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  * @author Aaron Schulz
  23  */
  24
  25 /**
  26  * Database cluster connection, tracking, load balancing, and transaction manager interface
  27  *
  28  * A "cluster" is considered to be one master database and zero or more replica databases.
  29  * Typically, the replica DBs replicate from the master asynchronously. The first node in the
  30  * "servers" configuration array is always considered the "master". However, this class can still
  31  * be used when all or some of the "replica" DBs are multi-master peers of the master or even
  32  * when all the DBs are non-replicating clones of each other holding read-only data. Thus, the
  33  * role of "master" is in some cases merely nominal.
  34  *
  35  * By default, each DB server uses DBO_DEFAULT for its 'flags' setting, unless explicitly set
  36  * otherwise in configuration. DBO_DEFAULT behavior depends on whether 'cliMode' is set:
  37  *   - In CLI mode, the flag has no effect with regards to LoadBalancer.
  38  *   - In non-CLI mode, the flag causes implicit transactions to be used; the first query on
  39  *     a database starts a transaction on that database. The transactions are meant to remain
  40  *     pending until either commitMasterChanges() or rollbackMasterChanges() is called. The
  41  *     application must have some point where it calls commitMasterChanges() near the end of
  42  *     the PHP request.
  43  * Every iteration of beginMasterChanges()/commitMasterChanges() is called a "transaction round".
  44  * Rounds are useful on the master DB connections because they make single-DB (and by and large
  45  * multi-DB) updates in web requests all-or-nothing. Also, transactions on replica DBs are useful
  46  * when REPEATABLE-READ or SERIALIZABLE isolation is used because all foriegn keys and constraints
  47  * hold across separate queries in the DB transaction since the data appears within a consistent
  48  * point-in-time snapshot.
  49  *
  50  * The typical caller will use LoadBalancer::getConnection( DB_* ) to yield a live database
  51  * connection handle. The choice of which DB server to use is based on pre-defined loads for
  52  * weighted random selection, adjustments thereof by LoadMonitor, and the amount of replication
  53  * lag on each DB server. Lag checks might cause problems in certain setups, so they should be
  54  * tuned in the server configuration maps as follows:
  55  *   - Master + N Replica(s): set 'max lag' to an appropriate threshold for avoiding any database
  56  *      lagged by this much or more. If all DBs are this lagged, then the load balancer considers
  57  *      the cluster to be read-only.
  58  *   - Galera Cluster: Seconds_Behind_Master will be 0, so there probably is nothing to tune.
  59  *      Note that lag is still possible depending on how wsrep-sync-wait is set server-side.
  60  *   - Read-only archive clones: set 'is static' in the server configuration maps. This will
  61  *      treat all such DBs as having 0 lag.
  62  *   - SQL load balancing proxy: any proxy should handle lag checks on its own, so the 'max lag'
  63  *      parameter should probably be set to INF in the server configuration maps. This will make
  64  *      the load balancer ignore whatever it detects as the lag of the logical replica is (which
  65  *      would probably just randomly bounce around).
  66  *
  67  * If using a SQL proxy service, it would probably be best to have two proxy hosts for the
  68  * load balancer to talk to. One would be the 'host' of the master server entry and another for
  69  * the (logical) replica server entry. The proxy could map the load balancer's "replica" DB to
  70  * any number of physical replica DBs.
  71  *
  72  * @since 1.28
  73  * @ingroup Database
  74  */
  75 interface ILoadBalancer {
  76         /** @var integer Request a replica DB connection */
  77         const DB_REPLICA = -1;
  78         /** @var integer Request a master DB connection */
  79         const DB_MASTER = -2;
  80
  81         /**
  82          * Construct a manager of IDatabase connection objects
  83          *
  84          * @param array $params Parameter map with keys:
  85          *  - servers : Required. Array of server info structures.
  86          *  - localDomain: A DatabaseDomain or domain ID string.
  87          *  - loadMonitor : Name of a class used to fetch server lag and load.
  88          *  - readOnlyReason : Reason the master DB is read-only if so [optional]
  89          *  - waitTimeout : Maximum time to wait for replicas for consistency [optional]
  90          *  - srvCache : BagOStuff object for server cache [optional]
  91          *  - memCache : BagOStuff object for cluster memory cache [optional]
  92          *  - wanCache : WANObjectCache object [optional]
  93          *  - hostname : The name of the current server [optional]
  94          *  - cliMode: Whether the execution context is a CLI script. [optional]
  95          *  - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
  96          *  - trxProfiler: TransactionProfiler instance. [optional]
  97          *  - replLogger: PSR-3 logger instance. [optional]
  98          *  - connLogger: PSR-3 logger instance. [optional]
  99          *  - queryLogger: PSR-3 logger instance. [optional]
 100          *  - perfLogger: PSR-3 logger instance. [optional]
 101          *  - errorLogger : Callback that takes an Exception and logs it. [optional]
 102          * @throws InvalidArgumentException
 103          */
 104         public function __construct( array $params );
 105
 106         /**
 107          * Get the index of the reader connection, which may be a replica DB
 108          * This takes into account load ratios and lag times. It should
 109          * always return a consistent index during a given invocation
 110          *
 111          * Side effect: opens connections to databases
 112          * @param string|bool $group Query group, or false for the generic reader
 113          * @param string|bool $domain Domain ID, or false for the current domain
 114          * @throws DBError
 115          * @return bool|int|string
 116          */
 117         public function getReaderIndex( $group = false, $domain = false );
 118
 119         /**
 120          * Set the master wait position
 121          * If a DB_REPLICA connection has been opened already, waits
 122          * Otherwise sets a variable telling it to wait if such a connection is opened
 123          * @param DBMasterPos $pos
 124          */
 125         public function waitFor( $pos );
 126
 127         /**
 128          * Set the master wait position and wait for a "generic" replica DB to catch up to it
 129          *
 130          * This can be used a faster proxy for waitForAll()
 131          *
 132          * @param DBMasterPos $pos
 133          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 134          * @return bool Success (able to connect and no timeouts reached)
 135          */
 136         public function waitForOne( $pos, $timeout = null );
 137
 138         /**
 139          * Set the master wait position and wait for ALL replica DBs to catch up to it
 140          * @param DBMasterPos $pos
 141          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 142          * @return bool Success (able to connect and no timeouts reached)
 143          */
 144         public function waitForAll( $pos, $timeout = null );
 145
 146         /**
 147          * Get any open connection to a given server index, local or foreign
 148          * Returns false if there is no connection open
 149          *
 150          * @param int $i Server index
 151          * @return IDatabase|bool False on failure
 152          */
 153         public function getAnyOpenConnection( $i );
 154
 155         /**
 156          * Get a connection by index
 157          * This is the main entry point for this class.
 158          *
 159          * @param int $i Server index
 160          * @param array|string|bool $groups Query group(s), or false for the generic reader
 161          * @param string|bool $domain Domain ID, or false for the current domain
 162          *
 163          * @throws DBError
 164          * @return IDatabase
 165          */
 166         public function getConnection( $i, $groups = [], $domain = false );
 167
 168         /**
 169          * Mark a foreign connection as being available for reuse under a different
 170          * DB name or prefix. This mechanism is reference-counted, and must be called
 171          * the same number of times as getConnection() to work.
 172          *
 173          * @param IDatabase $conn
 174          * @throws InvalidArgumentException
 175          */
 176         public function reuseConnection( $conn );
 177
 178         /**
 179          * Get a database connection handle reference
 180          *
 181          * The handle's methods wrap simply wrap those of a IDatabase handle
 182          *
 183          * @see LoadBalancer::getConnection() for parameter information
 184          *
 185          * @param int $db
 186          * @param array|string|bool $groups Query group(s), or false for the generic reader
 187          * @param string|bool $domain Domain ID, or false for the current domain
 188          * @return DBConnRef
 189          */
 190         public function getConnectionRef( $db, $groups = [], $domain = false );
 191
 192         /**
 193          * Get a database connection handle reference without connecting yet
 194          *
 195          * The handle's methods wrap simply wrap those of a IDatabase handle
 196          *
 197          * @see LoadBalancer::getConnection() for parameter information
 198          *
 199          * @param int $db
 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 getLazyConnectionRef( $db, $groups = [], $domain = false );
 205
 206         /**
 207          * Open a connection to the server given by the specified index
 208          * Index must be an actual index into the array.
 209          * If the server is already open, returns it.
 210          *
 211          * On error, returns false, and the connection which caused the
 212          * error will be available via $this->mErrorConnection.
 213          *
 214          * @note If disable() was called on this LoadBalancer, this method will throw a DBAccessError.
 215          *
 216          * @param int $i Server index
 217          * @param string|bool $domain Domain ID, or false for the current domain
 218          * @return IDatabase|bool Returns false on errors
 219          */
 220         public function openConnection( $i, $domain = false );
 221
 222         /**
 223          * @return int
 224          */
 225         public function getWriterIndex();
 226
 227         /**
 228          * Returns true if the specified index is a valid server index
 229          *
 230          * @param string $i
 231          * @return bool
 232          */
 233         public function haveIndex( $i );
 234
 235         /**
 236          * Returns true if the specified index is valid and has non-zero load
 237          *
 238          * @param string $i
 239          * @return bool
 240          */
 241         public function isNonZeroLoad( $i );
 242
 243         /**
 244          * Get the number of defined servers (not the number of open connections)
 245          *
 246          * @return int
 247          */
 248         public function getServerCount();
 249
 250         /**
 251          * Get the host name or IP address of the server with the specified index
 252          * Prefer a readable name if available.
 253          * @param string $i
 254          * @return string
 255          */
 256         public function getServerName( $i );
 257
 258         /**
 259          * Return the server info structure for a given index, or false if the index is invalid.
 260          * @param int $i
 261          * @return array|bool
 262          */
 263         public function getServerInfo( $i );
 264
 265         /**
 266          * Sets the server info structure for the given index. Entry at index $i
 267          * is created if it doesn't exist
 268          * @param int $i
 269          * @param array $serverInfo
 270          */
 271         public function setServerInfo( $i, array $serverInfo );
 272
 273         /**
 274          * Get the current master position for chronology control purposes
 275          * @return DBMasterPos|bool Returns false if not applicable
 276          */
 277         public function getMasterPos();
 278
 279         /**
 280          * Disable this load balancer. All connections are closed, and any attempt to
 281          * open a new connection will result in a DBAccessError.
 282          */
 283         public function disable();
 284
 285         /**
 286          * Close all open connections
 287          */
 288         public function closeAll();
 289
 290         /**
 291          * Close a connection
 292          *
 293          * Using this function makes sure the LoadBalancer knows the connection is closed.
 294          * If you use $conn->close() directly, the load balancer won't update its state.
 295          *
 296          * @param IDatabase $conn
 297          */
 298         public function closeConnection( IDatabase $conn );
 299
 300         /**
 301          * Commit transactions on all open connections
 302          * @param string $fname Caller name
 303          * @throws DBExpectedError
 304          */
 305         public function commitAll( $fname = __METHOD__ );
 306
 307         /**
 308          * Perform all pre-commit callbacks that remain part of the atomic transactions
 309          * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
 310          *
 311          * Use this only for mutli-database commits
 312          */
 313         public function finalizeMasterChanges();
 314
 315         /**
 316          * Perform all pre-commit checks for things like replication safety
 317          *
 318          * Use this only for mutli-database commits
 319          *
 320          * @param array $options Includes:
 321          *   - maxWriteDuration : max write query duration time in seconds
 322          * @throws DBTransactionError
 323          */
 324         public function approveMasterChanges( array $options );
 325
 326         /**
 327          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 328          *
 329          * The DBO_TRX setting will be reverted to the default in each of these methods:
 330          *   - commitMasterChanges()
 331          *   - rollbackMasterChanges()
 332          *   - commitAll()
 333          * This allows for custom transaction rounds from any outer transaction scope.
 334          *
 335          * @param string $fname
 336          * @throws DBExpectedError
 337          */
 338         public function beginMasterChanges( $fname = __METHOD__ );
 339
 340         /**
 341          * Issue COMMIT on all master connections where writes where done
 342          * @param string $fname Caller name
 343          * @throws DBExpectedError
 344          */
 345         public function commitMasterChanges( $fname = __METHOD__ );
 346
 347         /**
 348          * Issue all pending post-COMMIT/ROLLBACK callbacks
 349          *
 350          * Use this only for mutli-database commits
 351          *
 352          * @param integer $type IDatabase::TRIGGER_* constant
 353          * @return Exception|null The first exception or null if there were none
 354          */
 355         public function runMasterPostTrxCallbacks( $type );
 356
 357         /**
 358          * Issue ROLLBACK only on master, only if queries were done on connection
 359          * @param string $fname Caller name
 360          * @throws DBExpectedError
 361          */
 362         public function rollbackMasterChanges( $fname = __METHOD__ );
 363
 364         /**
 365          * Suppress all pending post-COMMIT/ROLLBACK callbacks
 366          *
 367          * Use this only for mutli-database commits
 368          *
 369          * @return Exception|null The first exception or null if there were none
 370          */
 371         public function suppressTransactionEndCallbacks();
 372
 373         /**
 374          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 375          *
 376          * @param string $fname Caller name
 377          */
 378         public function flushReplicaSnapshots( $fname = __METHOD__ );
 379
 380         /**
 381          * @return bool Whether a master connection is already open
 382          */
 383         public function hasMasterConnection();
 384
 385         /**
 386          * Determine if there are pending changes in a transaction by this thread
 387          * @return bool
 388          */
 389         public function hasMasterChanges();
 390
 391         /**
 392          * Get the timestamp of the latest write query done by this thread
 393          * @return float|bool UNIX timestamp or false
 394          */
 395         public function lastMasterChangeTimestamp();
 396
 397         /**
 398          * Check if this load balancer object had any recent or still
 399          * pending writes issued against it by this PHP thread
 400          *
 401          * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
 402          * @return bool
 403          */
 404         public function hasOrMadeRecentMasterChanges( $age = null );
 405
 406         /**
 407          * Get the list of callers that have pending master changes
 408          *
 409          * @return string[] List of method names
 410          */
 411         public function pendingMasterChangeCallers();
 412
 413         /**
 414          * @note This method will trigger a DB connection if not yet done
 415          * @param string|bool $domain Domain ID, or false for the current domain
 416          * @return bool Whether the generic connection for reads is highly "lagged"
 417          */
 418         public function getLaggedReplicaMode( $domain = false );
 419
 420         /**
 421          * @note This method will never cause a new DB connection
 422          * @return bool Whether any generic connection used for reads was highly "lagged"
 423          */
 424         public function laggedReplicaUsed();
 425
 426         /**
 427          * @note This method may trigger a DB connection if not yet done
 428          * @param string|bool $domain Domain ID, or false for the current domain
 429          * @param IDatabase|null DB master connection; used to avoid loops [optional]
 430          * @return string|bool Reason the master is read-only or false if it is not
 431          */
 432         public function getReadOnlyReason( $domain = false, IDatabase $conn = null );
 433
 434         /**
 435          * Disables/enables lag checks
 436          * @param null|bool $mode
 437          * @return bool
 438          */
 439         public function allowLagged( $mode = null );
 440
 441         /**
 442          * @return bool
 443          */
 444         public function pingAll();
 445
 446         /**
 447          * Call a function with each open connection object
 448          * @param callable $callback
 449          * @param array $params
 450          */
 451         public function forEachOpenConnection( $callback, array $params = [] );
 452
 453         /**
 454          * Call a function with each open connection object to a master
 455          * @param callable $callback
 456          * @param array $params
 457          */
 458         public function forEachOpenMasterConnection( $callback, array $params = [] );
 459
 460         /**
 461          * Call a function with each open replica DB connection object
 462          * @param callable $callback
 463          * @param array $params
 464          */
 465         public function forEachOpenReplicaConnection( $callback, array $params = [] );
 466
 467         /**
 468          * Get the hostname and lag time of the most-lagged replica DB
 469          *
 470          * This is useful for maintenance scripts that need to throttle their updates.
 471          * May attempt to open connections to replica DBs on the default DB. If there is
 472          * no lag, the maximum lag will be reported as -1.
 473          *
 474          * @param bool|string $domain Domain ID, or false for the default database
 475          * @return array ( host, max lag, index of max lagged host )
 476          */
 477         public function getMaxLag( $domain = false );
 478
 479         /**
 480          * Get an estimate of replication lag (in seconds) for each server
 481          *
 482          * Results are cached for a short time in memcached/process cache
 483          *
 484          * Values may be "false" if replication is too broken to estimate
 485          *
 486          * @param string|bool $domain
 487          * @return int[] Map of (server index => float|int|bool)
 488          */
 489         public function getLagTimes( $domain = false );
 490
 491         /**
 492          * Get the lag in seconds for a given connection, or zero if this load
 493          * balancer does not have replication enabled.
 494          *
 495          * This should be used in preference to Database::getLag() in cases where
 496          * replication may not be in use, since there is no way to determine if
 497          * replication is in use at the connection level without running
 498          * potentially restricted queries such as SHOW SLAVE STATUS. Using this
 499          * function instead of Database::getLag() avoids a fatal error in this
 500          * case on many installations.
 501          *
 502          * @param IDatabase $conn
 503          * @return int|bool Returns false on error
 504          */
 505         public function safeGetLag( IDatabase $conn );
 506
 507         /**
 508          * Wait for a replica DB to reach a specified master position
 509          *
 510          * This will connect to the master to get an accurate position if $pos is not given
 511          *
 512          * @param IDatabase $conn Replica DB
 513          * @param DBMasterPos|bool $pos Master position; default: current position
 514          * @param integer|null $timeout Timeout in seconds [optional]
 515          * @return bool Success
 516          */
 517         public function safeWaitForMasterPos( IDatabase $conn, $pos = false, $timeout = null );
 518
 519         /**
 520          * Clear the cache for slag lag delay times
 521          *
 522          * This is only used for testing
 523          */
 524         public function clearLagTimeCache();
 525
 526         /**
 527          * Set a callback via IDatabase::setTransactionListener() on
 528          * all current and future master connections of this load balancer
 529          *
 530          * @param string $name Callback name
 531          * @param callable|null $callback
 532          */
 533         public function setTransactionListener( $name, callable $callback = null );
 534
 535         /**
 536          * Set a new table prefix for the existing local domain ID for testing
 537          *
 538          * @param string $prefix
 539          */
 540         public function setDomainPrefix( $prefix );
 541
 542         /**
 543          * Make certain table names use their own database, schema, and table prefix
 544          * when passed into SQL queries pre-escaped and without a qualified database name
 545          *
 546          * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
 547          * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
 548          *
 549          * Calling this twice will completely clear any old table aliases. Also, note that
 550          * callers are responsible for making sure the schemas and databases actually exist.
 551          *
 552          * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
 553          */
 554         public function setTableAliases( array $aliases );
 555 }