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 integer Request a replica DB connection */
  80         const DB_REPLICA = -1;
  81         /** @var integer 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         /**
  88          * Construct a manager of IDatabase connection objects
  89          *
  90          * @param array $params Parameter map with keys:
  91          *  - servers : Required. Array of server info structures.
  92          *  - localDomain: A DatabaseDomain or domain ID string.
  93          *  - loadMonitor : Name of a class used to fetch server lag and load.
  94          *  - readOnlyReason : Reason the master DB is read-only if so [optional]
  95          *  - waitTimeout : Maximum time to wait for replicas for consistency [optional]
  96          *  - srvCache : BagOStuff object for server cache [optional]
  97          *  - memCache : BagOStuff object for cluster memory cache [optional]
  98          *  - wanCache : WANObjectCache object [optional]
  99          *  - chronologyProtector: ChronologyProtector object [optional]
 100          *  - hostname : The name of the current server [optional]
 101          *  - cliMode: Whether the execution context is a CLI script. [optional]
 102          *  - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
 103          *  - trxProfiler: TransactionProfiler instance. [optional]
 104          *  - replLogger: PSR-3 logger instance. [optional]
 105          *  - connLogger: PSR-3 logger instance. [optional]
 106          *  - queryLogger: PSR-3 logger instance. [optional]
 107          *  - perfLogger: PSR-3 logger instance. [optional]
 108          *  - errorLogger : Callback that takes an Exception and logs it. [optional]
 109          * @throws InvalidArgumentException
 110          */
 111         public function __construct( array $params );
 112
 113         /**
 114          * Get the index of the reader connection, which may be a replica DB
 115          *
 116          * This takes into account load ratios and lag times. It should
 117          * always return a consistent index during a given invocation.
 118          *
 119          * Side effect: opens connections to databases
 120          * @param string|bool $group Query group, or false for the generic reader
 121          * @param string|bool $domain Domain ID, or false for the current domain
 122          * @throws DBError
 123          * @return bool|int|string
 124          */
 125         public function getReaderIndex( $group = false, $domain = false );
 126
 127         /**
 128          * Set the master wait position
 129          *
 130          * If a DB_REPLICA connection has been opened already, then wait immediately.
 131          * Otherwise sets a variable telling it to wait if such a connection is opened.
 132          *
 133          * This only applies to connections to the generic replica DB for this request.
 134          * If a timeout happens when waiting, then getLaggedReplicaMode()/laggedReplicaUsed()
 135          * will return true.
 136          *
 137          * @param DBMasterPos|bool $pos Master position or false
 138          */
 139         public function waitFor( $pos );
 140
 141         /**
 142          * Set the master wait position and wait for a "generic" replica DB to catch up to it
 143          *
 144          * This can be used a faster proxy for waitForAll()
 145          *
 146          * @param DBMasterPos|bool $pos Master position or false
 147          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 148          * @return bool Success (able to connect and no timeouts reached)
 149          */
 150         public function waitForOne( $pos, $timeout = null );
 151
 152         /**
 153          * Set the master wait position and wait for ALL replica DBs to catch up to it
 154          *
 155          * @param DBMasterPos|bool $pos Master position or false
 156          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 157          * @return bool Success (able to connect and no timeouts reached)
 158          */
 159         public function waitForAll( $pos, $timeout = null );
 160
 161         /**
 162          * Get any open connection to a given server index, local or foreign
 163          *
 164          * @param int $i Server index or DB_MASTER/DB_REPLICA
 165          * @return Database|bool False if no such connection is open
 166          */
 167         public function getAnyOpenConnection( $i );
 168
 169         /**
 170          * Get a connection by index
 171          *
 172          * @param int $i Server index or DB_MASTER/DB_REPLICA
 173          * @param array|string|bool $groups Query group(s), or false for the generic reader
 174          * @param string|bool $domain Domain ID, or false for the current domain
 175          *
 176          * @throws DBError
 177          * @return Database
 178          */
 179         public function getConnection( $i, $groups = [], $domain = false );
 180
 181         /**
 182          * Mark a foreign connection as being available for reuse under a different DB domain
 183          *
 184          * This mechanism is reference-counted, and must be called the same number of times
 185          * as getConnection() to work.
 186          *
 187          * @param IDatabase $conn
 188          * @throws InvalidArgumentException
 189          */
 190         public function reuseConnection( $conn );
 191
 192         /**
 193          * Get a database connection handle reference
 194          *
 195          * The handle's methods simply wrap those of a Database handle
 196          *
 197          * @see ILoadBalancer::getConnection() for parameter information
 198          *
 199          * @param int $i Server index or DB_MASTER/DB_REPLICA
 200          * @param array|string|bool $groups Query group(s), or false for the generic reader
 201          * @param string|bool $domain Domain ID, or false for the current domain
 202          * @return DBConnRef
 203          */
 204         public function getConnectionRef( $i, $groups = [], $domain = false );
 205
 206         /**
 207          * Get a database connection handle reference without connecting yet
 208          *
 209          * The handle's methods simply wrap those of a Database handle
 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          * @return DBConnRef
 217          */
 218         public function getLazyConnectionRef( $i, $groups = [], $domain = false );
 219
 220         /**
 221          * Get a maintenance database connection handle reference for migrations and schema changes
 222          *
 223          * The handle's methods simply wrap those of a Database handle
 224          *
 225          * @see ILoadBalancer::getConnection() for parameter information
 226          *
 227          * @param int $db Server index or DB_MASTER/DB_REPLICA
 228          * @param array|string|bool $groups Query group(s), or false for the generic reader
 229          * @param string|bool $domain Domain ID, or false for the current domain
 230          * @return MaintainableDBConnRef
 231          */
 232         public function getMaintenanceConnectionRef( $db, $groups = [], $domain = false );
 233
 234         /**
 235          * Open a connection to the server given by the specified index
 236          * Index must be an actual index into the array.
 237          * If the server is already open, returns it.
 238          *
 239          * @note If disable() was called on this LoadBalancer, this method will throw a DBAccessError.
 240          *
 241          * @param int $i Server index or DB_MASTER/DB_REPLICA
 242          * @param string|bool $domain Domain ID, or false for the current domain
 243          * @return Database|bool Returns false on errors
 244          * @throws DBAccessError
 245          */
 246         public function openConnection( $i, $domain = false );
 247
 248         /**
 249          * @return int
 250          */
 251         public function getWriterIndex();
 252
 253         /**
 254          * Returns true if the specified index is a valid server index
 255          *
 256          * @param string $i
 257          * @return bool
 258          */
 259         public function haveIndex( $i );
 260
 261         /**
 262          * Returns true if the specified index is valid and has non-zero load
 263          *
 264          * @param string $i
 265          * @return bool
 266          */
 267         public function isNonZeroLoad( $i );
 268
 269         /**
 270          * Get the number of defined servers (not the number of open connections)
 271          *
 272          * @return int
 273          */
 274         public function getServerCount();
 275
 276         /**
 277          * Get the host name or IP address of the server with the specified index
 278          * Prefer a readable name if available.
 279          * @param string $i
 280          * @return string
 281          */
 282         public function getServerName( $i );
 283
 284         /**
 285          * Return the server info structure for a given index, or false if the index is invalid.
 286          * @param int $i
 287          * @return array|bool
 288          *
 289          * @deprecated Since 1.30, no alternative
 290          */
 291         public function getServerInfo( $i );
 292
 293         /**
 294          * Sets the server info structure for the given index. Entry at index $i
 295          * is created if it doesn't exist
 296          * @param int $i
 297          * @param array $serverInfo
 298          *
 299          * @deprecated Since 1.30, construct new object
 300          */
 301         public function setServerInfo( $i, array $serverInfo );
 302
 303         /**
 304          * Get the current master position for chronology control purposes
 305          * @return DBMasterPos|bool Returns false if not applicable
 306          */
 307         public function getMasterPos();
 308
 309         /**
 310          * Disable this load balancer. All connections are closed, and any attempt to
 311          * open a new connection will result in a DBAccessError.
 312          */
 313         public function disable();
 314
 315         /**
 316          * Close all open connections
 317          */
 318         public function closeAll();
 319
 320         /**
 321          * Close a connection
 322          *
 323          * Using this function makes sure the LoadBalancer knows the connection is closed.
 324          * If you use $conn->close() directly, the load balancer won't update its state.
 325          *
 326          * @param IDatabase $conn
 327          */
 328         public function closeConnection( IDatabase $conn );
 329
 330         /**
 331          * Commit transactions on all open connections
 332          * @param string $fname Caller name
 333          * @throws DBExpectedError
 334          */
 335         public function commitAll( $fname = __METHOD__ );
 336
 337         /**
 338          * Perform all pre-commit callbacks that remain part of the atomic transactions
 339          * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
 340          *
 341          * Use this only for mutli-database commits
 342          */
 343         public function finalizeMasterChanges();
 344
 345         /**
 346          * Perform all pre-commit checks for things like replication safety
 347          *
 348          * Use this only for mutli-database commits
 349          *
 350          * @param array $options Includes:
 351          *   - maxWriteDuration : max write query duration time in seconds
 352          * @throws DBTransactionError
 353          */
 354         public function approveMasterChanges( array $options );
 355
 356         /**
 357          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 358          *
 359          * The DBO_TRX setting will be reverted to the default in each of these methods:
 360          *   - commitMasterChanges()
 361          *   - rollbackMasterChanges()
 362          *   - commitAll()
 363          * This allows for custom transaction rounds from any outer transaction scope.
 364          *
 365          * @param string $fname
 366          * @throws DBExpectedError
 367          */
 368         public function beginMasterChanges( $fname = __METHOD__ );
 369
 370         /**
 371          * Issue COMMIT on all master connections where writes where done
 372          * @param string $fname Caller name
 373          * @throws DBExpectedError
 374          */
 375         public function commitMasterChanges( $fname = __METHOD__ );
 376
 377         /**
 378          * Issue all pending post-COMMIT/ROLLBACK callbacks
 379          *
 380          * Use this only for mutli-database commits
 381          *
 382          * @param int $type IDatabase::TRIGGER_* constant
 383          * @return Exception|null The first exception or null if there were none
 384          */
 385         public function runMasterPostTrxCallbacks( $type );
 386
 387         /**
 388          * Issue ROLLBACK only on master, only if queries were done on connection
 389          * @param string $fname Caller name
 390          * @throws DBExpectedError
 391          */
 392         public function rollbackMasterChanges( $fname = __METHOD__ );
 393
 394         /**
 395          * Suppress all pending post-COMMIT/ROLLBACK callbacks
 396          *
 397          * Use this only for mutli-database commits
 398          *
 399          * @return Exception|null The first exception or null if there were none
 400          */
 401         public function suppressTransactionEndCallbacks();
 402
 403         /**
 404          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 405          *
 406          * @param string $fname Caller name
 407          */
 408         public function flushReplicaSnapshots( $fname = __METHOD__ );
 409
 410         /**
 411          * @return bool Whether a master connection is already open
 412          */
 413         public function hasMasterConnection();
 414
 415         /**
 416          * Determine if there are pending changes in a transaction by this thread
 417          * @return bool
 418          */
 419         public function hasMasterChanges();
 420
 421         /**
 422          * Get the timestamp of the latest write query done by this thread
 423          * @return float|bool UNIX timestamp or false
 424          */
 425         public function lastMasterChangeTimestamp();
 426
 427         /**
 428          * Check if this load balancer object had any recent or still
 429          * pending writes issued against it by this PHP thread
 430          *
 431          * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
 432          * @return bool
 433          */
 434         public function hasOrMadeRecentMasterChanges( $age = null );
 435
 436         /**
 437          * Get the list of callers that have pending master changes
 438          *
 439          * @return string[] List of method names
 440          */
 441         public function pendingMasterChangeCallers();
 442
 443         /**
 444          * @note This method will trigger a DB connection if not yet done
 445          * @param string|bool $domain Domain ID, or false for the current domain
 446          * @return bool Whether the database for generic connections this request is highly "lagged"
 447          */
 448         public function getLaggedReplicaMode( $domain = false );
 449
 450         /**
 451          * Checks whether the database for generic connections this request was both:
 452          *   - a) Already choosen due to a prior connection attempt
 453          *   - b) Considered highly "lagged"
 454          *
 455          * @note This method will never cause a new DB connection
 456          * @return bool
 457          */
 458         public function laggedReplicaUsed();
 459
 460         /**
 461          * @note This method may trigger a DB connection if not yet done
 462          * @param string|bool $domain Domain ID, or false for the current domain
 463          * @param IDatabase|null $conn DB master connection; used to avoid loops [optional]
 464          * @return string|bool Reason the master is read-only or false if it is not
 465          */
 466         public function getReadOnlyReason( $domain = false, IDatabase $conn = null );
 467
 468         /**
 469          * Disables/enables lag checks
 470          * @param null|bool $mode
 471          * @return bool
 472          */
 473         public function allowLagged( $mode = null );
 474
 475         /**
 476          * @return bool
 477          */
 478         public function pingAll();
 479
 480         /**
 481          * Call a function with each open connection object
 482          * @param callable $callback
 483          * @param array $params
 484          */
 485         public function forEachOpenConnection( $callback, array $params = [] );
 486
 487         /**
 488          * Call a function with each open connection object to a master
 489          * @param callable $callback
 490          * @param array $params
 491          */
 492         public function forEachOpenMasterConnection( $callback, array $params = [] );
 493
 494         /**
 495          * Call a function with each open replica DB connection object
 496          * @param callable $callback
 497          * @param array $params
 498          */
 499         public function forEachOpenReplicaConnection( $callback, array $params = [] );
 500
 501         /**
 502          * Get the hostname and lag time of the most-lagged replica DB
 503          *
 504          * This is useful for maintenance scripts that need to throttle their updates.
 505          * May attempt to open connections to replica DBs on the default DB. If there is
 506          * no lag, the maximum lag will be reported as -1.
 507          *
 508          * @param bool|string $domain Domain ID, or false for the default database
 509          * @return array ( host, max lag, index of max lagged host )
 510          */
 511         public function getMaxLag( $domain = false );
 512
 513         /**
 514          * Get an estimate of replication lag (in seconds) for each server
 515          *
 516          * Results are cached for a short time in memcached/process cache
 517          *
 518          * Values may be "false" if replication is too broken to estimate
 519          *
 520          * @param string|bool $domain
 521          * @return int[] Map of (server index => float|int|bool)
 522          */
 523         public function getLagTimes( $domain = false );
 524
 525         /**
 526          * Get the lag in seconds for a given connection, or zero if this load
 527          * balancer does not have replication enabled.
 528          *
 529          * This should be used in preference to Database::getLag() in cases where
 530          * replication may not be in use, since there is no way to determine if
 531          * replication is in use at the connection level without running
 532          * potentially restricted queries such as SHOW SLAVE STATUS. Using this
 533          * function instead of Database::getLag() avoids a fatal error in this
 534          * case on many installations.
 535          *
 536          * @param IDatabase $conn
 537          * @return int|bool Returns false on error
 538          */
 539         public function safeGetLag( IDatabase $conn );
 540
 541         /**
 542          * Wait for a replica DB to reach a specified master position
 543          *
 544          * This will connect to the master to get an accurate position if $pos is not given
 545          *
 546          * @param IDatabase $conn Replica DB
 547          * @param DBMasterPos|bool $pos Master position; default: current position
 548          * @param int $timeout Timeout in seconds [optional]
 549          * @return bool Success
 550          */
 551         public function safeWaitForMasterPos( IDatabase $conn, $pos = false, $timeout = 10 );
 552
 553         /**
 554          * Set a callback via IDatabase::setTransactionListener() on
 555          * all current and future master connections of this load balancer
 556          *
 557          * @param string $name Callback name
 558          * @param callable|null $callback
 559          */
 560         public function setTransactionListener( $name, callable $callback = null );
 561
 562         /**
 563          * Set a new table prefix for the existing local domain ID for testing
 564          *
 565          * @param string $prefix
 566          */
 567         public function setDomainPrefix( $prefix );
 568
 569         /**
 570          * Make certain table names use their own database, schema, and table prefix
 571          * when passed into SQL queries pre-escaped and without a qualified database name
 572          *
 573          * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
 574          * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
 575          *
 576          * Calling this twice will completely clear any old table aliases. Also, note that
 577          * callers are responsible for making sure the schemas and databases actually exist.
 578          *
 579          * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
 580          */
 581         public function setTableAliases( array $aliases );
 582 }