 275         }
 276
 277         /**
 278          * Fetch multiple blob items out of the database
 279          *
 280          * @param string $cluster A cluster name valid for use with LBFactory
 281          * @param array $ids A map from the blob_id's to look for to the requested itemIDs in the blobs
 282          * @return array A map from the blob_id's requested to their content.
 283          *   Unlocated ids are not represented
 284          */
 285         private function batchFetchBlobs( $cluster, array $ids ) {
 286                 $dbr = $this->getSlave( $cluster );
 287                 $res = $dbr->select(
 288                         $this->getTable( $dbr ),
 289                         [ 'blob_id', 'blob_text' ],
 290                         [ 'blob_id' => array_keys( $ids ) ],
 291                         __METHOD__
 292                 );
 293
 294                 $ret = [];
 295                 if ( $res !== false ) {
 296                         $this->mergeBatchResult( $ret, $ids, $res );
 297                 }
 298                 if ( $ids ) {
 299                         $this->logger->info(
 300                                 __METHOD__ . ": master fallback on '$cluster' for: " .
 301                                 implode( ',', array_keys( $ids ) )
 302                         );
 303                         // Try the master
 304                         $dbw = $this->getMaster( $cluster );
 305                         $res = $dbw->select( $this->getTable( $dbr ),
 306                                 [ 'blob_id', 'blob_text' ],
 307                                 [ 'blob_id' => array_keys( $ids ) ],
 308                                 __METHOD__ );
 309                         if ( $res === false ) {
 310                                 $this->logger->error( __METHOD__ . ": master failed on '$cluster'" );
 311                         } else {
 312                                 $this->mergeBatchResult( $ret, $ids, $res );
 313                         }
 314                 }
 315                 if ( $ids ) {
 316                         $this->logger->error(
 317                                 __METHOD__ . ": master on '$cluster' failed locating items: " .
 318                                 implode( ',', array_keys( $ids ) )
 319                         );
 320                 }
 321
 322                 return $ret;
 323         }
 324
 325         /**
 326          * Helper function for self::batchFetchBlobs for merging master/replica DB results
 327          * @param array &$ret Current self::batchFetchBlobs return value
 328          * @param array &$ids Map from blob_id to requested itemIDs
 329          * @param mixed $res DB result from Database::select
 330          */
 331         private function mergeBatchResult( array &$ret, array &$ids, $res ) {
 332                 foreach ( $res as $row ) {
 333                         $id = $row->blob_id;
 334                         $itemIDs = $ids[$id];
 335                         unset( $ids[$id] ); // to track if everything is found
 336                         if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
 337                                 // single result stored per blob
 338                                 $ret[$id] = $row->blob_text;
 339                         } else {
 340                                 // multi result stored per blob
 341                                 $ret[$id] = unserialize( $row->blob_text );
 342                         }
 343                 }
 344         }
 345
 346         /**
 347          * @param string $url
 348          * @return array
 349          */
 350         protected function parseURL( $url ) {
 351                 $path = explode( '/', $url );
 352
 353                 return [
 354                         $path[2], // cluster
 355                         $path[3], // id
 356                         $path[4] ?? false // itemID
 357                 ];
 358         }
 359 }