includes/libs/rdbms/lbfactory/LBFactory.php

   1 <?php
   2 /**
   3  * Generator and manager 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 Psr\Log\LoggerInterface;
  25 use Wikimedia\ScopedCallback;
  26
  27 /**
  28  * An interface for generating database load balancers
  29  * @ingroup Database
  30  */
  31 abstract class LBFactory implements ILBFactory {
  32         /** @var ChronologyProtector */
  33         protected $chronProt;
  34         /** @var object|string Class name or object With profileIn/profileOut methods */
  35         protected $profiler;
  36         /** @var TransactionProfiler */
  37         protected $trxProfiler;
  38         /** @var LoggerInterface */
  39         protected $replLogger;
  40         /** @var LoggerInterface */
  41         protected $connLogger;
  42         /** @var LoggerInterface */
  43         protected $queryLogger;
  44         /** @var LoggerInterface */
  45         protected $perfLogger;
  46         /** @var callable Error logger */
  47         protected $errorLogger;
  48         /** @var BagOStuff */
  49         protected $srvCache;
  50         /** @var BagOStuff */
  51         protected $memCache;
  52         /** @var WANObjectCache */
  53         protected $wanCache;
  54
  55         /** @var DatabaseDomain Local domain */
  56         protected $localDomain;
  57         /** @var string Local hostname of the app server */
  58         protected $hostname;
  59         /** @var array Web request information about the client */
  60         protected $requestInfo;
  61
  62         /** @var mixed */
  63         protected $ticket;
  64         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  65         protected $trxRoundId = false;
  66         /** @var string|bool Reason all LBs are read-only or false if not */
  67         protected $readOnlyReason = false;
  68         /** @var callable[] */
  69         protected $replicationWaitCallbacks = [];
  70
  71         /** @var bool Whether this PHP instance is for a CLI script */
  72         protected $cliMode;
  73         /** @var string Agent name for query profiling */
  74         protected $agent;
  75
  76         private static $loggerFields =
  77                 [ 'replLogger', 'connLogger', 'queryLogger', 'perfLogger' ];
  78
  79         public function __construct( array $conf ) {
  80                 $this->localDomain = isset( $conf['localDomain'] )
  81                         ? DatabaseDomain::newFromId( $conf['localDomain'] )
  82                         : DatabaseDomain::newUnspecified();
  83
  84                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  85                         $this->readOnlyReason = $conf['readOnlyReason'];
  86                 }
  87
  88                 $this->srvCache = isset( $conf['srvCache'] ) ? $conf['srvCache'] : new EmptyBagOStuff();
  89                 $this->memCache = isset( $conf['memCache'] ) ? $conf['memCache'] : new EmptyBagOStuff();
  90                 $this->wanCache = isset( $conf['wanCache'] )
  91                         ? $conf['wanCache']
  92                         : WANObjectCache::newEmpty();
  93
  94                 foreach ( self::$loggerFields as $key ) {
  95                         $this->$key = isset( $conf[$key] ) ? $conf[$key] : new \Psr\Log\NullLogger();
  96                 }
  97                 $this->errorLogger = isset( $conf['errorLogger'] )
  98                         ? $conf['errorLogger']
  99                         : function ( Exception $e ) {
 100                                 trigger_error( E_USER_WARNING, get_class( $e ) . ': ' . $e->getMessage() );
 101                         };
 102
 103                 $this->profiler = isset( $params['profiler'] ) ? $params['profiler'] : null;
 104                 $this->trxProfiler = isset( $conf['trxProfiler'] )
 105                         ? $conf['trxProfiler']
 106                         : new TransactionProfiler();
 107
 108                 $this->requestInfo = [
 109                         'IPAddress' => isset( $_SERVER[ 'REMOTE_ADDR' ] ) ? $_SERVER[ 'REMOTE_ADDR' ] : '',
 110                         'UserAgent' => isset( $_SERVER['HTTP_USER_AGENT'] ) ? $_SERVER['HTTP_USER_AGENT'] : '',
 111                         'ChronologyProtection' => 'true'
 112                 ];
 113
 114                 $this->cliMode = isset( $params['cliMode'] ) ? $params['cliMode'] : PHP_SAPI === 'cli';
 115                 $this->hostname = isset( $conf['hostname'] ) ? $conf['hostname'] : gethostname();
 116                 $this->agent = isset( $params['agent'] ) ? $params['agent'] : '';
 117
 118                 $this->ticket = mt_rand();
 119         }
 120
 121         public function destroy() {
 122                 $this->shutdown( self::SHUTDOWN_NO_CHRONPROT );
 123                 $this->forEachLBCallMethod( 'disable' );
 124         }
 125
 126         public function shutdown(
 127                 $mode = self::SHUTDOWN_CHRONPROT_SYNC, callable $workCallback = null
 128         ) {
 129                 $chronProt = $this->getChronologyProtector();
 130                 if ( $mode === self::SHUTDOWN_CHRONPROT_SYNC ) {
 131                         $this->shutdownChronologyProtector( $chronProt, $workCallback, 'sync' );
 132                 } elseif ( $mode === self::SHUTDOWN_CHRONPROT_ASYNC ) {
 133                         $this->shutdownChronologyProtector( $chronProt, null, 'async' );
 134                 }
 135
 136                 $this->commitMasterChanges( __METHOD__ ); // sanity
 137         }
 138
 139         /**
 140          * @see ILBFactory::newMainLB()
 141          * @param bool $domain
 142          * @return LoadBalancer
 143          */
 144         abstract public function newMainLB( $domain = false );
 145
 146         /**
 147          * @see ILBFactory::getMainLB()
 148          * @param bool $domain
 149          * @return LoadBalancer
 150          */
 151         abstract public function getMainLB( $domain = false );
 152
 153         /**
 154          * @see ILBFactory::newExternalLB()
 155          * @param string $cluster
 156          * @return LoadBalancer
 157          */
 158         abstract public function newExternalLB( $cluster );
 159
 160         /**
 161          * @see ILBFactory::getExternalLB()
 162          * @param string $cluster
 163          * @return LoadBalancer
 164          */
 165         abstract public function getExternalLB( $cluster );
 166
 167         /**
 168          * Call a method of each tracked load balancer
 169          *
 170          * @param string $methodName
 171          * @param array $args
 172          */
 173         protected function forEachLBCallMethod( $methodName, array $args = [] ) {
 174                 $this->forEachLB(
 175                         function ( ILoadBalancer $loadBalancer, $methodName, array $args ) {
 176                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 177                         },
 178                         [ $methodName, $args ]
 179                 );
 180         }
 181
 182         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 183                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 184         }
 185
 186         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 187                 $this->commitMasterChanges( $fname, $options );
 188                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 189         }
 190
 191         public function beginMasterChanges( $fname = __METHOD__ ) {
 192                 if ( $this->trxRoundId !== false ) {
 193                         throw new DBTransactionError(
 194                                 null,
 195                                 "$fname: transaction round '{$this->trxRoundId}' already started."
 196                         );
 197                 }
 198                 $this->trxRoundId = $fname;
 199                 // Set DBO_TRX flags on all appropriate DBs
 200                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 201         }
 202
 203         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 204                 if ( $this->trxRoundId !== false && $this->trxRoundId !== $fname ) {
 205                         throw new DBTransactionError(
 206                                 null,
 207                                 "$fname: transaction round '{$this->trxRoundId}' still running."
 208                         );
 209                 }
 210                 /** @noinspection PhpUnusedLocalVariableInspection */
 211                 $scope = $this->getScopedPHPBehaviorForCommit(); // try to ignore client aborts
 212                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 213                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 214                 $this->trxRoundId = false;
 215                 // Perform pre-commit checks, aborting on failure
 216                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 217                 // Log the DBs and methods involved in multi-DB transactions
 218                 $this->logIfMultiDbTransaction();
 219                 // Actually perform the commit on all master DB connections and revert DBO_TRX
 220                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 221                 // Run all post-commit callbacks
 222                 /** @var Exception $e */
 223                 $e = null; // first callback exception
 224                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$e ) {
 225                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 226                         $e = $e ?: $ex;
 227                 } );
 228                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 229                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 230                 // Throw any last post-commit callback error
 231                 if ( $e instanceof Exception ) {
 232                         throw $e;
 233                 }
 234         }
 235
 236         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 237                 $this->trxRoundId = false;
 238                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 239                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 240                 // Run all post-rollback callbacks
 241                 $this->forEachLB( function ( ILoadBalancer $lb ) {
 242                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 243                 } );
 244         }
 245
 246         /**
 247          * Log query info if multi DB transactions are going to be committed now
 248          */
 249         private function logIfMultiDbTransaction() {
 250                 $callersByDB = [];
 251                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$callersByDB ) {
 252                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 253                         $callers = $lb->pendingMasterChangeCallers();
 254                         if ( $callers ) {
 255                                 $callersByDB[$masterName] = $callers;
 256                         }
 257                 } );
 258
 259                 if ( count( $callersByDB ) >= 2 ) {
 260                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 261                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 262                         foreach ( $callersByDB as $db => $callers ) {
 263                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 264                         }
 265                         $this->queryLogger->info( $msg );
 266                 }
 267         }
 268
 269         public function hasMasterChanges() {
 270                 $ret = false;
 271                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 272                         $ret = $ret || $lb->hasMasterChanges();
 273                 } );
 274
 275                 return $ret;
 276         }
 277
 278         public function laggedReplicaUsed() {
 279                 $ret = false;
 280                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 281                         $ret = $ret || $lb->laggedReplicaUsed();
 282                 } );
 283
 284                 return $ret;
 285         }
 286
 287         public function hasOrMadeRecentMasterChanges( $age = null ) {
 288                 $ret = false;
 289                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $age, &$ret ) {
 290                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges( $age );
 291                 } );
 292                 return $ret;
 293         }
 294
 295         public function waitForReplication( array $opts = [] ) {
 296                 $opts += [
 297                         'domain' => false,
 298                         'cluster' => false,
 299                         'timeout' => 60,
 300                         'ifWritesSince' => null
 301                 ];
 302
 303                 if ( $opts['domain'] === false && isset( $opts['wiki'] ) ) {
 304                         $opts['domain'] = $opts['wiki']; // b/c
 305                 }
 306
 307                 // Figure out which clusters need to be checked
 308                 /** @var ILoadBalancer[] $lbs */
 309                 $lbs = [];
 310                 if ( $opts['cluster'] !== false ) {
 311                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 312                 } elseif ( $opts['domain'] !== false ) {
 313                         $lbs[] = $this->getMainLB( $opts['domain'] );
 314                 } else {
 315                         $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$lbs ) {
 316                                 $lbs[] = $lb;
 317                         } );
 318                         if ( !$lbs ) {
 319                                 return; // nothing actually used
 320                         }
 321                 }
 322
 323                 // Get all the master positions of applicable DBs right now.
 324                 // This can be faster since waiting on one cluster reduces the
 325                 // time needed to wait on the next clusters.
 326                 $masterPositions = array_fill( 0, count( $lbs ), false );
 327                 foreach ( $lbs as $i => $lb ) {
 328                         if ( $lb->getServerCount() <= 1 ) {
 329                                 // Bug 27975 - Don't try to wait for replica DBs if there are none
 330                                 // Prevents permission error when getting master position
 331                                 continue;
 332                         } elseif ( $opts['ifWritesSince']
 333                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 334                         ) {
 335                                 continue; // no writes since the last wait
 336                         }
 337                         $masterPositions[$i] = $lb->getMasterPos();
 338                 }
 339
 340                 // Run any listener callbacks *after* getting the DB positions. The more
 341                 // time spent in the callbacks, the less time is spent in waitForAll().
 342                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 343                         $callback();
 344                 }
 345
 346                 $failed = [];
 347                 foreach ( $lbs as $i => $lb ) {
 348                         if ( $masterPositions[$i] ) {
 349                                 // The DBMS may not support getMasterPos()
 350                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 351                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 352                                 }
 353                         }
 354                 }
 355
 356                 if ( $failed ) {
 357                         throw new DBReplicationWaitError(
 358                                 null,
 359                                 "Could not wait for replica DBs to catch up to " .
 360                                 implode( ', ', $failed )
 361                         );
 362                 }
 363         }
 364
 365         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 366                 if ( $callback ) {
 367                         $this->replicationWaitCallbacks[$name] = $callback;
 368                 } else {
 369                         unset( $this->replicationWaitCallbacks[$name] );
 370                 }
 371         }
 372
 373         public function getEmptyTransactionTicket( $fname ) {
 374                 if ( $this->hasMasterChanges() ) {
 375                         $this->queryLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 376                                 ( new RuntimeException() )->getTraceAsString() );
 377
 378                         return null;
 379                 }
 380
 381                 return $this->ticket;
 382         }
 383
 384         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 385                 if ( $ticket !== $this->ticket ) {
 386                         $this->perfLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 387                                 ( new RuntimeException() )->getTraceAsString() );
 388
 389                         return;
 390                 }
 391
 392                 // The transaction owner and any caller with the empty transaction ticket can commit
 393                 // so that getEmptyTransactionTicket() callers don't risk seeing DBTransactionError.
 394                 if ( $this->trxRoundId !== false && $fname !== $this->trxRoundId ) {
 395                         $this->queryLogger->info( "$fname: committing on behalf of {$this->trxRoundId}." );
 396                         $fnameEffective = $this->trxRoundId;
 397                 } else {
 398                         $fnameEffective = $fname;
 399                 }
 400
 401                 $this->commitMasterChanges( $fnameEffective );
 402                 $this->waitForReplication( $opts );
 403                 // If a nested caller committed on behalf of $fname, start another empty $fname
 404                 // transaction, leaving the caller with the same empty transaction state as before.
 405                 if ( $fnameEffective !== $fname ) {
 406                         $this->beginMasterChanges( $fnameEffective );
 407                 }
 408         }
 409
 410         public function getChronologyProtectorTouched( $dbName ) {
 411                 return $this->getChronologyProtector()->getTouched( $dbName );
 412         }
 413
 414         public function disableChronologyProtection() {
 415                 $this->getChronologyProtector()->setEnabled( false );
 416         }
 417
 418         /**
 419          * @return ChronologyProtector
 420          */
 421         protected function getChronologyProtector() {
 422                 if ( $this->chronProt ) {
 423                         return $this->chronProt;
 424                 }
 425
 426                 $this->chronProt = new ChronologyProtector(
 427                         $this->memCache,
 428                         [
 429                                 'ip' => $this->requestInfo['IPAddress'],
 430                                 'agent' => $this->requestInfo['UserAgent'],
 431                         ],
 432                         isset( $_GET['cpPosTime'] ) ? $_GET['cpPosTime'] : null
 433                 );
 434                 $this->chronProt->setLogger( $this->replLogger );
 435
 436                 if ( $this->cliMode ) {
 437                         $this->chronProt->setEnabled( false );
 438                 } elseif ( $this->requestInfo['ChronologyProtection'] === 'false' ) {
 439                         // Request opted out of using position wait logic. This is useful for requests
 440                         // done by the job queue or background ETL that do not have a meaningful session.
 441                         $this->chronProt->setWaitEnabled( false );
 442                 }
 443
 444                 $this->replLogger->debug( __METHOD__ . ': using request info ' .
 445                         json_encode( $this->requestInfo, JSON_PRETTY_PRINT ) );
 446
 447                 return $this->chronProt;
 448         }
 449
 450         /**
 451          * Get and record all of the staged DB positions into persistent memory storage
 452          *
 453          * @param ChronologyProtector $cp
 454          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 455          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 456          */
 457         protected function shutdownChronologyProtector(
 458                 ChronologyProtector $cp, $workCallback, $mode
 459         ) {
 460                 // Record all the master positions needed
 461                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $cp ) {
 462                         $cp->shutdownLB( $lb );
 463                 } );
 464                 // Write them to the persistent stash. Try to do something useful by running $work
 465                 // while ChronologyProtector waits for the stash write to replicate to all DCs.
 466                 $unsavedPositions = $cp->shutdown( $workCallback, $mode );
 467                 if ( $unsavedPositions && $workCallback ) {
 468                         // Invoke callback in case it did not cache the result yet
 469                         $workCallback(); // work now to block for less time in waitForAll()
 470                 }
 471                 // If the positions failed to write to the stash, at least wait on local datacenter
 472                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 473                 // the chance that the user will see their own changes immediately afterwards. As long
 474                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 475                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $unsavedPositions ) {
 476                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 477                         if ( isset( $unsavedPositions[$masterName] ) ) {
 478                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 479                         }
 480                 } );
 481         }
 482
 483         /**
 484          * Base parameters to LoadBalancer::__construct()
 485          * @return array
 486          */
 487         final protected function baseLoadBalancerParams() {
 488                 return [
 489                         'localDomain' => $this->localDomain,
 490                         'readOnlyReason' => $this->readOnlyReason,
 491                         'srvCache' => $this->srvCache,
 492                         'wanCache' => $this->wanCache,
 493                         'profiler' => $this->profiler,
 494                         'trxProfiler' => $this->trxProfiler,
 495                         'queryLogger' => $this->queryLogger,
 496                         'connLogger' => $this->connLogger,
 497                         'replLogger' => $this->replLogger,
 498                         'errorLogger' => $this->errorLogger,
 499                         'hostname' => $this->hostname,
 500                         'cliMode' => $this->cliMode,
 501                         'agent' => $this->agent
 502                 ];
 503         }
 504
 505         /**
 506          * @param ILoadBalancer $lb
 507          */
 508         protected function initLoadBalancer( ILoadBalancer $lb ) {
 509                 if ( $this->trxRoundId !== false ) {
 510                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 511                 }
 512         }
 513
 514         public function setDomainPrefix( $prefix ) {
 515                 $this->localDomain = new DatabaseDomain(
 516                         $this->localDomain->getDatabase(),
 517                         null,
 518                         $prefix
 519                 );
 520
 521                 $this->forEachLB( function( ILoadBalancer $lb ) use ( $prefix ) {
 522                         $lb->setDomainPrefix( $prefix );
 523                 } );
 524         }
 525
 526         public function closeAll() {
 527                 $this->forEachLBCallMethod( 'closeAll', [] );
 528         }
 529
 530         public function setAgentName( $agent ) {
 531                 $this->agent = $agent;
 532         }
 533
 534         public function appendPreShutdownTimeAsQuery( $url, $time ) {
 535                 $usedCluster = 0;
 536                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$usedCluster ) {
 537                         $usedCluster |= ( $lb->getServerCount() > 1 );
 538                 } );
 539
 540                 if ( !$usedCluster ) {
 541                         return $url; // no master/replica clusters touched
 542                 }
 543
 544                 return strpos( $url, '?' ) === false ? "$url?cpPosTime=$time" : "$url&cpPosTime=$time";
 545         }
 546
 547         public function setRequestInfo( array $info ) {
 548                 $this->requestInfo = $info + $this->requestInfo;
 549         }
 550
 551         /**
 552          * Make PHP ignore user aborts/disconnects until the returned
 553          * value leaves scope. This returns null and does nothing in CLI mode.
 554          *
 555          * @return ScopedCallback|null
 556          */
 557         final protected function getScopedPHPBehaviorForCommit() {
 558                 if ( PHP_SAPI != 'cli' ) { // https://bugs.php.net/bug.php?id=47540
 559                         $old = ignore_user_abort( true ); // avoid half-finished operations
 560                         return new ScopedCallback( function () use ( $old ) {
 561                                 ignore_user_abort( $old );
 562                         } );
 563                 }
 564
 565                 return null;
 566         }
 567
 568         function __destruct() {
 569                 $this->destroy();
 570         }
 571 }