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          * Avoid using CONN_TRX_AUTO with sqlite (e.g. check getServerType() first)
 176          *
 177          * If the caller uses $domain or sets CONN_TRX_AUTO in $flags, then it must also
 178          * call ILoadBalancer::reuseConnection() on the handle when finished using it.
 179          * In all other cases, this is not necessary, though not harmful either.
 180          *
 181          * @param int $i Server index or DB_MASTER/DB_REPLICA
 182          * @param array|string|bool $groups Query group(s), or false for the generic reader
 183          * @param string|bool $domain Domain ID, or false for the current domain
 184          * @param int $flags Bitfield of CONN_* class constants
 185          *
 186          * @note This method throws DBAccessError if ILoadBalancer::disable() was called
 187          *
 188          * @throws DBError
 189          * @return Database
 190          */
 191         public function getConnection( $i, $groups = [], $domain = false, $flags = 0 );
 192
 193         /**
 194          * Mark a foreign connection as being available for reuse under a different DB domain
 195          *
 196          * This mechanism is reference-counted, and must be called the same number of times
 197          * as getConnection() to work.
 198          *
 199          * @param IDatabase $conn
 200          * @throws InvalidArgumentException
 201          */
 202         public function reuseConnection( $conn );
 203
 204         /**
 205          * Get a database connection handle reference
 206          *
 207          * The handle's methods simply wrap those of a Database handle
 208          *
 209          * Avoid using CONN_TRX_AUTO with sqlite (e.g. check getServerType() first)
 210          *
 211          * @see ILoadBalancer::getConnection() for parameter information
 212          *
 213          * @param int $i Server index or DB_MASTER/DB_REPLICA
 214          * @param array|string|bool $groups Query group(s), or false for the generic reader
 215          * @param string|bool $domain Domain ID, or false for the current domain
 216          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 217          * @return DBConnRef
 218          */
 219         public function getConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
 220
 221         /**
 222          * Get a database connection handle reference without connecting yet
 223          *
 224          * The handle's methods simply wrap those of a Database handle
 225          *
 226          * Avoid using CONN_TRX_AUTO with sqlite (e.g. check getServerType() first)
 227          *
 228          * @see ILoadBalancer::getConnection() for parameter information
 229          *
 230          * @param int $i Server index or DB_MASTER/DB_REPLICA
 231          * @param array|string|bool $groups Query group(s), or false for the generic reader
 232          * @param string|bool $domain Domain ID, or false for the current domain
 233          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 234          * @return DBConnRef
 235          */
 236         public function getLazyConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
 237
 238         /**
 239          * Get a maintenance database connection handle reference for migrations and schema changes
 240          *
 241          * The handle's methods simply wrap those of a Database handle
 242          *
 243          * Avoid using CONN_TRX_AUTO with sqlite (e.g. check getServerType() first)
 244          *
 245          * @see ILoadBalancer::getConnection() for parameter information
 246          *
 247          * @param int $db Server index or DB_MASTER/DB_REPLICA
 248          * @param array|string|bool $groups Query group(s), or false for the generic reader
 249          * @param string|bool $domain Domain ID, or false for the current domain
 250          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 251          * @return MaintainableDBConnRef
 252          */
 253         public function getMaintenanceConnectionRef( $db, $groups = [], $domain = false, $flags = 0 );
 254
 255         /**
 256          * Open a connection to the server given by the specified index
 257          *
 258          * The index must be an actual index into the array. If a connection to the server is
 259          * already open and not considered an "in use" foreign connection, this simply returns it.
 260          *
 261          * Avoid using CONN_TRX_AUTO with sqlite (e.g. check getServerType() first)
 262          *
 263          * If the caller uses $domain or sets CONN_TRX_AUTO in $flags, then it must also
 264          * call ILoadBalancer::reuseConnection() on the handle when finished using it.
 265          * In all other cases, this is not necessary, though not harmful either.
 266          *
 267          * @note This method throws DBAccessError if ILoadBalancer::disable() was called
 268          *
 269          * @param int $i Server index (does not support DB_MASTER/DB_REPLICA)
 270          * @param string|bool $domain Domain ID, or false for the current domain
 271          * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTO)
 272          * @return Database|bool Returns false on errors
 273          * @throws DBAccessError
 274          */
 275         public function openConnection( $i, $domain = false, $flags = 0 );
 276
 277         /**
 278          * @return int
 279          */
 280         public function getWriterIndex();
 281
 282         /**
 283          * Returns true if the specified index is a valid server index
 284          *
 285          * @param int $i
 286          * @return bool
 287          */
 288         public function haveIndex( $i );
 289
 290         /**
 291          * Returns true if the specified index is valid and has non-zero load
 292          *
 293          * @param int $i
 294          * @return bool
 295          */
 296         public function isNonZeroLoad( $i );
 297
 298         /**
 299          * Get the number of defined servers (not the number of open connections)
 300          *
 301          * @return int
 302          */
 303         public function getServerCount();
 304
 305         /**
 306          * Get the host name or IP address of the server with the specified index
 307          *
 308          * @param int $i
 309          * @return string Readable name if available or IP/host otherwise
 310          */
 311         public function getServerName( $i );
 312
 313         /**
 314          * Get DB type of the server with the specified index
 315          *
 316          * @param int $i
 317          * @return string One of (mysql,postgres,sqlite,...) or "unknown" for bad indexes
 318          * @since 1.30
 319          */
 320         public function getServerType( $i );
 321
 322         /**
 323          * Get the current master position for chronology control purposes
 324          * @return DBMasterPos|bool Returns false if not applicable
 325          */
 326         public function getMasterPos();
 327
 328         /**
 329          * Disable this load balancer. All connections are closed, and any attempt to
 330          * open a new connection will result in a DBAccessError.
 331          */
 332         public function disable();
 333
 334         /**
 335          * Close all open connections
 336          */
 337         public function closeAll();
 338
 339         /**
 340          * Close a connection
 341          *
 342          * Using this function makes sure the LoadBalancer knows the connection is closed.
 343          * If you use $conn->close() directly, the load balancer won't update its state.
 344          *
 345          * @param IDatabase $conn
 346          */
 347         public function closeConnection( IDatabase $conn );
 348
 349         /**
 350          * Commit transactions on all open connections
 351          * @param string $fname Caller name
 352          * @throws DBExpectedError
 353          */
 354         public function commitAll( $fname = __METHOD__ );
 355
 356         /**
 357          * Perform all pre-commit callbacks that remain part of the atomic transactions
 358          * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
 359          *
 360          * Use this only for mutli-database commits
 361          */
 362         public function finalizeMasterChanges();
 363
 364         /**
 365          * Perform all pre-commit checks for things like replication safety
 366          *
 367          * Use this only for mutli-database commits
 368          *
 369          * @param array $options Includes:
 370          *   - maxWriteDuration : max write query duration time in seconds
 371          * @throws DBTransactionError
 372          */
 373         public function approveMasterChanges( array $options );
 374
 375         /**
 376          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 377          *
 378          * The DBO_TRX setting will be reverted to the default in each of these methods:
 379          *   - commitMasterChanges()
 380          *   - rollbackMasterChanges()
 381          *   - commitAll()
 382          * This allows for custom transaction rounds from any outer transaction scope.
 383          *
 384          * @param string $fname
 385          * @throws DBExpectedError
 386          */
 387         public function beginMasterChanges( $fname = __METHOD__ );
 388
 389         /**
 390          * Issue COMMIT on all master connections where writes where done
 391          * @param string $fname Caller name
 392          * @throws DBExpectedError
 393          */
 394         public function commitMasterChanges( $fname = __METHOD__ );
 395
 396         /**
 397          * Issue all pending post-COMMIT/ROLLBACK callbacks
 398          *
 399          * Use this only for mutli-database commits
 400          *
 401          * @param int $type IDatabase::TRIGGER_* constant
 402          * @return Exception|null The first exception or null if there were none
 403          */
 404         public function runMasterPostTrxCallbacks( $type );
 405
 406         /**
 407          * Issue ROLLBACK only on master, only if queries were done on connection
 408          * @param string $fname Caller name
 409          * @throws DBExpectedError
 410          */
 411         public function rollbackMasterChanges( $fname = __METHOD__ );
 412
 413         /**
 414          * Suppress all pending post-COMMIT/ROLLBACK callbacks
 415          *
 416          * Use this only for mutli-database commits
 417          *
 418          * @return Exception|null The first exception or null if there were none
 419          */
 420         public function suppressTransactionEndCallbacks();
 421
 422         /**
 423          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 424          *
 425          * @param string $fname Caller name
 426          */
 427         public function flushReplicaSnapshots( $fname = __METHOD__ );
 428
 429         /**
 430          * @return bool Whether a master connection is already open
 431          */
 432         public function hasMasterConnection();
 433
 434         /**
 435          * Determine if there are pending changes in a transaction by this thread
 436          * @return bool
 437          */
 438         public function hasMasterChanges();
 439
 440         /**
 441          * Get the timestamp of the latest write query done by this thread
 442          * @return float|bool UNIX timestamp or false
 443          */
 444         public function lastMasterChangeTimestamp();
 445
 446         /**
 447          * Check if this load balancer object had any recent or still
 448          * pending writes issued against it by this PHP thread
 449          *
 450          * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
 451          * @return bool
 452          */
 453         public function hasOrMadeRecentMasterChanges( $age = null );
 454
 455         /**
 456          * Get the list of callers that have pending master changes
 457          *
 458          * @return string[] List of method names
 459          */
 460         public function pendingMasterChangeCallers();
 461
 462         /**
 463          * @note This method will trigger a DB connection if not yet done
 464          * @param string|bool $domain Domain ID, or false for the current domain
 465          * @return bool Whether the database for generic connections this request is highly "lagged"
 466          */
 467         public function getLaggedReplicaMode( $domain = false );
 468
 469         /**
 470          * Checks whether the database for generic connections this request was both:
 471          *   - a) Already choosen due to a prior connection attempt
 472          *   - b) Considered highly "lagged"
 473          *
 474          * @note This method will never cause a new DB connection
 475          * @return bool
 476          */
 477         public function laggedReplicaUsed();
 478
 479         /**
 480          * @note This method may trigger a DB connection if not yet done
 481          * @param string|bool $domain Domain ID, or false for the current domain
 482          * @param IDatabase|null $conn DB master connection; used to avoid loops [optional]
 483          * @return string|bool Reason the master is read-only or false if it is not
 484          */
 485         public function getReadOnlyReason( $domain = false, IDatabase $conn = null );
 486
 487         /**
 488          * Disables/enables lag checks
 489          * @param null|bool $mode
 490          * @return bool
 491          */
 492         public function allowLagged( $mode = null );
 493
 494         /**
 495          * @return bool
 496          */
 497         public function pingAll();
 498
 499         /**
 500          * Call a function with each open connection object
 501          * @param callable $callback
 502          * @param array $params
 503          */
 504         public function forEachOpenConnection( $callback, array $params = [] );
 505
 506         /**
 507          * Call a function with each open connection object to a master
 508          * @param callable $callback
 509          * @param array $params
 510          */
 511         public function forEachOpenMasterConnection( $callback, array $params = [] );
 512
 513         /**
 514          * Call a function with each open replica DB connection object
 515          * @param callable $callback
 516          * @param array $params
 517          */
 518         public function forEachOpenReplicaConnection( $callback, array $params = [] );
 519
 520         /**
 521          * Get the hostname and lag time of the most-lagged replica DB
 522          *
 523          * This is useful for maintenance scripts that need to throttle their updates.
 524          * May attempt to open connections to replica DBs on the default DB. If there is
 525          * no lag, the maximum lag will be reported as -1.
 526          *
 527          * @param bool|string $domain Domain ID, or false for the default database
 528          * @return array ( host, max lag, index of max lagged host )
 529          */
 530         public function getMaxLag( $domain = false );
 531
 532         /**
 533          * Get an estimate of replication lag (in seconds) for each server
 534          *
 535          * Results are cached for a short time in memcached/process cache
 536          *
 537          * Values may be "false" if replication is too broken to estimate
 538          *
 539          * @param string|bool $domain
 540          * @return int[] Map of (server index => float|int|bool)
 541          */
 542         public function getLagTimes( $domain = false );
 543
 544         /**
 545          * Get the lag in seconds for a given connection, or zero if this load
 546          * balancer does not have replication enabled.
 547          *
 548          * This should be used in preference to Database::getLag() in cases where
 549          * replication may not be in use, since there is no way to determine if
 550          * replication is in use at the connection level without running
 551          * potentially restricted queries such as SHOW SLAVE STATUS. Using this
 552          * function instead of Database::getLag() avoids a fatal error in this
 553          * case on many installations.
 554          *
 555          * @param IDatabase $conn
 556          * @return int|bool Returns false on error
 557          */
 558         public function safeGetLag( IDatabase $conn );
 559
 560         /**
 561          * Wait for a replica DB to reach a specified master position
 562          *
 563          * This will connect to the master to get an accurate position if $pos is not given
 564          *
 565          * @param IDatabase $conn Replica DB
 566          * @param DBMasterPos|bool $pos Master position; default: current position
 567          * @param int $timeout Timeout in seconds [optional]
 568          * @return bool Success
 569          */
 570         public function safeWaitForMasterPos( IDatabase $conn, $pos = false, $timeout = 10 );
 571
 572         /**
 573          * Set a callback via IDatabase::setTransactionListener() on
 574          * all current and future master connections of this load balancer
 575          *
 576          * @param string $name Callback name
 577          * @param callable|null $callback
 578          */
 579         public function setTransactionListener( $name, callable $callback = null );
 580
 581         /**
 582          * Set a new table prefix for the existing local domain ID for testing
 583          *
 584          * @param string $prefix
 585          */
 586         public function setDomainPrefix( $prefix );
 587
 588         /**
 589          * Make certain table names use their own database, schema, and table prefix
 590          * when passed into SQL queries pre-escaped and without a qualified database name
 591          *
 592          * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
 593          * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
 594          *
 595          * Calling this twice will completely clear any old table aliases. Also, note that
 596          * callers are responsible for making sure the schemas and databases actually exist.
 597          *
 598          * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
 599          */
 600         public function setTableAliases( array $aliases );
 601 }