includes/libs/rdbms/loadbalancer/ILoadBalancer.php

   1 <?php
   2 /**
   3  * Database load balancing interface
   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 namespace Wikimedia\Rdbms;
  24
  25 use Exception;
  26 use InvalidArgumentException;
  27
  28 /**
  29  * Database cluster connection, tracking, load balancing, and transaction manager interface
  30  *
  31  * A "cluster" is considered to be one master database and zero or more replica databases.
  32  * Typically, the replica DBs replicate from the master asynchronously. The first node in the
  33  * "servers" configuration array is always considered the "master". However, this class can still
  34  * be used when all or some of the "replica" DBs are multi-master peers of the master or even
  35  * when all the DBs are non-replicating clones of each other holding read-only data. Thus, the
  36  * role of "master" is in some cases merely nominal.
  37  *
  38  * By default, each DB server uses DBO_DEFAULT for its 'flags' setting, unless explicitly set
  39  * otherwise in configuration. DBO_DEFAULT behavior depends on whether 'cliMode' is set:
  40  *   - In CLI mode, the flag has no effect with regards to LoadBalancer.
  41  *   - In non-CLI mode, the flag causes implicit transactions to be used; the first query on
  42  *     a database starts a transaction on that database. The transactions are meant to remain
  43  *     pending until either commitMasterChanges() or rollbackMasterChanges() is called. The
  44  *     application must have some point where it calls commitMasterChanges() near the end of
  45  *     the PHP request.
  46  * Every iteration of beginMasterChanges()/commitMasterChanges() is called a "transaction round".
  47  * Rounds are useful on the master DB connections because they make single-DB (and by and large
  48  * multi-DB) updates in web requests all-or-nothing. Also, transactions on replica DBs are useful
  49  * when REPEATABLE-READ or SERIALIZABLE isolation is used because all foriegn keys and constraints
  50  * hold across separate queries in the DB transaction since the data appears within a consistent
  51  * point-in-time snapshot.
  52  *
  53  * The typical caller will use LoadBalancer::getConnection( DB_* ) to yield a live database
  54  * connection handle. The choice of which DB server to use is based on pre-defined loads for
  55  * weighted random selection, adjustments thereof by LoadMonitor, and the amount of replication
  56  * lag on each DB server. Lag checks might cause problems in certain setups, so they should be
  57  * tuned in the server configuration maps as follows:
  58  *   - Master + N Replica(s): set 'max lag' to an appropriate threshold for avoiding any database
  59  *      lagged by this much or more. If all DBs are this lagged, then the load balancer considers
  60  *      the cluster to be read-only.
  61  *   - Galera Cluster: Seconds_Behind_Master will be 0, so there probably is nothing to tune.
  62  *      Note that lag is still possible depending on how wsrep-sync-wait is set server-side.
  63  *   - Read-only archive clones: set 'is static' in the server configuration maps. This will
  64  *      treat all such DBs as having 0 lag.
  65  *   - SQL load balancing proxy: any proxy should handle lag checks on its own, so the 'max lag'
  66  *      parameter should probably be set to INF in the server configuration maps. This will make
  67  *      the load balancer ignore whatever it detects as the lag of the logical replica is (which
  68  *      would probably just randomly bounce around).
  69  *
  70  * If using a SQL proxy service, it would probably be best to have two proxy hosts for the
  71  * load balancer to talk to. One would be the 'host' of the master server entry and another for
  72  * the (logical) replica server entry. The proxy could map the load balancer's "replica" DB to
  73  * any number of physical replica DBs.
  74  *
  75  * @since 1.28
  76  * @ingroup Database
  77  */
  78 interface ILoadBalancer {
  79         /** @var int Request a replica DB connection */
  80         const DB_REPLICA = -1;
  81         /** @var int Request a master DB connection */
  82         const DB_MASTER = -2;
  83
  84         /** @var string Domain specifier when no specific database needs to be selected */
  85         const DOMAIN_ANY = '';
  86
  87         /** @var int DB handle should have DBO_TRX disabled and the caller will leave it as such */
  88         const CONN_TRX_AUTO = 1;
  89
  90         /**
  91          * Construct a manager of IDatabase connection objects
  92          *
  93          * @param array $params Parameter map with keys:
  94          *  - servers : Required. Array of server info structures.
  95          *  - localDomain: A DatabaseDomain or domain ID string.
  96          *  - loadMonitor : Name of a class used to fetch server lag and load.
  97          *  - readOnlyReason : Reason the master DB is read-only if so [optional]
  98          *  - waitTimeout : Maximum time to wait for replicas for consistency [optional]
  99          *  - maxLag: Avoid replica DB servers with more lag than this [optional]
 100          *  - srvCache : BagOStuff object for server cache [optional]
 101          *  - wanCache : WANObjectCache object [optional]
 102          *  - chronologyProtector: ChronologyProtector object [optional]
 103          *  - hostname : The name of the current server [optional]
 104          *  - cliMode: Whether the execution context is a CLI script. [optional]
 105          *  - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
 106          *  - trxProfiler: TransactionProfiler instance. [optional]
 107          *  - replLogger: PSR-3 logger instance. [optional]
 108          *  - connLogger: PSR-3 logger instance. [optional]
 109          *  - queryLogger: PSR-3 logger instance. [optional]
 110          *  - perfLogger: PSR-3 logger instance. [optional]
 111          *  - errorLogger : Callback that takes an Exception and logs it. [optional]
 112          * @throws InvalidArgumentException
 113          */
 114         public function __construct( array $params );
 115
 116         /**
 117          * Get the index of the reader connection, which may be a replica DB
 118          *
 119          * This takes into account load ratios and lag times. It should
 120          * always return a consistent index during a given invocation.
 121          *
 122          * Side effect: opens connections to databases
 123          * @param string|bool $group Query group, or false for the generic reader
 124          * @param string|bool $domain Domain ID, or false for the current domain
 125          * @throws DBError
 126          * @return bool|int|string
 127          */
 128         public function getReaderIndex( $group = false, $domain = false );
 129
 130         /**
 131          * Set the master wait position
 132          *
 133          * If a DB_REPLICA connection has been opened already, then wait immediately.
 134          * Otherwise sets a variable telling it to wait if such a connection is opened.
 135          *
 136          * This only applies to connections to the generic replica DB for this request.
 137          * If a timeout happens when waiting, then getLaggedReplicaMode()/laggedReplicaUsed()
 138          * will return true.
 139          *
 140          * @param DBMasterPos|bool $pos Master position or false
 141          */
 142         public function waitFor( $pos );
 143
 144         /**
 145          * Set the master wait position and wait for a "generic" replica DB to catch up to it
 146          *
 147          * This can be used a faster proxy for waitForAll()
 148          *
 149          * @param DBMasterPos|bool $pos Master position or false
 150          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 151          * @return bool Success (able to connect and no timeouts reached)
 152          */
 153         public function waitForOne( $pos, $timeout = null );
 154
 155         /**
 156          * Set the master wait position and wait for ALL replica DBs to catch up to it
 157          *
 158          * @param DBMasterPos|bool $pos Master position or false
 159          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 160          * @return bool Success (able to connect and no timeouts reached)
 161          */
 162         public function waitForAll( $pos, $timeout = null );
 163
 164         /**
 165          * Get any open connection to a given server index, local or foreign
 166          *
 167          * @param int $i Server index or DB_MASTER/DB_REPLICA
 168          * @return Database|bool False if no such connection is open
 169          */
 170         public function getAnyOpenConnection( $i );
 171
 172         /**
 173          * Get a connection handle by server index
 174          *
 175          * The CONN_TRX_AUTO flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
 176          * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
 177          * can be used to check such flags beforehand.
 178          *
 179          * If the caller uses $domain or sets CONN_TRX_AUTO in $flags, then it must also
 180          * call ILoadBalancer::reuseConnection() on the handle when finished using it.
 181          * In all other cases, this is not necessary, though not harmful either.
 182          *
 183          * @param int $i Server index or DB_MASTER/DB_REPLICA
 184          * @param array|string|bool $groups Query group(s), or false for the generic reader
 185          * @param string|bool $domain Domain ID, or false for the current domain
 186          * @param int $flags Bitfield of CONN_* class constants
 187          *
 188          * @note This method throws DBAccessError if ILoadBalancer::disable() was called
 189          *
 190          * @throws DBError
 191          * @return Database
 192          */
 193         public function getConnection( $i, $groups = [], $domain = false, $flags = 0 );
 194
 195         /**
 196          * Mark a foreign connection as being available for reuse under a different DB domain
 197          *
 198          * This mechanism is reference-counted, and must be called the same number of times
 199          * as getConnection() to work.
 200          *
 201          * @param IDatabase $conn
 202          * @throws InvalidArgumentException
 203          */
 204         public function reuseConnection( IDatabase $conn );
 205
 206         /**
 207          * Get a database connection handle reference
 208          *
 209          * The handle's methods simply wrap those of a Database handle
 210          *
 211          * The CONN_TRX_AUTO flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
 212          * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
 213          * can be used to check such flags beforehand.
 214          *
 215          * @see ILoadBalancer::getConnection() for parameter information
 216          *
 217          * @param int $i Server index or DB_MASTER/DB_REPLICA
 218          * @param array|string|bool $groups Query group(s), or false for the generic reader
 219          * @param string|bool $domain Domain ID, or false for the current domain
 220          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 221          * @return DBConnRef
 222          */
 223         public function getConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
 224
 225         /**
 226          * Get a database connection handle reference without connecting yet
 227          *
 228          * The handle's methods simply wrap those of a Database handle
 229          *
 230          * The CONN_TRX_AUTO flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
 231          * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
 232          * can be used to check such flags beforehand.
 233          *
 234          * @see ILoadBalancer::getConnection() for parameter information
 235          *
 236          * @param int $i Server index or DB_MASTER/DB_REPLICA
 237          * @param array|string|bool $groups Query group(s), or false for the generic reader
 238          * @param string|bool $domain Domain ID, or false for the current domain
 239          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 240          * @return DBConnRef
 241          */
 242         public function getLazyConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
 243
 244         /**
 245          * Get a maintenance database connection handle reference for migrations and schema changes
 246          *
 247          * The handle's methods simply wrap those of a Database handle
 248          *
 249          * The CONN_TRX_AUTO flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
 250          * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
 251          * can be used to check such flags beforehand.
 252          *
 253          * @see ILoadBalancer::getConnection() for parameter information
 254          *
 255          * @param int $db Server index or DB_MASTER/DB_REPLICA
 256          * @param array|string|bool $groups Query group(s), or false for the generic reader
 257          * @param string|bool $domain Domain ID, or false for the current domain
 258          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 259          * @return MaintainableDBConnRef
 260          */
 261         public function getMaintenanceConnectionRef( $db, $groups = [], $domain = false, $flags = 0 );
 262
 263         /**
 264          * Open a connection to the server given by the specified index
 265          *
 266          * The index must be an actual index into the array. If a connection to the server is
 267          * already open and not considered an "in use" foreign connection, this simply returns it.
 268          *
 269          * Avoid using CONN_TRX_AUTO for databases with ATTR_DB_LEVEL_LOCKING (e.g. sqlite) in
 270          * order to avoid deadlocks. ILoadBalancer::getServerAttributes() can be used to check
 271          * such flags beforehand.
 272          *
 273          * If the caller uses $domain or sets CONN_TRX_AUTO in $flags, then it must also
 274          * call ILoadBalancer::reuseConnection() on the handle when finished using it.
 275          * In all other cases, this is not necessary, though not harmful either.
 276          *
 277          * @note This method throws DBAccessError if ILoadBalancer::disable() was called
 278          *
 279          * @param int $i Server index (does not support DB_MASTER/DB_REPLICA)
 280          * @param string|bool $domain Domain ID, or false for the current domain
 281          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 282          * @return Database|bool Returns false on errors
 283          * @throws DBAccessError
 284          */
 285         public function openConnection( $i, $domain = false, $flags = 0 );
 286
 287         /**
 288          * @return int
 289          */
 290         public function getWriterIndex();
 291
 292         /**
 293          * Returns true if the specified index is a valid server index
 294          *
 295          * @param int $i
 296          * @return bool
 297          */
 298         public function haveIndex( $i );
 299
 300         /**
 301          * Returns true if the specified index is valid and has non-zero load
 302          *
 303          * @param int $i
 304          * @return bool
 305          */
 306         public function isNonZeroLoad( $i );
 307
 308         /**
 309          * Get the number of defined servers (not the number of open connections)
 310          *
 311          * @return int
 312          */
 313         public function getServerCount();
 314
 315         /**
 316          * Get the host name or IP address of the server with the specified index
 317          *
 318          * @param int $i
 319          * @return string Readable name if available or IP/host otherwise
 320          */
 321         public function getServerName( $i );
 322
 323         /**
 324          * Get DB type of the server with the specified index
 325          *
 326          * @param int $i
 327          * @return string One of (mysql,postgres,sqlite,...) or "unknown" for bad indexes
 328          * @since 1.30
 329          */
 330         public function getServerType( $i );
 331
 332         /**
 333          * @param int $i Server index
 334          * @return array (Database::ATTRIBUTE_* constant => value) for all such constants
 335          * @since 1.31
 336          */
 337         public function getServerAttributes( $i );
 338
 339         /**
 340          * Get the current master position for chronology control purposes
 341          * @return DBMasterPos|bool Returns false if not applicable
 342          */
 343         public function getMasterPos();
 344
 345         /**
 346          * Disable this load balancer. All connections are closed, and any attempt to
 347          * open a new connection will result in a DBAccessError.
 348          */
 349         public function disable();
 350
 351         /**
 352          * Close all open connections
 353          */
 354         public function closeAll();
 355
 356         /**
 357          * Close a connection
 358          *
 359          * Using this function makes sure the LoadBalancer knows the connection is closed.
 360          * If you use $conn->close() directly, the load balancer won't update its state.
 361          *
 362          * @param IDatabase $conn
 363          */
 364         public function closeConnection( IDatabase $conn );
 365
 366         /**
 367          * Commit transactions on all open connections
 368          * @param string $fname Caller name
 369          * @throws DBExpectedError
 370          */
 371         public function commitAll( $fname = __METHOD__ );
 372
 373         /**
 374          * Perform all pre-commit callbacks that remain part of the atomic transactions
 375          * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
 376          *
 377          * Use this only for mutli-database commits
 378          */
 379         public function finalizeMasterChanges();
 380
 381         /**
 382          * Perform all pre-commit checks for things like replication safety
 383          *
 384          * Use this only for mutli-database commits
 385          *
 386          * @param array $options Includes:
 387          *   - maxWriteDuration : max write query duration time in seconds
 388          * @throws DBTransactionError
 389          */
 390         public function approveMasterChanges( array $options );
 391
 392         /**
 393          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 394          *
 395          * The DBO_TRX setting will be reverted to the default in each of these methods:
 396          *   - commitMasterChanges()
 397          *   - rollbackMasterChanges()
 398          *   - commitAll()
 399          * This allows for custom transaction rounds from any outer transaction scope.
 400          *
 401          * @param string $fname
 402          * @throws DBExpectedError
 403          */
 404         public function beginMasterChanges( $fname = __METHOD__ );
 405
 406         /**
 407          * Issue COMMIT on all master connections where writes where done
 408          * @param string $fname Caller name
 409          * @throws DBExpectedError
 410          */
 411         public function commitMasterChanges( $fname = __METHOD__ );
 412
 413         /**
 414          * Issue all pending post-COMMIT/ROLLBACK callbacks
 415          *
 416          * Use this only for mutli-database commits
 417          *
 418          * @param int $type IDatabase::TRIGGER_* constant
 419          * @return Exception|null The first exception or null if there were none
 420          */
 421         public function runMasterPostTrxCallbacks( $type );
 422
 423         /**
 424          * Issue ROLLBACK only on master, only if queries were done on connection
 425          * @param string $fname Caller name
 426          * @throws DBExpectedError
 427          */
 428         public function rollbackMasterChanges( $fname = __METHOD__ );
 429
 430         /**
 431          * Suppress all pending post-COMMIT/ROLLBACK callbacks
 432          *
 433          * Use this only for mutli-database commits
 434          *
 435          * @return Exception|null The first exception or null if there were none
 436          */
 437         public function suppressTransactionEndCallbacks();
 438
 439         /**
 440          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 441          *
 442          * @param string $fname Caller name
 443          */
 444         public function flushReplicaSnapshots( $fname = __METHOD__ );
 445
 446         /**
 447          * @return bool Whether a master connection is already open
 448          */
 449         public function hasMasterConnection();
 450
 451         /**
 452          * Determine if there are pending changes in a transaction by this thread
 453          * @return bool
 454          */
 455         public function hasMasterChanges();
 456
 457         /**
 458          * Get the timestamp of the latest write query done by this thread
 459          * @return float|bool UNIX timestamp or false
 460          */
 461         public function lastMasterChangeTimestamp();
 462
 463         /**
 464          * Check if this load balancer object had any recent or still
 465          * pending writes issued against it by this PHP thread
 466          *
 467          * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
 468          * @return bool
 469          */
 470         public function hasOrMadeRecentMasterChanges( $age = null );
 471
 472         /**
 473          * Get the list of callers that have pending master changes
 474          *
 475          * @return string[] List of method names
 476          */
 477         public function pendingMasterChangeCallers();
 478
 479         /**
 480          * @note This method will trigger a DB connection if not yet done
 481          * @param string|bool $domain Domain ID, or false for the current domain
 482          * @return bool Whether the database for generic connections this request is highly "lagged"
 483          */
 484         public function getLaggedReplicaMode( $domain = false );
 485
 486         /**
 487          * Checks whether the database for generic connections this request was both:
 488          *   - a) Already choosen due to a prior connection attempt
 489          *   - b) Considered highly "lagged"
 490          *
 491          * @note This method will never cause a new DB connection
 492          * @return bool
 493          */
 494         public function laggedReplicaUsed();
 495
 496         /**
 497          * @note This method may trigger a DB connection if not yet done
 498          * @param string|bool $domain Domain ID, or false for the current domain
 499          * @param IDatabase|null $conn DB master connection; used to avoid loops [optional]
 500          * @return string|bool Reason the master is read-only or false if it is not
 501          */
 502         public function getReadOnlyReason( $domain = false, IDatabase $conn = null );
 503
 504         /**
 505          * Disables/enables lag checks
 506          * @param null|bool $mode
 507          * @return bool
 508          */
 509         public function allowLagged( $mode = null );
 510
 511         /**
 512          * @return bool
 513          */
 514         public function pingAll();
 515
 516         /**
 517          * Call a function with each open connection object
 518          * @param callable $callback
 519          * @param array $params
 520          */
 521         public function forEachOpenConnection( $callback, array $params = [] );
 522
 523         /**
 524          * Call a function with each open connection object to a master
 525          * @param callable $callback
 526          * @param array $params
 527          */
 528         public function forEachOpenMasterConnection( $callback, array $params = [] );
 529
 530         /**
 531          * Call a function with each open replica DB connection object
 532          * @param callable $callback
 533          * @param array $params
 534          */
 535         public function forEachOpenReplicaConnection( $callback, array $params = [] );
 536
 537         /**
 538          * Get the hostname and lag time of the most-lagged replica DB
 539          *
 540          * This is useful for maintenance scripts that need to throttle their updates.
 541          * May attempt to open connections to replica DBs on the default DB. If there is
 542          * no lag, the maximum lag will be reported as -1.
 543          *
 544          * @param bool|string $domain Domain ID, or false for the default database
 545          * @return array ( host, max lag, index of max lagged host )
 546          */
 547         public function getMaxLag( $domain = false );
 548
 549         /**
 550          * Get an estimate of replication lag (in seconds) for each server
 551          *
 552          * Results are cached for a short time in memcached/process cache
 553          *
 554          * Values may be "false" if replication is too broken to estimate
 555          *
 556          * @param string|bool $domain
 557          * @return int[] Map of (server index => float|int|bool)
 558          */
 559         public function getLagTimes( $domain = false );
 560
 561         /**
 562          * Get the lag in seconds for a given connection, or zero if this load
 563          * balancer does not have replication enabled.
 564          *
 565          * This should be used in preference to Database::getLag() in cases where
 566          * replication may not be in use, since there is no way to determine if
 567          * replication is in use at the connection level without running
 568          * potentially restricted queries such as SHOW SLAVE STATUS. Using this
 569          * function instead of Database::getLag() avoids a fatal error in this
 570          * case on many installations.
 571          *
 572          * @param IDatabase $conn
 573          * @return int|bool Returns false on error
 574          */
 575         public function safeGetLag( IDatabase $conn );
 576
 577         /**
 578          * Wait for a replica DB to reach a specified master position
 579          *
 580          * This will connect to the master to get an accurate position if $pos is not given
 581          *
 582          * @param IDatabase $conn Replica DB
 583          * @param DBMasterPos|bool $pos Master position; default: current position
 584          * @param int $timeout Timeout in seconds [optional]
 585          * @return bool Success
 586          */
 587         public function safeWaitForMasterPos( IDatabase $conn, $pos = false, $timeout = 10 );
 588
 589         /**
 590          * Set a callback via IDatabase::setTransactionListener() on
 591          * all current and future master connections of this load balancer
 592          *
 593          * @param string $name Callback name
 594          * @param callable|null $callback
 595          */
 596         public function setTransactionListener( $name, callable $callback = null );
 597
 598         /**
 599          * Set a new table prefix for the existing local domain ID for testing
 600          *
 601          * @param string $prefix
 602          */
 603         public function setDomainPrefix( $prefix );
 604
 605         /**
 606          * Make certain table names use their own database, schema, and table prefix
 607          * when passed into SQL queries pre-escaped and without a qualified database name
 608          *
 609          * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
 610          * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
 611          *
 612          * Calling this twice will completely clear any old table aliases. Also, note that
 613          * callers are responsible for making sure the schemas and databases actually exist.
 614          *
 615          * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
 616          */
 617         public function setTableAliases( array $aliases );
 618 }