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