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         /** @var string Domain specifier when no specific database needs to be selected */
  82         const DOMAIN_ANY = '';
  83
  84         /**
  85          * Construct a manager of IDatabase connection objects
  86          *
  87          * @param array $params Parameter map with keys:
  88          *  - servers : Required. Array of server info structures.
  89          *  - localDomain: A DatabaseDomain or domain ID string.
  90          *  - loadMonitor : Name of a class used to fetch server lag and load.
  91          *  - readOnlyReason : Reason the master DB is read-only if so [optional]
  92          *  - waitTimeout : Maximum time to wait for replicas for consistency [optional]
  93          *  - srvCache : BagOStuff object for server cache [optional]
  94          *  - memCache : BagOStuff object for cluster memory cache [optional]
  95          *  - wanCache : WANObjectCache object [optional]
  96          *  - hostname : The name of the current server [optional]
  97          *  - cliMode: Whether the execution context is a CLI script. [optional]
  98          *  - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
  99          *  - trxProfiler: TransactionProfiler instance. [optional]
 100          *  - replLogger: PSR-3 logger instance. [optional]
 101          *  - connLogger: PSR-3 logger instance. [optional]
 102          *  - queryLogger: PSR-3 logger instance. [optional]
 103          *  - perfLogger: PSR-3 logger instance. [optional]
 104          *  - errorLogger : Callback that takes an Exception and logs it. [optional]
 105          * @throws InvalidArgumentException
 106          */
 107         public function __construct( array $params );
 108
 109         /**
 110          * Get the index of the reader connection, which may be a replica DB
 111          *
 112          * This takes into account load ratios and lag times. It should
 113          * always return a consistent index during a given invocation.
 114          *
 115          * Side effect: opens connections to databases
 116          * @param string|bool $group Query group, or false for the generic reader
 117          * @param string|bool $domain Domain ID, or false for the current domain
 118          * @throws DBError
 119          * @return bool|int|string
 120          */
 121         public function getReaderIndex( $group = false, $domain = false );
 122
 123         /**
 124          * Set the master wait position
 125          *
 126          * If a DB_REPLICA connection has been opened already, then wait immediately.
 127          * Otherwise sets a variable telling it to wait if such a connection is opened.
 128          *
 129          * @param DBMasterPos $pos
 130          */
 131         public function waitFor( $pos );
 132
 133         /**
 134          * Set the master wait position and wait for a "generic" replica DB to catch up to it
 135          *
 136          * This can be used a faster proxy for waitForAll()
 137          *
 138          * @param DBMasterPos $pos
 139          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 140          * @return bool Success (able to connect and no timeouts reached)
 141          */
 142         public function waitForOne( $pos, $timeout = null );
 143
 144         /**
 145          * Set the master wait position and wait for ALL replica DBs to catch up to it
 146          *
 147          * @param DBMasterPos $pos
 148          * @param int $timeout Max seconds to wait; default is mWaitTimeout
 149          * @return bool Success (able to connect and no timeouts reached)
 150          */
 151         public function waitForAll( $pos, $timeout = null );
 152
 153         /**
 154          * Get any open connection to a given server index, local or foreign
 155          *
 156          * @param int $i Server index or DB_MASTER/DB_REPLICA
 157          * @return Database|bool False if no such connection is open
 158          */
 159         public function getAnyOpenConnection( $i );
 160
 161         /**
 162          * Get a connection by index
 163          *
 164          * @param int $i Server index or DB_MASTER/DB_REPLICA
 165          * @param array|string|bool $groups Query group(s), or false for the generic reader
 166          * @param string|bool $domain Domain ID, or false for the current domain
 167          *
 168          * @throws DBError
 169          * @return Database
 170          */
 171         public function getConnection( $i, $groups = [], $domain = false );
 172
 173         /**
 174          * Mark a foreign connection as being available for reuse under a different DB domain
 175          *
 176          * This mechanism is reference-counted, and must be called the same number of times
 177          * as getConnection() to work.
 178          *
 179          * @param IDatabase $conn
 180          * @throws InvalidArgumentException
 181          */
 182         public function reuseConnection( $conn );
 183
 184         /**
 185          * Get a database connection handle reference
 186          *
 187          * The handle's methods simply wrap those of a Database handle
 188          *
 189          * @see ILoadBalancer::getConnection() for parameter information
 190          *
 191          * @param int $i Server index or DB_MASTER/DB_REPLICA
 192          * @param array|string|bool $groups Query group(s), or false for the generic reader
 193          * @param string|bool $domain Domain ID, or false for the current domain
 194          * @return DBConnRef
 195          */
 196         public function getConnectionRef( $i, $groups = [], $domain = false );
 197
 198         /**
 199          * Get a database connection handle reference without connecting yet
 200          *
 201          * The handle's methods simply wrap those of a Database handle
 202          *
 203          * @see ILoadBalancer::getConnection() for parameter information
 204          *
 205          * @param int $i Server index or DB_MASTER/DB_REPLICA
 206          * @param array|string|bool $groups Query group(s), or false for the generic reader
 207          * @param string|bool $domain Domain ID, or false for the current domain
 208          * @return DBConnRef
 209          */
 210         public function getLazyConnectionRef( $i, $groups = [], $domain = false );
 211
 212         /**
 213          * Get a maintenance database connection handle reference for migrations and schema changes
 214          *
 215          * The handle's methods simply wrap those of a Database handle
 216          *
 217          * @see ILoadBalancer::getConnection() for parameter information
 218          *
 219          * @param int $db Server index or DB_MASTER/DB_REPLICA
 220          * @param array|string|bool $groups Query group(s), or false for the generic reader
 221          * @param string|bool $domain Domain ID, or false for the current domain
 222          * @return MaintainableDBConnRef
 223          */
 224         public function getMaintenanceConnectionRef( $db, $groups = [], $domain = false );
 225
 226         /**
 227          * Open a connection to the server given by the specified index
 228          * Index must be an actual index into the array.
 229          * If the server is already open, returns it.
 230          *
 231          * On error, returns false, and the connection which caused the
 232          * error will be available via $this->mErrorConnection.
 233          *
 234          * @note If disable() was called on this LoadBalancer, this method will throw a DBAccessError.
 235          *
 236          * @param int $i Server index or DB_MASTER/DB_REPLICA
 237          * @param string|bool $domain Domain ID, or false for the current domain
 238          * @return Database|bool Returns false on errors
 239          * @throws DBAccessError
 240          */
 241         public function openConnection( $i, $domain = false );
 242
 243         /**
 244          * @return int
 245          */
 246         public function getWriterIndex();
 247
 248         /**
 249          * Returns true if the specified index is a valid server index
 250          *
 251          * @param string $i
 252          * @return bool
 253          */
 254         public function haveIndex( $i );
 255
 256         /**
 257          * Returns true if the specified index is valid and has non-zero load
 258          *
 259          * @param string $i
 260          * @return bool
 261          */
 262         public function isNonZeroLoad( $i );
 263
 264         /**
 265          * Get the number of defined servers (not the number of open connections)
 266          *
 267          * @return int
 268          */
 269         public function getServerCount();
 270
 271         /**
 272          * Get the host name or IP address of the server with the specified index
 273          * Prefer a readable name if available.
 274          * @param string $i
 275          * @return string
 276          */
 277         public function getServerName( $i );
 278
 279         /**
 280          * Return the server info structure for a given index, or false if the index is invalid.
 281          * @param int $i
 282          * @return array|bool
 283          */
 284         public function getServerInfo( $i );
 285
 286         /**
 287          * Sets the server info structure for the given index. Entry at index $i
 288          * is created if it doesn't exist
 289          * @param int $i
 290          * @param array $serverInfo
 291          */
 292         public function setServerInfo( $i, array $serverInfo );
 293
 294         /**
 295          * Get the current master position for chronology control purposes
 296          * @return DBMasterPos|bool Returns false if not applicable
 297          */
 298         public function getMasterPos();
 299
 300         /**
 301          * Disable this load balancer. All connections are closed, and any attempt to
 302          * open a new connection will result in a DBAccessError.
 303          */
 304         public function disable();
 305
 306         /**
 307          * Close all open connections
 308          */
 309         public function closeAll();
 310
 311         /**
 312          * Close a connection
 313          *
 314          * Using this function makes sure the LoadBalancer knows the connection is closed.
 315          * If you use $conn->close() directly, the load balancer won't update its state.
 316          *
 317          * @param IDatabase $conn
 318          */
 319         public function closeConnection( IDatabase $conn );
 320
 321         /**
 322          * Commit transactions on all open connections
 323          * @param string $fname Caller name
 324          * @throws DBExpectedError
 325          */
 326         public function commitAll( $fname = __METHOD__ );
 327
 328         /**
 329          * Perform all pre-commit callbacks that remain part of the atomic transactions
 330          * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
 331          *
 332          * Use this only for mutli-database commits
 333          */
 334         public function finalizeMasterChanges();
 335
 336         /**
 337          * Perform all pre-commit checks for things like replication safety
 338          *
 339          * Use this only for mutli-database commits
 340          *
 341          * @param array $options Includes:
 342          *   - maxWriteDuration : max write query duration time in seconds
 343          * @throws DBTransactionError
 344          */
 345         public function approveMasterChanges( array $options );
 346
 347         /**
 348          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 349          *
 350          * The DBO_TRX setting will be reverted to the default in each of these methods:
 351          *   - commitMasterChanges()
 352          *   - rollbackMasterChanges()
 353          *   - commitAll()
 354          * This allows for custom transaction rounds from any outer transaction scope.
 355          *
 356          * @param string $fname
 357          * @throws DBExpectedError
 358          */
 359         public function beginMasterChanges( $fname = __METHOD__ );
 360
 361         /**
 362          * Issue COMMIT on all master connections where writes where done
 363          * @param string $fname Caller name
 364          * @throws DBExpectedError
 365          */
 366         public function commitMasterChanges( $fname = __METHOD__ );
 367
 368         /**
 369          * Issue all pending post-COMMIT/ROLLBACK callbacks
 370          *
 371          * Use this only for mutli-database commits
 372          *
 373          * @param integer $type IDatabase::TRIGGER_* constant
 374          * @return Exception|null The first exception or null if there were none
 375          */
 376         public function runMasterPostTrxCallbacks( $type );
 377
 378         /**
 379          * Issue ROLLBACK only on master, only if queries were done on connection
 380          * @param string $fname Caller name
 381          * @throws DBExpectedError
 382          */
 383         public function rollbackMasterChanges( $fname = __METHOD__ );
 384
 385         /**
 386          * Suppress all pending post-COMMIT/ROLLBACK callbacks
 387          *
 388          * Use this only for mutli-database commits
 389          *
 390          * @return Exception|null The first exception or null if there were none
 391          */
 392         public function suppressTransactionEndCallbacks();
 393
 394         /**
 395          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 396          *
 397          * @param string $fname Caller name
 398          */
 399         public function flushReplicaSnapshots( $fname = __METHOD__ );
 400
 401         /**
 402          * @return bool Whether a master connection is already open
 403          */
 404         public function hasMasterConnection();
 405
 406         /**
 407          * Determine if there are pending changes in a transaction by this thread
 408          * @return bool
 409          */
 410         public function hasMasterChanges();
 411
 412         /**
 413          * Get the timestamp of the latest write query done by this thread
 414          * @return float|bool UNIX timestamp or false
 415          */
 416         public function lastMasterChangeTimestamp();
 417
 418         /**
 419          * Check if this load balancer object had any recent or still
 420          * pending writes issued against it by this PHP thread
 421          *
 422          * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
 423          * @return bool
 424          */
 425         public function hasOrMadeRecentMasterChanges( $age = null );
 426
 427         /**
 428          * Get the list of callers that have pending master changes
 429          *
 430          * @return string[] List of method names
 431          */
 432         public function pendingMasterChangeCallers();
 433
 434         /**
 435          * @note This method will trigger a DB connection if not yet done
 436          * @param string|bool $domain Domain ID, or false for the current domain
 437          * @return bool Whether the generic connection for reads is highly "lagged"
 438          */
 439         public function getLaggedReplicaMode( $domain = false );
 440
 441         /**
 442          * @note This method will never cause a new DB connection
 443          * @return bool Whether any generic connection used for reads was highly "lagged"
 444          */
 445         public function laggedReplicaUsed();
 446
 447         /**
 448          * @note This method may trigger a DB connection if not yet done
 449          * @param string|bool $domain Domain ID, or false for the current domain
 450          * @param IDatabase|null DB master connection; used to avoid loops [optional]
 451          * @return string|bool Reason the master is read-only or false if it is not
 452          */
 453         public function getReadOnlyReason( $domain = false, IDatabase $conn = null );
 454
 455         /**
 456          * Disables/enables lag checks
 457          * @param null|bool $mode
 458          * @return bool
 459          */
 460         public function allowLagged( $mode = null );
 461
 462         /**
 463          * @return bool
 464          */
 465         public function pingAll();
 466
 467         /**
 468          * Call a function with each open connection object
 469          * @param callable $callback
 470          * @param array $params
 471          */
 472         public function forEachOpenConnection( $callback, array $params = [] );
 473
 474         /**
 475          * Call a function with each open connection object to a master
 476          * @param callable $callback
 477          * @param array $params
 478          */
 479         public function forEachOpenMasterConnection( $callback, array $params = [] );
 480
 481         /**
 482          * Call a function with each open replica DB connection object
 483          * @param callable $callback
 484          * @param array $params
 485          */
 486         public function forEachOpenReplicaConnection( $callback, array $params = [] );
 487
 488         /**
 489          * Get the hostname and lag time of the most-lagged replica DB
 490          *
 491          * This is useful for maintenance scripts that need to throttle their updates.
 492          * May attempt to open connections to replica DBs on the default DB. If there is
 493          * no lag, the maximum lag will be reported as -1.
 494          *
 495          * @param bool|string $domain Domain ID, or false for the default database
 496          * @return array ( host, max lag, index of max lagged host )
 497          */
 498         public function getMaxLag( $domain = false );
 499
 500         /**
 501          * Get an estimate of replication lag (in seconds) for each server
 502          *
 503          * Results are cached for a short time in memcached/process cache
 504          *
 505          * Values may be "false" if replication is too broken to estimate
 506          *
 507          * @param string|bool $domain
 508          * @return int[] Map of (server index => float|int|bool)
 509          */
 510         public function getLagTimes( $domain = false );
 511
 512         /**
 513          * Get the lag in seconds for a given connection, or zero if this load
 514          * balancer does not have replication enabled.
 515          *
 516          * This should be used in preference to Database::getLag() in cases where
 517          * replication may not be in use, since there is no way to determine if
 518          * replication is in use at the connection level without running
 519          * potentially restricted queries such as SHOW SLAVE STATUS. Using this
 520          * function instead of Database::getLag() avoids a fatal error in this
 521          * case on many installations.
 522          *
 523          * @param IDatabase $conn
 524          * @return int|bool Returns false on error
 525          */
 526         public function safeGetLag( IDatabase $conn );
 527
 528         /**
 529          * Wait for a replica DB to reach a specified master position
 530          *
 531          * This will connect to the master to get an accurate position if $pos is not given
 532          *
 533          * @param IDatabase $conn Replica DB
 534          * @param DBMasterPos|bool $pos Master position; default: current position
 535          * @param integer|null $timeout Timeout in seconds [optional]
 536          * @return bool Success
 537          */
 538         public function safeWaitForMasterPos( IDatabase $conn, $pos = false, $timeout = null );
 539
 540         /**
 541          * Set a callback via IDatabase::setTransactionListener() on
 542          * all current and future master connections of this load balancer
 543          *
 544          * @param string $name Callback name
 545          * @param callable|null $callback
 546          */
 547         public function setTransactionListener( $name, callable $callback = null );
 548
 549         /**
 550          * Set a new table prefix for the existing local domain ID for testing
 551          *
 552          * @param string $prefix
 553          */
 554         public function setDomainPrefix( $prefix );
 555
 556         /**
 557          * Make certain table names use their own database, schema, and table prefix
 558          * when passed into SQL queries pre-escaped and without a qualified database name
 559          *
 560          * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
 561          * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
 562          *
 563          * Calling this twice will completely clear any old table aliases. Also, note that
 564          * callers are responsible for making sure the schemas and databases actually exist.
 565          *
 566          * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
 567          */
 568         public function setTableAliases( array $aliases );
 569 }