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 use Wikimedia\Rdbms\TransactionProfiler;
  27 use Wikimedia\Rdbms\ChronologyProtector;
  28
  29 /**
  30  * An interface for generating database load balancers
  31  * @ingroup Database
  32  */
  33 abstract class LBFactory implements ILBFactory {
  34         /** @var ChronologyProtector */
  35         protected $chronProt;
  36         /** @var object|string Class name or object With profileIn/profileOut methods */
  37         protected $profiler;
  38         /** @var TransactionProfiler */
  39         protected $trxProfiler;
  40         /** @var LoggerInterface */
  41         protected $replLogger;
  42         /** @var LoggerInterface */
  43         protected $connLogger;
  44         /** @var LoggerInterface */
  45         protected $queryLogger;
  46         /** @var LoggerInterface */
  47         protected $perfLogger;
  48         /** @var callable Error logger */
  49         protected $errorLogger;
  50         /** @var BagOStuff */
  51         protected $srvCache;
  52         /** @var BagOStuff */
  53         protected $memCache;
  54         /** @var WANObjectCache */
  55         protected $wanCache;
  56
  57         /** @var DatabaseDomain Local domain */
  58         protected $localDomain;
  59         /** @var string Local hostname of the app server */
  60         protected $hostname;
  61         /** @var array Web request information about the client */
  62         protected $requestInfo;
  63
  64         /** @var mixed */
  65         protected $ticket;
  66         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  67         protected $trxRoundId = false;
  68         /** @var string|bool Reason all LBs are read-only or false if not */
  69         protected $readOnlyReason = false;
  70         /** @var callable[] */
  71         protected $replicationWaitCallbacks = [];
  72
  73         /** @var bool Whether this PHP instance is for a CLI script */
  74         protected $cliMode;
  75         /** @var string Agent name for query profiling */
  76         protected $agent;
  77
  78         private static $loggerFields =
  79                 [ 'replLogger', 'connLogger', 'queryLogger', 'perfLogger' ];
  80
  81         public function __construct( array $conf ) {
  82                 $this->localDomain = isset( $conf['localDomain'] )
  83                         ? DatabaseDomain::newFromId( $conf['localDomain'] )
  84                         : DatabaseDomain::newUnspecified();
  85
  86                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  87                         $this->readOnlyReason = $conf['readOnlyReason'];
  88                 }
  89
  90                 $this->srvCache = isset( $conf['srvCache'] ) ? $conf['srvCache'] : new EmptyBagOStuff();
  91                 $this->memCache = isset( $conf['memCache'] ) ? $conf['memCache'] : new EmptyBagOStuff();
  92                 $this->wanCache = isset( $conf['wanCache'] )
  93                         ? $conf['wanCache']
  94                         : WANObjectCache::newEmpty();
  95
  96                 foreach ( self::$loggerFields as $key ) {
  97                         $this->$key = isset( $conf[$key] ) ? $conf[$key] : new \Psr\Log\NullLogger();
  98                 }
  99                 $this->errorLogger = isset( $conf['errorLogger'] )
 100                         ? $conf['errorLogger']
 101                         : function ( Exception $e ) {
 102                                 trigger_error( E_USER_WARNING, get_class( $e ) . ': ' . $e->getMessage() );
 103                         };
 104
 105                 $this->profiler = isset( $conf['profiler'] ) ? $conf['profiler'] : null;
 106                 $this->trxProfiler = isset( $conf['trxProfiler'] )
 107                         ? $conf['trxProfiler']
 108                         : new TransactionProfiler();
 109
 110                 $this->requestInfo = [
 111                         'IPAddress' => isset( $_SERVER[ 'REMOTE_ADDR' ] ) ? $_SERVER[ 'REMOTE_ADDR' ] : '',
 112                         'UserAgent' => isset( $_SERVER['HTTP_USER_AGENT'] ) ? $_SERVER['HTTP_USER_AGENT'] : '',
 113                         'ChronologyProtection' => 'true'
 114                 ];
 115
 116                 $this->cliMode = isset( $conf['cliMode'] ) ? $conf['cliMode'] : PHP_SAPI === 'cli';
 117                 $this->hostname = isset( $conf['hostname'] ) ? $conf['hostname'] : gethostname();
 118                 $this->agent = isset( $conf['agent'] ) ? $conf['agent'] : '';
 119
 120                 $this->ticket = mt_rand();
 121         }
 122
 123         public function destroy() {
 124                 $this->shutdown( self::SHUTDOWN_NO_CHRONPROT );
 125                 $this->forEachLBCallMethod( 'disable' );
 126         }
 127
 128         public function shutdown(
 129                 $mode = self::SHUTDOWN_CHRONPROT_SYNC, callable $workCallback = null
 130         ) {
 131                 $chronProt = $this->getChronologyProtector();
 132                 if ( $mode === self::SHUTDOWN_CHRONPROT_SYNC ) {
 133                         $this->shutdownChronologyProtector( $chronProt, $workCallback, 'sync' );
 134                 } elseif ( $mode === self::SHUTDOWN_CHRONPROT_ASYNC ) {
 135                         $this->shutdownChronologyProtector( $chronProt, null, 'async' );
 136                 }
 137
 138                 $this->commitMasterChanges( __METHOD__ ); // sanity
 139         }
 140
 141         /**
 142          * @see ILBFactory::newMainLB()
 143          * @param bool $domain
 144          * @return LoadBalancer
 145          */
 146         abstract public function newMainLB( $domain = false );
 147
 148         /**
 149          * @see ILBFactory::getMainLB()
 150          * @param bool $domain
 151          * @return LoadBalancer
 152          */
 153         abstract public function getMainLB( $domain = false );
 154
 155         /**
 156          * @see ILBFactory::newExternalLB()
 157          * @param string $cluster
 158          * @return LoadBalancer
 159          */
 160         abstract public function newExternalLB( $cluster );
 161
 162         /**
 163          * @see ILBFactory::getExternalLB()
 164          * @param string $cluster
 165          * @return LoadBalancer
 166          */
 167         abstract public function getExternalLB( $cluster );
 168
 169         /**
 170          * Call a method of each tracked load balancer
 171          *
 172          * @param string $methodName
 173          * @param array $args
 174          */
 175         protected function forEachLBCallMethod( $methodName, array $args = [] ) {
 176                 $this->forEachLB(
 177                         function ( ILoadBalancer $loadBalancer, $methodName, array $args ) {
 178                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 179                         },
 180                         [ $methodName, $args ]
 181                 );
 182         }
 183
 184         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 185                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 186         }
 187
 188         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 189                 $this->commitMasterChanges( $fname, $options );
 190                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 191         }
 192
 193         public function beginMasterChanges( $fname = __METHOD__ ) {
 194                 if ( $this->trxRoundId !== false ) {
 195                         throw new DBTransactionError(
 196                                 null,
 197                                 "$fname: transaction round '{$this->trxRoundId}' already started."
 198                         );
 199                 }
 200                 $this->trxRoundId = $fname;
 201                 // Set DBO_TRX flags on all appropriate DBs
 202                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 203         }
 204
 205         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 206                 if ( $this->trxRoundId !== false && $this->trxRoundId !== $fname ) {
 207                         throw new DBTransactionError(
 208                                 null,
 209                                 "$fname: transaction round '{$this->trxRoundId}' still running."
 210                         );
 211                 }
 212                 /** @noinspection PhpUnusedLocalVariableInspection */
 213                 $scope = $this->getScopedPHPBehaviorForCommit(); // try to ignore client aborts
 214                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 215                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 216                 $this->trxRoundId = false;
 217                 // Perform pre-commit checks, aborting on failure
 218                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 219                 // Log the DBs and methods involved in multi-DB transactions
 220                 $this->logIfMultiDbTransaction();
 221                 // Actually perform the commit on all master DB connections and revert DBO_TRX
 222                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 223                 // Run all post-commit callbacks
 224                 /** @var Exception $e */
 225                 $e = null; // first callback exception
 226                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$e ) {
 227                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 228                         $e = $e ?: $ex;
 229                 } );
 230                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 231                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 232                 // Throw any last post-commit callback error
 233                 if ( $e instanceof Exception ) {
 234                         throw $e;
 235                 }
 236         }
 237
 238         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 239                 $this->trxRoundId = false;
 240                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 241                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 242                 // Run all post-rollback callbacks
 243                 $this->forEachLB( function ( ILoadBalancer $lb ) {
 244                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 245                 } );
 246         }
 247
 248         /**
 249          * Log query info if multi DB transactions are going to be committed now
 250          */
 251         private function logIfMultiDbTransaction() {
 252                 $callersByDB = [];
 253                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$callersByDB ) {
 254                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 255                         $callers = $lb->pendingMasterChangeCallers();
 256                         if ( $callers ) {
 257                                 $callersByDB[$masterName] = $callers;
 258                         }
 259                 } );
 260
 261                 if ( count( $callersByDB ) >= 2 ) {
 262                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 263                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 264                         foreach ( $callersByDB as $db => $callers ) {
 265                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 266                         }
 267                         $this->queryLogger->info( $msg );
 268                 }
 269         }
 270
 271         public function hasMasterChanges() {
 272                 $ret = false;
 273                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 274                         $ret = $ret || $lb->hasMasterChanges();
 275                 } );
 276
 277                 return $ret;
 278         }
 279
 280         public function laggedReplicaUsed() {
 281                 $ret = false;
 282                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 283                         $ret = $ret || $lb->laggedReplicaUsed();
 284                 } );
 285
 286                 return $ret;
 287         }
 288
 289         public function hasOrMadeRecentMasterChanges( $age = null ) {
 290                 $ret = false;
 291                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $age, &$ret ) {
 292                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges( $age );
 293                 } );
 294                 return $ret;
 295         }
 296
 297         public function waitForReplication( array $opts = [] ) {
 298                 $opts += [
 299                         'domain' => false,
 300                         'cluster' => false,
 301                         'timeout' => 60,
 302                         'ifWritesSince' => null
 303                 ];
 304
 305                 if ( $opts['domain'] === false && isset( $opts['wiki'] ) ) {
 306                         $opts['domain'] = $opts['wiki']; // b/c
 307                 }
 308
 309                 // Figure out which clusters need to be checked
 310                 /** @var ILoadBalancer[] $lbs */
 311                 $lbs = [];
 312                 if ( $opts['cluster'] !== false ) {
 313                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 314                 } elseif ( $opts['domain'] !== false ) {
 315                         $lbs[] = $this->getMainLB( $opts['domain'] );
 316                 } else {
 317                         $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$lbs ) {
 318                                 $lbs[] = $lb;
 319                         } );
 320                         if ( !$lbs ) {
 321                                 return; // nothing actually used
 322                         }
 323                 }
 324
 325                 // Get all the master positions of applicable DBs right now.
 326                 // This can be faster since waiting on one cluster reduces the
 327                 // time needed to wait on the next clusters.
 328                 $masterPositions = array_fill( 0, count( $lbs ), false );
 329                 foreach ( $lbs as $i => $lb ) {
 330                         if ( $lb->getServerCount() <= 1 ) {
 331                                 // Bug 27975 - Don't try to wait for replica DBs if there are none
 332                                 // Prevents permission error when getting master position
 333                                 continue;
 334                         } elseif ( $opts['ifWritesSince']
 335                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 336                         ) {
 337                                 continue; // no writes since the last wait
 338                         }
 339                         $masterPositions[$i] = $lb->getMasterPos();
 340                 }
 341
 342                 // Run any listener callbacks *after* getting the DB positions. The more
 343                 // time spent in the callbacks, the less time is spent in waitForAll().
 344                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 345                         $callback();
 346                 }
 347
 348                 $failed = [];
 349                 foreach ( $lbs as $i => $lb ) {
 350                         if ( $masterPositions[$i] ) {
 351                                 // The DBMS may not support getMasterPos()
 352                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 353                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 354                                 }
 355                         }
 356                 }
 357
 358                 if ( $failed ) {
 359                         throw new DBReplicationWaitError(
 360                                 null,
 361                                 "Could not wait for replica DBs to catch up to " .
 362                                 implode( ', ', $failed )
 363                         );
 364                 }
 365         }
 366
 367         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 368                 if ( $callback ) {
 369                         $this->replicationWaitCallbacks[$name] = $callback;
 370                 } else {
 371                         unset( $this->replicationWaitCallbacks[$name] );
 372                 }
 373         }
 374
 375         public function getEmptyTransactionTicket( $fname ) {
 376                 if ( $this->hasMasterChanges() ) {
 377                         $this->queryLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 378                                 ( new RuntimeException() )->getTraceAsString() );
 379
 380                         return null;
 381                 }
 382
 383                 return $this->ticket;
 384         }
 385
 386         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 387                 if ( $ticket !== $this->ticket ) {
 388                         $this->perfLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 389                                 ( new RuntimeException() )->getTraceAsString() );
 390
 391                         return;
 392                 }
 393
 394                 // The transaction owner and any caller with the empty transaction ticket can commit
 395                 // so that getEmptyTransactionTicket() callers don't risk seeing DBTransactionError.
 396                 if ( $this->trxRoundId !== false && $fname !== $this->trxRoundId ) {
 397                         $this->queryLogger->info( "$fname: committing on behalf of {$this->trxRoundId}." );
 398                         $fnameEffective = $this->trxRoundId;
 399                 } else {
 400                         $fnameEffective = $fname;
 401                 }
 402
 403                 $this->commitMasterChanges( $fnameEffective );
 404                 $this->waitForReplication( $opts );
 405                 // If a nested caller committed on behalf of $fname, start another empty $fname
 406                 // transaction, leaving the caller with the same empty transaction state as before.
 407                 if ( $fnameEffective !== $fname ) {
 408                         $this->beginMasterChanges( $fnameEffective );
 409                 }
 410         }
 411
 412         public function getChronologyProtectorTouched( $dbName ) {
 413                 return $this->getChronologyProtector()->getTouched( $dbName );
 414         }
 415
 416         public function disableChronologyProtection() {
 417                 $this->getChronologyProtector()->setEnabled( false );
 418         }
 419
 420         /**
 421          * @return ChronologyProtector
 422          */
 423         protected function getChronologyProtector() {
 424                 if ( $this->chronProt ) {
 425                         return $this->chronProt;
 426                 }
 427
 428                 $this->chronProt = new ChronologyProtector(
 429                         $this->memCache,
 430                         [
 431                                 'ip' => $this->requestInfo['IPAddress'],
 432                                 'agent' => $this->requestInfo['UserAgent'],
 433                         ],
 434                         isset( $_GET['cpPosTime'] ) ? $_GET['cpPosTime'] : null
 435                 );
 436                 $this->chronProt->setLogger( $this->replLogger );
 437
 438                 if ( $this->cliMode ) {
 439                         $this->chronProt->setEnabled( false );
 440                 } elseif ( $this->requestInfo['ChronologyProtection'] === 'false' ) {
 441                         // Request opted out of using position wait logic. This is useful for requests
 442                         // done by the job queue or background ETL that do not have a meaningful session.
 443                         $this->chronProt->setWaitEnabled( false );
 444                 }
 445
 446                 $this->replLogger->debug( __METHOD__ . ': using request info ' .
 447                         json_encode( $this->requestInfo, JSON_PRETTY_PRINT ) );
 448
 449                 return $this->chronProt;
 450         }
 451
 452         /**
 453          * Get and record all of the staged DB positions into persistent memory storage
 454          *
 455          * @param ChronologyProtector $cp
 456          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 457          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 458          */
 459         protected function shutdownChronologyProtector(
 460                 ChronologyProtector $cp, $workCallback, $mode
 461         ) {
 462                 // Record all the master positions needed
 463                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $cp ) {
 464                         $cp->shutdownLB( $lb );
 465                 } );
 466                 // Write them to the persistent stash. Try to do something useful by running $work
 467                 // while ChronologyProtector waits for the stash write to replicate to all DCs.
 468                 $unsavedPositions = $cp->shutdown( $workCallback, $mode );
 469                 if ( $unsavedPositions && $workCallback ) {
 470                         // Invoke callback in case it did not cache the result yet
 471                         $workCallback(); // work now to block for less time in waitForAll()
 472                 }
 473                 // If the positions failed to write to the stash, at least wait on local datacenter
 474                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 475                 // the chance that the user will see their own changes immediately afterwards. As long
 476                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 477                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $unsavedPositions ) {
 478                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 479                         if ( isset( $unsavedPositions[$masterName] ) ) {
 480                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 481                         }
 482                 } );
 483         }
 484
 485         /**
 486          * Base parameters to LoadBalancer::__construct()
 487          * @return array
 488          */
 489         final protected function baseLoadBalancerParams() {
 490                 return [
 491                         'localDomain' => $this->localDomain,
 492                         'readOnlyReason' => $this->readOnlyReason,
 493                         'srvCache' => $this->srvCache,
 494                         'wanCache' => $this->wanCache,
 495                         'profiler' => $this->profiler,
 496                         'trxProfiler' => $this->trxProfiler,
 497                         'queryLogger' => $this->queryLogger,
 498                         'connLogger' => $this->connLogger,
 499                         'replLogger' => $this->replLogger,
 500                         'errorLogger' => $this->errorLogger,
 501                         'hostname' => $this->hostname,
 502                         'cliMode' => $this->cliMode,
 503                         'agent' => $this->agent
 504                 ];
 505         }
 506
 507         /**
 508          * @param ILoadBalancer $lb
 509          */
 510         protected function initLoadBalancer( ILoadBalancer $lb ) {
 511                 if ( $this->trxRoundId !== false ) {
 512                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 513                 }
 514         }
 515
 516         public function setDomainPrefix( $prefix ) {
 517                 $this->localDomain = new DatabaseDomain(
 518                         $this->localDomain->getDatabase(),
 519                         null,
 520                         $prefix
 521                 );
 522
 523                 $this->forEachLB( function( ILoadBalancer $lb ) use ( $prefix ) {
 524                         $lb->setDomainPrefix( $prefix );
 525                 } );
 526         }
 527
 528         public function closeAll() {
 529                 $this->forEachLBCallMethod( 'closeAll', [] );
 530         }
 531
 532         public function setAgentName( $agent ) {
 533                 $this->agent = $agent;
 534         }
 535
 536         public function appendPreShutdownTimeAsQuery( $url, $time ) {
 537                 $usedCluster = 0;
 538                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$usedCluster ) {
 539                         $usedCluster |= ( $lb->getServerCount() > 1 );
 540                 } );
 541
 542                 if ( !$usedCluster ) {
 543                         return $url; // no master/replica clusters touched
 544                 }
 545
 546                 return strpos( $url, '?' ) === false ? "$url?cpPosTime=$time" : "$url&cpPosTime=$time";
 547         }
 548
 549         public function setRequestInfo( array $info ) {
 550                 $this->requestInfo = $info + $this->requestInfo;
 551         }
 552
 553         /**
 554          * Make PHP ignore user aborts/disconnects until the returned
 555          * value leaves scope. This returns null and does nothing in CLI mode.
 556          *
 557          * @return ScopedCallback|null
 558          */
 559         final protected function getScopedPHPBehaviorForCommit() {
 560                 if ( PHP_SAPI != 'cli' ) { // https://bugs.php.net/bug.php?id=47540
 561                         $old = ignore_user_abort( true ); // avoid half-finished operations
 562                         return new ScopedCallback( function () use ( $old ) {
 563                                 ignore_user_abort( $old );
 564                         } );
 565                 }
 566
 567                 return null;
 568         }
 569
 570         function __destruct() {
 571                 $this->destroy();
 572         }
 573 }