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 namespace Wikimedia\Rdbms;
  25
  26 use Psr\Log\LoggerInterface;
  27 use Wikimedia\ScopedCallback;
  28 use BagOStuff;
  29 use EmptyBagOStuff;
  30 use WANObjectCache;
  31 use DatabaseDomain;
  32 use Exception;
  33 use RuntimeException;
  34 use ILoadBalancer;
  35 use IDatabase;
  36 use LoadBalancer;
  37 use DBTransactionError;
  38 use DBReplicationWaitError;
  39
  40 /**
  41  * An interface for generating database load balancers
  42  * @ingroup Database
  43  */
  44 abstract class LBFactory implements ILBFactory {
  45         /** @var ChronologyProtector */
  46         protected $chronProt;
  47         /** @var object|string Class name or object With profileIn/profileOut methods */
  48         protected $profiler;
  49         /** @var TransactionProfiler */
  50         protected $trxProfiler;
  51         /** @var LoggerInterface */
  52         protected $replLogger;
  53         /** @var LoggerInterface */
  54         protected $connLogger;
  55         /** @var LoggerInterface */
  56         protected $queryLogger;
  57         /** @var LoggerInterface */
  58         protected $perfLogger;
  59         /** @var callable Error logger */
  60         protected $errorLogger;
  61         /** @var BagOStuff */
  62         protected $srvCache;
  63         /** @var BagOStuff */
  64         protected $memCache;
  65         /** @var WANObjectCache */
  66         protected $wanCache;
  67
  68         /** @var DatabaseDomain Local domain */
  69         protected $localDomain;
  70         /** @var string Local hostname of the app server */
  71         protected $hostname;
  72         /** @var array Web request information about the client */
  73         protected $requestInfo;
  74
  75         /** @var mixed */
  76         protected $ticket;
  77         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  78         protected $trxRoundId = false;
  79         /** @var string|bool Reason all LBs are read-only or false if not */
  80         protected $readOnlyReason = false;
  81         /** @var callable[] */
  82         protected $replicationWaitCallbacks = [];
  83
  84         /** @var bool Whether this PHP instance is for a CLI script */
  85         protected $cliMode;
  86         /** @var string Agent name for query profiling */
  87         protected $agent;
  88
  89         private static $loggerFields =
  90                 [ 'replLogger', 'connLogger', 'queryLogger', 'perfLogger' ];
  91
  92         public function __construct( array $conf ) {
  93                 $this->localDomain = isset( $conf['localDomain'] )
  94                         ? DatabaseDomain::newFromId( $conf['localDomain'] )
  95                         : DatabaseDomain::newUnspecified();
  96
  97                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  98                         $this->readOnlyReason = $conf['readOnlyReason'];
  99                 }
 100
 101                 $this->srvCache = isset( $conf['srvCache'] ) ? $conf['srvCache'] : new EmptyBagOStuff();
 102                 $this->memCache = isset( $conf['memCache'] ) ? $conf['memCache'] : new EmptyBagOStuff();
 103                 $this->wanCache = isset( $conf['wanCache'] )
 104                         ? $conf['wanCache']
 105                         : WANObjectCache::newEmpty();
 106
 107                 foreach ( self::$loggerFields as $key ) {
 108                         $this->$key = isset( $conf[$key] ) ? $conf[$key] : new \Psr\Log\NullLogger();
 109                 }
 110                 $this->errorLogger = isset( $conf['errorLogger'] )
 111                         ? $conf['errorLogger']
 112                         : function ( Exception $e ) {
 113                                 trigger_error( E_USER_WARNING, get_class( $e ) . ': ' . $e->getMessage() );
 114                         };
 115
 116                 $this->profiler = isset( $conf['profiler'] ) ? $conf['profiler'] : null;
 117                 $this->trxProfiler = isset( $conf['trxProfiler'] )
 118                         ? $conf['trxProfiler']
 119                         : new TransactionProfiler();
 120
 121                 $this->requestInfo = [
 122                         'IPAddress' => isset( $_SERVER[ 'REMOTE_ADDR' ] ) ? $_SERVER[ 'REMOTE_ADDR' ] : '',
 123                         'UserAgent' => isset( $_SERVER['HTTP_USER_AGENT'] ) ? $_SERVER['HTTP_USER_AGENT'] : '',
 124                         'ChronologyProtection' => 'true'
 125                 ];
 126
 127                 $this->cliMode = isset( $conf['cliMode'] ) ? $conf['cliMode'] : PHP_SAPI === 'cli';
 128                 $this->hostname = isset( $conf['hostname'] ) ? $conf['hostname'] : gethostname();
 129                 $this->agent = isset( $conf['agent'] ) ? $conf['agent'] : '';
 130
 131                 $this->ticket = mt_rand();
 132         }
 133
 134         public function destroy() {
 135                 $this->shutdown( self::SHUTDOWN_NO_CHRONPROT );
 136                 $this->forEachLBCallMethod( 'disable' );
 137         }
 138
 139         public function shutdown(
 140                 $mode = self::SHUTDOWN_CHRONPROT_SYNC, callable $workCallback = null
 141         ) {
 142                 $chronProt = $this->getChronologyProtector();
 143                 if ( $mode === self::SHUTDOWN_CHRONPROT_SYNC ) {
 144                         $this->shutdownChronologyProtector( $chronProt, $workCallback, 'sync' );
 145                 } elseif ( $mode === self::SHUTDOWN_CHRONPROT_ASYNC ) {
 146                         $this->shutdownChronologyProtector( $chronProt, null, 'async' );
 147                 }
 148
 149                 $this->commitMasterChanges( __METHOD__ ); // sanity
 150         }
 151
 152         /**
 153          * @see ILBFactory::newMainLB()
 154          * @param bool $domain
 155          * @return LoadBalancer
 156          */
 157         abstract public function newMainLB( $domain = false );
 158
 159         /**
 160          * @see ILBFactory::getMainLB()
 161          * @param bool $domain
 162          * @return LoadBalancer
 163          */
 164         abstract public function getMainLB( $domain = false );
 165
 166         /**
 167          * @see ILBFactory::newExternalLB()
 168          * @param string $cluster
 169          * @return LoadBalancer
 170          */
 171         abstract public function newExternalLB( $cluster );
 172
 173         /**
 174          * @see ILBFactory::getExternalLB()
 175          * @param string $cluster
 176          * @return LoadBalancer
 177          */
 178         abstract public function getExternalLB( $cluster );
 179
 180         /**
 181          * Call a method of each tracked load balancer
 182          *
 183          * @param string $methodName
 184          * @param array $args
 185          */
 186         protected function forEachLBCallMethod( $methodName, array $args = [] ) {
 187                 $this->forEachLB(
 188                         function ( ILoadBalancer $loadBalancer, $methodName, array $args ) {
 189                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 190                         },
 191                         [ $methodName, $args ]
 192                 );
 193         }
 194
 195         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 196                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 197         }
 198
 199         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 200                 $this->commitMasterChanges( $fname, $options );
 201                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 202         }
 203
 204         public function beginMasterChanges( $fname = __METHOD__ ) {
 205                 if ( $this->trxRoundId !== false ) {
 206                         throw new DBTransactionError(
 207                                 null,
 208                                 "$fname: transaction round '{$this->trxRoundId}' already started."
 209                         );
 210                 }
 211                 $this->trxRoundId = $fname;
 212                 // Set DBO_TRX flags on all appropriate DBs
 213                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 214         }
 215
 216         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 217                 if ( $this->trxRoundId !== false && $this->trxRoundId !== $fname ) {
 218                         throw new DBTransactionError(
 219                                 null,
 220                                 "$fname: transaction round '{$this->trxRoundId}' still running."
 221                         );
 222                 }
 223                 /** @noinspection PhpUnusedLocalVariableInspection */
 224                 $scope = $this->getScopedPHPBehaviorForCommit(); // try to ignore client aborts
 225                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 226                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 227                 $this->trxRoundId = false;
 228                 // Perform pre-commit checks, aborting on failure
 229                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 230                 // Log the DBs and methods involved in multi-DB transactions
 231                 $this->logIfMultiDbTransaction();
 232                 // Actually perform the commit on all master DB connections and revert DBO_TRX
 233                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 234                 // Run all post-commit callbacks
 235                 /** @var Exception $e */
 236                 $e = null; // first callback exception
 237                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$e ) {
 238                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 239                         $e = $e ?: $ex;
 240                 } );
 241                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 242                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 243                 // Throw any last post-commit callback error
 244                 if ( $e instanceof Exception ) {
 245                         throw $e;
 246                 }
 247         }
 248
 249         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 250                 $this->trxRoundId = false;
 251                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 252                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 253                 // Run all post-rollback callbacks
 254                 $this->forEachLB( function ( ILoadBalancer $lb ) {
 255                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 256                 } );
 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 ( ILoadBalancer $lb ) use ( &$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->queryLogger->info( $msg );
 279                 }
 280         }
 281
 282         public function hasMasterChanges() {
 283                 $ret = false;
 284                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 285                         $ret = $ret || $lb->hasMasterChanges();
 286                 } );
 287
 288                 return $ret;
 289         }
 290
 291         public function laggedReplicaUsed() {
 292                 $ret = false;
 293                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 294                         $ret = $ret || $lb->laggedReplicaUsed();
 295                 } );
 296
 297                 return $ret;
 298         }
 299
 300         public function hasOrMadeRecentMasterChanges( $age = null ) {
 301                 $ret = false;
 302                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $age, &$ret ) {
 303                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges( $age );
 304                 } );
 305                 return $ret;
 306         }
 307
 308         public function waitForReplication( array $opts = [] ) {
 309                 $opts += [
 310                         'domain' => false,
 311                         'cluster' => false,
 312                         'timeout' => 60,
 313                         'ifWritesSince' => null
 314                 ];
 315
 316                 if ( $opts['domain'] === false && isset( $opts['wiki'] ) ) {
 317                         $opts['domain'] = $opts['wiki']; // b/c
 318                 }
 319
 320                 // Figure out which clusters need to be checked
 321                 /** @var ILoadBalancer[] $lbs */
 322                 $lbs = [];
 323                 if ( $opts['cluster'] !== false ) {
 324                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 325                 } elseif ( $opts['domain'] !== false ) {
 326                         $lbs[] = $this->getMainLB( $opts['domain'] );
 327                 } else {
 328                         $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$lbs ) {
 329                                 $lbs[] = $lb;
 330                         } );
 331                         if ( !$lbs ) {
 332                                 return; // nothing actually used
 333                         }
 334                 }
 335
 336                 // Get all the master positions of applicable DBs right now.
 337                 // This can be faster since waiting on one cluster reduces the
 338                 // time needed to wait on the next clusters.
 339                 $masterPositions = array_fill( 0, count( $lbs ), false );
 340                 foreach ( $lbs as $i => $lb ) {
 341                         if ( $lb->getServerCount() <= 1 ) {
 342                                 // Bug 27975 - Don't try to wait for replica DBs if there are none
 343                                 // Prevents permission error when getting master position
 344                                 continue;
 345                         } elseif ( $opts['ifWritesSince']
 346                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 347                         ) {
 348                                 continue; // no writes since the last wait
 349                         }
 350                         $masterPositions[$i] = $lb->getMasterPos();
 351                 }
 352
 353                 // Run any listener callbacks *after* getting the DB positions. The more
 354                 // time spent in the callbacks, the less time is spent in waitForAll().
 355                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 356                         $callback();
 357                 }
 358
 359                 $failed = [];
 360                 foreach ( $lbs as $i => $lb ) {
 361                         if ( $masterPositions[$i] ) {
 362                                 // The DBMS may not support getMasterPos()
 363                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 364                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 365                                 }
 366                         }
 367                 }
 368
 369                 if ( $failed ) {
 370                         throw new DBReplicationWaitError(
 371                                 null,
 372                                 "Could not wait for replica DBs to catch up to " .
 373                                 implode( ', ', $failed )
 374                         );
 375                 }
 376         }
 377
 378         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 379                 if ( $callback ) {
 380                         $this->replicationWaitCallbacks[$name] = $callback;
 381                 } else {
 382                         unset( $this->replicationWaitCallbacks[$name] );
 383                 }
 384         }
 385
 386         public function getEmptyTransactionTicket( $fname ) {
 387                 if ( $this->hasMasterChanges() ) {
 388                         $this->queryLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 389                                 ( new RuntimeException() )->getTraceAsString() );
 390
 391                         return null;
 392                 }
 393
 394                 return $this->ticket;
 395         }
 396
 397         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 398                 if ( $ticket !== $this->ticket ) {
 399                         $this->perfLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 400                                 ( new RuntimeException() )->getTraceAsString() );
 401
 402                         return;
 403                 }
 404
 405                 // The transaction owner and any caller with the empty transaction ticket can commit
 406                 // so that getEmptyTransactionTicket() callers don't risk seeing DBTransactionError.
 407                 if ( $this->trxRoundId !== false && $fname !== $this->trxRoundId ) {
 408                         $this->queryLogger->info( "$fname: committing on behalf of {$this->trxRoundId}." );
 409                         $fnameEffective = $this->trxRoundId;
 410                 } else {
 411                         $fnameEffective = $fname;
 412                 }
 413
 414                 $this->commitMasterChanges( $fnameEffective );
 415                 $this->waitForReplication( $opts );
 416                 // If a nested caller committed on behalf of $fname, start another empty $fname
 417                 // transaction, leaving the caller with the same empty transaction state as before.
 418                 if ( $fnameEffective !== $fname ) {
 419                         $this->beginMasterChanges( $fnameEffective );
 420                 }
 421         }
 422
 423         public function getChronologyProtectorTouched( $dbName ) {
 424                 return $this->getChronologyProtector()->getTouched( $dbName );
 425         }
 426
 427         public function disableChronologyProtection() {
 428                 $this->getChronologyProtector()->setEnabled( false );
 429         }
 430
 431         /**
 432          * @return ChronologyProtector
 433          */
 434         protected function getChronologyProtector() {
 435                 if ( $this->chronProt ) {
 436                         return $this->chronProt;
 437                 }
 438
 439                 $this->chronProt = new ChronologyProtector(
 440                         $this->memCache,
 441                         [
 442                                 'ip' => $this->requestInfo['IPAddress'],
 443                                 'agent' => $this->requestInfo['UserAgent'],
 444                         ],
 445                         isset( $_GET['cpPosTime'] ) ? $_GET['cpPosTime'] : null
 446                 );
 447                 $this->chronProt->setLogger( $this->replLogger );
 448
 449                 if ( $this->cliMode ) {
 450                         $this->chronProt->setEnabled( false );
 451                 } elseif ( $this->requestInfo['ChronologyProtection'] === 'false' ) {
 452                         // Request opted out of using position wait logic. This is useful for requests
 453                         // done by the job queue or background ETL that do not have a meaningful session.
 454                         $this->chronProt->setWaitEnabled( false );
 455                 }
 456
 457                 $this->replLogger->debug( __METHOD__ . ': using request info ' .
 458                         json_encode( $this->requestInfo, JSON_PRETTY_PRINT ) );
 459
 460                 return $this->chronProt;
 461         }
 462
 463         /**
 464          * Get and record all of the staged DB positions into persistent memory storage
 465          *
 466          * @param ChronologyProtector $cp
 467          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 468          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 469          */
 470         protected function shutdownChronologyProtector(
 471                 ChronologyProtector $cp, $workCallback, $mode
 472         ) {
 473                 // Record all the master positions needed
 474                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $cp ) {
 475                         $cp->shutdownLB( $lb );
 476                 } );
 477                 // Write them to the persistent stash. Try to do something useful by running $work
 478                 // while ChronologyProtector waits for the stash write to replicate to all DCs.
 479                 $unsavedPositions = $cp->shutdown( $workCallback, $mode );
 480                 if ( $unsavedPositions && $workCallback ) {
 481                         // Invoke callback in case it did not cache the result yet
 482                         $workCallback(); // work now to block for less time in waitForAll()
 483                 }
 484                 // If the positions failed to write to the stash, at least wait on local datacenter
 485                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 486                 // the chance that the user will see their own changes immediately afterwards. As long
 487                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 488                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $unsavedPositions ) {
 489                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 490                         if ( isset( $unsavedPositions[$masterName] ) ) {
 491                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 492                         }
 493                 } );
 494         }
 495
 496         /**
 497          * Base parameters to LoadBalancer::__construct()
 498          * @return array
 499          */
 500         final protected function baseLoadBalancerParams() {
 501                 return [
 502                         'localDomain' => $this->localDomain,
 503                         'readOnlyReason' => $this->readOnlyReason,
 504                         'srvCache' => $this->srvCache,
 505                         'wanCache' => $this->wanCache,
 506                         'profiler' => $this->profiler,
 507                         'trxProfiler' => $this->trxProfiler,
 508                         'queryLogger' => $this->queryLogger,
 509                         'connLogger' => $this->connLogger,
 510                         'replLogger' => $this->replLogger,
 511                         'errorLogger' => $this->errorLogger,
 512                         'hostname' => $this->hostname,
 513                         'cliMode' => $this->cliMode,
 514                         'agent' => $this->agent
 515                 ];
 516         }
 517
 518         /**
 519          * @param ILoadBalancer $lb
 520          */
 521         protected function initLoadBalancer( ILoadBalancer $lb ) {
 522                 if ( $this->trxRoundId !== false ) {
 523                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 524                 }
 525         }
 526
 527         public function setDomainPrefix( $prefix ) {
 528                 $this->localDomain = new DatabaseDomain(
 529                         $this->localDomain->getDatabase(),
 530                         null,
 531                         $prefix
 532                 );
 533
 534                 $this->forEachLB( function( ILoadBalancer $lb ) use ( $prefix ) {
 535                         $lb->setDomainPrefix( $prefix );
 536                 } );
 537         }
 538
 539         public function closeAll() {
 540                 $this->forEachLBCallMethod( 'closeAll', [] );
 541         }
 542
 543         public function setAgentName( $agent ) {
 544                 $this->agent = $agent;
 545         }
 546
 547         public function appendPreShutdownTimeAsQuery( $url, $time ) {
 548                 $usedCluster = 0;
 549                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$usedCluster ) {
 550                         $usedCluster |= ( $lb->getServerCount() > 1 );
 551                 } );
 552
 553                 if ( !$usedCluster ) {
 554                         return $url; // no master/replica clusters touched
 555                 }
 556
 557                 return strpos( $url, '?' ) === false ? "$url?cpPosTime=$time" : "$url&cpPosTime=$time";
 558         }
 559
 560         public function setRequestInfo( array $info ) {
 561                 $this->requestInfo = $info + $this->requestInfo;
 562         }
 563
 564         /**
 565          * Make PHP ignore user aborts/disconnects until the returned
 566          * value leaves scope. This returns null and does nothing in CLI mode.
 567          *
 568          * @return ScopedCallback|null
 569          */
 570         final protected function getScopedPHPBehaviorForCommit() {
 571                 if ( PHP_SAPI != 'cli' ) { // https://bugs.php.net/bug.php?id=47540
 572                         $old = ignore_user_abort( true ); // avoid half-finished operations
 573                         return new ScopedCallback( function () use ( $old ) {
 574                                 ignore_user_abort( $old );
 575                         } );
 576                 }
 577
 578                 return null;
 579         }
 580
 581         function __destruct() {
 582                 $this->destroy();
 583         }
 584 }