includes/db/loadbalancer/LBFactory.php

   1 <?php
   2 /**
   3  * Generator of database load balancing objects.
   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  * @ingroup Database
  22  */
  23
  24 use MediaWiki\MediaWikiServices;
  25 use MediaWiki\Services\DestructibleService;
  26 use Psr\Log\LoggerInterface;
  27 use MediaWiki\Logger\LoggerFactory;
  28
  29 /**
  30  * An interface for generating database load balancers
  31  * @ingroup Database
  32  */
  33 abstract class LBFactory implements DestructibleService {
  34         /** @var ChronologyProtector */
  35         protected $chronProt;
  36         /** @var TransactionProfiler */
  37         protected $trxProfiler;
  38         /** @var LoggerInterface */
  39         protected $trxLogger;
  40         /** @var BagOStuff */
  41         protected $srvCache;
  42         /** @var WANObjectCache */
  43         protected $wanCache;
  44
  45         /** @var string|bool Reason all LBs are read-only or false if not */
  46         protected $readOnlyReason = false;
  47
  48         const SHUTDOWN_NO_CHRONPROT = 1; // don't save ChronologyProtector positions (for async code)
  49
  50         /**
  51          * Construct a factory based on a configuration array (typically from $wgLBFactoryConf)
  52          * @param array $conf
  53          * @TODO: inject objects via dependency framework
  54          */
  55         public function __construct( array $conf ) {
  56                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  57                         $this->readOnlyReason = $conf['readOnlyReason'];
  58                 }
  59                 $this->chronProt = $this->newChronologyProtector();
  60                 $this->trxProfiler = Profiler::instance()->getTransactionProfiler();
  61                 // Use APC/memcached style caching, but avoids loops with CACHE_DB (T141804)
  62                 $cache = ObjectCache::getLocalServerInstance();
  63                 if ( $cache->getQoS( $cache::ATTR_EMULATION ) > $cache::QOS_EMULATION_SQL ) {
  64                         $this->srvCache = $cache;
  65                 } else {
  66                         $this->srvCache = new EmptyBagOStuff();
  67                 }
  68                 $wCache = ObjectCache::getMainWANInstance();
  69                 if ( $wCache->getQoS( $wCache::ATTR_EMULATION ) > $wCache::QOS_EMULATION_SQL ) {
  70                         $this->wanCache = $wCache;
  71                 } else {
  72                         $this->wanCache = WANObjectCache::newEmpty();
  73                 }
  74                 $this->trxLogger = LoggerFactory::getInstance( 'DBTransaction' );
  75         }
  76
  77         /**
  78          * Disables all load balancers. All connections are closed, and any attempt to
  79          * open a new connection will result in a DBAccessError.
  80          * @see LoadBalancer::disable()
  81          */
  82         public function destroy() {
  83                 $this->shutdown();
  84                 $this->forEachLBCallMethod( 'disable' );
  85         }
  86
  87         /**
  88          * Disables all access to the load balancer, will cause all database access
  89          * to throw a DBAccessError
  90          */
  91         public static function disableBackend() {
  92                 MediaWikiServices::disableStorageBackend();
  93         }
  94
  95         /**
  96          * Get an LBFactory instance
  97          *
  98          * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
  99          *
 100          * @return LBFactory
 101          */
 102         public static function singleton() {
 103                 return MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 104         }
 105
 106         /**
 107          * Returns the LBFactory class to use and the load balancer configuration.
 108          *
 109          * @todo instead of this, use a ServiceContainer for managing the different implementations.
 110          *
 111          * @param array $config (e.g. $wgLBFactoryConf)
 112          * @return string Class name
 113          */
 114         public static function getLBFactoryClass( array $config ) {
 115                 // For configuration backward compatibility after removing
 116                 // underscores from class names in MediaWiki 1.23.
 117                 $bcClasses = [
 118                         'LBFactory_Simple' => 'LBFactorySimple',
 119                         'LBFactory_Single' => 'LBFactorySingle',
 120                         'LBFactory_Multi' => 'LBFactoryMulti',
 121                         'LBFactory_Fake' => 'LBFactoryFake',
 122                 ];
 123
 124                 $class = $config['class'];
 125
 126                 if ( isset( $bcClasses[$class] ) ) {
 127                         $class = $bcClasses[$class];
 128                         wfDeprecated(
 129                                 '$wgLBFactoryConf must be updated. See RELEASE-NOTES for details',
 130                                 '1.23'
 131                         );
 132                 }
 133
 134                 return $class;
 135         }
 136
 137         /**
 138          * Shut down, close connections and destroy the cached instance.
 139          *
 140          * @deprecated since 1.27, use LBFactory::destroy()
 141          */
 142         public static function destroyInstance() {
 143                 self::singleton()->destroy();
 144         }
 145
 146         /**
 147          * Create a new load balancer object. The resulting object will be untracked,
 148          * not chronology-protected, and the caller is responsible for cleaning it up.
 149          *
 150          * @param bool|string $wiki Wiki ID, or false for the current wiki
 151          * @return LoadBalancer
 152          */
 153         abstract public function newMainLB( $wiki = false );
 154
 155         /**
 156          * Get a cached (tracked) load balancer object.
 157          *
 158          * @param bool|string $wiki Wiki ID, or false for the current wiki
 159          * @return LoadBalancer
 160          */
 161         abstract public function getMainLB( $wiki = false );
 162
 163         /**
 164          * Create a new load balancer for external storage. The resulting object will be
 165          * untracked, not chronology-protected, and the caller is responsible for
 166          * cleaning it up.
 167          *
 168          * @param string $cluster External storage cluster, or false for core
 169          * @param bool|string $wiki Wiki ID, or false for the current wiki
 170          * @return LoadBalancer
 171          */
 172         abstract protected function newExternalLB( $cluster, $wiki = false );
 173
 174         /**
 175          * Get a cached (tracked) load balancer for external storage
 176          *
 177          * @param string $cluster External storage cluster, or false for core
 178          * @param bool|string $wiki Wiki ID, or false for the current wiki
 179          * @return LoadBalancer
 180          */
 181         abstract public function &getExternalLB( $cluster, $wiki = false );
 182
 183         /**
 184          * Execute a function for each tracked load balancer
 185          * The callback is called with the load balancer as the first parameter,
 186          * and $params passed as the subsequent parameters.
 187          *
 188          * @param callable $callback
 189          * @param array $params
 190          */
 191         abstract public function forEachLB( $callback, array $params = [] );
 192
 193         /**
 194          * Prepare all tracked load balancers for shutdown
 195          * @param integer $flags Supports SHUTDOWN_* flags
 196          * STUB
 197          */
 198         public function shutdown( $flags = 0 ) {
 199         }
 200
 201         /**
 202          * Call a method of each tracked load balancer
 203          *
 204          * @param string $methodName
 205          * @param array $args
 206          */
 207         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 208                 $this->forEachLB(
 209                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 210                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 211                         },
 212                         [ $methodName, $args ]
 213                 );
 214         }
 215
 216         /**
 217          * Commit on all connections. Done for two reasons:
 218          * 1. To commit changes to the masters.
 219          * 2. To release the snapshot on all connections, master and slave.
 220          * @param string $fname Caller name
 221          * @param array $options Options map:
 222          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 223          */
 224         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 225                 $this->commitMasterChanges( $fname, $options );
 226                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 227         }
 228
 229         /**
 230          * Commit changes on all master connections
 231          * @param string $fname Caller name
 232          * @param array $options Options map:
 233          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 234          */
 235         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 236                 // Perform all pre-commit callbacks, aborting on failure
 237                 $this->forEachLBCallMethod( 'runMasterPreCommitCallbacks' );
 238                 // Perform all pre-commit checks, aborting on failure
 239                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 240                 // Log the DBs and methods involved in multi-DB transactions
 241                 $this->logIfMultiDbTransaction();
 242                 // Actually perform the commit on all master DB connections
 243                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 244                 // Run all post-commit callbacks
 245                 $this->forEachLBCallMethod( 'runMasterPostCommitCallbacks' );
 246                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 247                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 248         }
 249
 250         /**
 251          * Rollback changes on all master connections
 252          * @param string $fname Caller name
 253          * @since 1.23
 254          */
 255         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 256                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 257         }
 258
 259         /**
 260          * Log query info if multi DB transactions are going to be committed now
 261          */
 262         private function logIfMultiDbTransaction() {
 263                 $callersByDB = [];
 264                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 265                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 266                         $callers = $lb->pendingMasterChangeCallers();
 267                         if ( $callers ) {
 268                                 $callersByDB[$masterName] = $callers;
 269                         }
 270                 } );
 271
 272                 if ( count( $callersByDB ) >= 2 ) {
 273                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 274                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 275                         foreach ( $callersByDB as $db => $callers ) {
 276                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 277                         }
 278                         $this->trxLogger->info( $msg );
 279                 }
 280         }
 281
 282         /**
 283          * Determine if any master connection has pending changes
 284          * @return bool
 285          * @since 1.23
 286          */
 287         public function hasMasterChanges() {
 288                 $ret = false;
 289                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 290                         $ret = $ret || $lb->hasMasterChanges();
 291                 } );
 292
 293                 return $ret;
 294         }
 295
 296         /**
 297          * Detemine if any lagged slave connection was used
 298          * @since 1.27
 299          * @return bool
 300          */
 301         public function laggedSlaveUsed() {
 302                 $ret = false;
 303                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 304                         $ret = $ret || $lb->laggedSlaveUsed();
 305                 } );
 306
 307                 return $ret;
 308         }
 309
 310         /**
 311          * Determine if any master connection has pending/written changes from this request
 312          * @return bool
 313          * @since 1.27
 314          */
 315         public function hasOrMadeRecentMasterChanges() {
 316                 $ret = false;
 317                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 318                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 319                 } );
 320                 return $ret;
 321         }
 322
 323         /**
 324          * Waits for the slave DBs to catch up to the current master position
 325          *
 326          * Use this when updating very large numbers of rows, as in maintenance scripts,
 327          * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
 328          *
 329          * By default this waits on all DB clusters actually used in this request.
 330          * This makes sense when lag being waiting on is caused by the code that does this check.
 331          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 332          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 333          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 334          * use the "cluster" parameter.
 335          *
 336          * Never call this function after a large DB write that is *still* in a transaction.
 337          * It only makes sense to call this after the possible lag inducing changes were committed.
 338          *
 339          * @param array $opts Optional fields that include:
 340          *   - wiki : wait on the load balancer DBs that handles the given wiki
 341          *   - cluster : wait on the given external load balancer DBs
 342          *   - timeout : Max wait time. Default: ~60 seconds
 343          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 344          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 345          * @since 1.27
 346          */
 347         public function waitForReplication( array $opts = [] ) {
 348                 $opts += [
 349                         'wiki' => false,
 350                         'cluster' => false,
 351                         'timeout' => 60,
 352                         'ifWritesSince' => null
 353                 ];
 354
 355                 // Figure out which clusters need to be checked
 356                 /** @var LoadBalancer[] $lbs */
 357                 $lbs = [];
 358                 if ( $opts['cluster'] !== false ) {
 359                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 360                 } elseif ( $opts['wiki'] !== false ) {
 361                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 362                 } else {
 363                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 364                                 $lbs[] = $lb;
 365                         } );
 366                         if ( !$lbs ) {
 367                                 return; // nothing actually used
 368                         }
 369                 }
 370
 371                 // Get all the master positions of applicable DBs right now.
 372                 // This can be faster since waiting on one cluster reduces the
 373                 // time needed to wait on the next clusters.
 374                 $masterPositions = array_fill( 0, count( $lbs ), false );
 375                 foreach ( $lbs as $i => $lb ) {
 376                         if ( $lb->getServerCount() <= 1 ) {
 377                                 // Bug 27975 - Don't try to wait for slaves if there are none
 378                                 // Prevents permission error when getting master position
 379                                 continue;
 380                         } elseif ( $opts['ifWritesSince']
 381                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 382                         ) {
 383                                 continue; // no writes since the last wait
 384                         }
 385                         $masterPositions[$i] = $lb->getMasterPos();
 386                 }
 387
 388                 $failed = [];
 389                 foreach ( $lbs as $i => $lb ) {
 390                         if ( $masterPositions[$i] ) {
 391                                 // The DBMS may not support getMasterPos() or the whole
 392                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 393                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 394                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 395                                 }
 396                         }
 397                 }
 398
 399                 if ( $failed ) {
 400                         throw new DBReplicationWaitError(
 401                                 "Could not wait for slaves to catch up to " .
 402                                 implode( ', ', $failed )
 403                         );
 404                 }
 405         }
 406
 407         /**
 408          * Disable the ChronologyProtector for all load balancers
 409          *
 410          * This can be called at the start of special API entry points
 411          *
 412          * @since 1.27
 413          */
 414         public function disableChronologyProtection() {
 415                 $this->chronProt->setEnabled( false );
 416         }
 417
 418         /**
 419          * @return ChronologyProtector
 420          */
 421         protected function newChronologyProtector() {
 422                 $request = RequestContext::getMain()->getRequest();
 423                 $chronProt = new ChronologyProtector(
 424                         ObjectCache::getMainStashInstance(),
 425                         [
 426                                 'ip' => $request->getIP(),
 427                                 'agent' => $request->getHeader( 'User-Agent' )
 428                         ]
 429                 );
 430                 if ( PHP_SAPI === 'cli' ) {
 431                         $chronProt->setEnabled( false );
 432                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 433                         // Request opted out of using position wait logic. This is useful for requests
 434                         // done by the job queue or background ETL that do not have a meaningful session.
 435                         $chronProt->setWaitEnabled( false );
 436                 }
 437
 438                 return $chronProt;
 439         }
 440
 441         /**
 442          * @param ChronologyProtector $cp
 443          */
 444         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 445                 // Get all the master positions needed
 446                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 447                         $cp->shutdownLB( $lb );
 448                 } );
 449                 // Write them to the stash
 450                 $unsavedPositions = $cp->shutdown();
 451                 // If the positions failed to write to the stash, at least wait on local datacenter
 452                 // slaves to catch up before responding. Even if there are several DCs, this increases
 453                 // the chance that the user will see their own changes immediately afterwards. As long
 454                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 455                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 456                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 457                         if ( isset( $unsavedPositions[$masterName] ) ) {
 458                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 459                         }
 460                 } );
 461         }
 462
 463         /**
 464          * Close all open database connections on all open load balancers.
 465          * @since 1.28
 466          */
 467         public function closeAll() {
 468                 $this->forEachLBCallMethod( 'closeAll', [] );
 469         }
 470
 471 }
 472
 473 /**
 474  * Exception class for attempted DB access
 475  */
 476 class DBAccessError extends MWException {
 477         public function __construct() {
 478                 parent::__construct( "Mediawiki tried to access the database via wfGetDB(). " .
 479                         "This is not allowed, because database access has been disabled." );
 480         }
 481 }
 482
 483 /**
 484  * Exception class for replica DB wait timeouts
 485  */
 486 class DBReplicationWaitError extends Exception {
 487 }