<?php
-# Database load balancing object
-
+/**
+ *
+ * @package MediaWiki
+ */
+
+/**
+ * Depends on the database object
+ */
require_once( 'Database.php' );
# Valid database indexes
define( 'DB_READ', -1 );
define( 'DB_WRITE', -2 );
-# Task-based indexes
-# ***NOT USED YET, EXPERIMENTAL***
-# These may be defined in $wgDBservers. If they aren't, the default reader or writer will be used
-# Even numbers are always readers, odd numbers are writers
-define( 'DB_TASK_FIRST', 1000 ); # First in list
-define( 'DB_SEARCH_R', 1000 ); # Search read
-define( 'DB_SEARCH_W', 1001 ); # Search write
-define( 'DB_ASKSQL_R', 1002 ); # Special:Asksql read
-define( 'DB_WATCHLIST_R', 1004 ); # Watchlist read
-define( 'DB_TASK_LAST', 1004) ; # Last in list
-
-define( 'MASTER_WAIT_TIMEOUT', 15 ); # Time to wait for a slave to synchronise
-
+/**
+ * Database load balancing object
+ *
+ * @todo document
+ * @package MediaWiki
+ */
class LoadBalancer {
- /* private */ var $mServers, $mConnections, $mLoads;
+ /* private */ var $mServers, $mConnections, $mLoads, $mGroupLoads;
/* private */ var $mFailFunction;
/* private */ var $mForce, $mReadIndex, $mLastIndex;
- /* private */ var $mWaitForFile, $mWaitForPos;
+ /* private */ var $mWaitForFile, $mWaitForPos, $mWaitTimeout;
+ /* private */ var $mLaggedSlaveMode;
function LoadBalancer()
{
$this->mLastIndex = -1;
}
- function newFromParams( $servers, $failFunction = false )
+ function newFromParams( $servers, $failFunction = false, $waitTimeout = 10 )
{
$lb = new LoadBalancer;
- $lb->initialise( $servers, $failFunction = false );
+ $lb->initialise( $servers, $failFunction, $waitTimeout );
return $lb;
}
- function initialise( $servers, $failFunction = false )
+ function initialise( $servers, $failFunction = false, $waitTimeout = 10 )
{
$this->mServers = $servers;
$this->mFailFunction = $failFunction;
$this->mLoads = array();
$this->mWaitForFile = false;
$this->mWaitForPos = false;
+ $this->mWaitTimeout = $waitTimeout;
+ $this->mLaggedSlaveMode = false;
foreach( $servers as $i => $server ) {
$this->mLoads[$i] = $server['load'];
- }
+ if ( isset( $server['groupLoads'] ) ) {
+ foreach ( $server['groupLoads'] as $group => $ratio ) {
+ if ( !isset( $this->mGroupLoads[$group] ) ) {
+ $this->mGroupLoads[$group] = array();
+ }
+ $this->mGroupLoads[$group][$i] = $ratio;
+ }
+ }
+ }
}
- # Given an array of non-normalised probabilities, this function will select
- # an element and return the appropriate key
+ /**
+ * Given an array of non-normalised probabilities, this function will select
+ * an element and return the appropriate key
+ */
function pickRandom( $weights )
{
if ( !is_array( $weights ) || count( $weights ) == 0 ) {
foreach ( $weights as $w ) {
$sum += $w;
}
+
+ if ( $sum == 0 ) {
+ # No loads on any of them
+ # Just pick one at random
+ foreach ( $weights as $i => $w ) {
+ $weights[$i] = 1;
+ }
+ }
$max = mt_getrandmax();
$rand = mt_rand(0, $max) / $max * $sum;
return $i;
}
+ function getRandomNonLagged( $loads ) {
+ # Unset excessively lagged servers
+ $lags = $this->getLagTimes();
+ foreach ( $lags as $i => $lag ) {
+ if ( isset( $this->mServers[$i]['max lag'] ) && $lag > $this->mServers[$i]['max lag'] ) {
+ unset( $loads[$i] );
+ }
+ }
+
+
+ # Find out if all the slaves with non-zero load are lagged
+ $sum = 0;
+ foreach ( $loads as $load ) {
+ $sum += $load;
+ }
+ if ( $sum == 0 ) {
+ # No appropriate DB servers except maybe the master and some slaves with zero load
+ # Do NOT use the master
+ # Instead, this function will return false, triggering read-only mode,
+ # and a lagged slave will be used instead.
+ unset ( $loads[0] );
+ }
+
+ if ( count( $loads ) == 0 ) {
+ return false;
+ }
+
+ #wfDebug( var_export( $loads, true ) );
+
+ # Return a random representative of the remainder
+ return $this->pickRandom( $loads );
+ }
+
+ /**
+ * Get the index of the reader connection, which may be a slave
+ * This takes into account load ratios and lag times. It should
+ * always return a consistent index during a given invocation
+ *
+ * Side effect: opens connections to databases
+ */
function getReaderIndex()
{
+ global $wgMaxLag, $wgReadOnly, $wgDBClusterTimeout;
+
$fname = 'LoadBalancer::getReaderIndex';
wfProfileIn( $fname );
# $loads is $this->mLoads except with elements knocked out if they
# don't work
$loads = $this->mLoads;
+ $done = false;
+ $totalElapsed = 0;
do {
- $i = $this->pickRandom( $loads );
+ if ( $wgReadOnly ) {
+ $i = $this->pickRandom( $loads );
+ } else {
+ $i = $this->getRandomNonLagged( $loads );
+ if ( $i === false && count( $loads ) != 0 ) {
+ # All slaves lagged. Switch to read-only mode
+ $wgReadOnly = wfMsgNoDB( 'readonly_lag' );
+ $i = $this->pickRandom( $loads );
+ }
+ }
if ( $i !== false ) {
- wfDebug( "Using reader #$i: {$this->mServers[$i]['host']}\n" );
-
+ wfDebug( "Using reader #$i: {$this->mServers[$i]['host']}...\n" );
$this->openConnection( $i );
-
+
if ( !$this->isOpen( $i ) ) {
+ wfDebug( "Failed\n" );
unset( $loads[$i] );
+ $sleepTime = 0;
+ } else {
+ $status = $this->mConnections[$i]->getStatus();
+ if ( isset( $this->mServers[$i]['max threads'] ) &&
+ $status['Threads_running'] > $this->mServers[$i]['max threads'] )
+ {
+ # Slave is lagged, wait for a while
+ $sleepTime = 5000 * $status['Threads_connected'];
+
+ # If we reach the timeout and exit the loop, don't use it
+ $i = false;
+ } else {
+ $done = true;
+ $sleepTime = 0;
+ }
}
+ } else {
+ $sleepTime = 500000;
}
- } while ( $i !== false && !$this->isOpen( $i ) );
+ if ( $sleepTime ) {
+ $totalElapsed += $sleepTime;
+ usleep( $sleepTime );
+ }
+ } while ( count( $loads ) && !$done && $totalElapsed / 1e6 < $wgDBClusterTimeout );
- if ( $this->isOpen( $i ) ) {
- $this->mReadIndex = $i;
+ if ( $i !== false && $this->isOpen( $i ) ) {
+ # Wait for the session master pos for a short time
+ if ( $this->mWaitForFile ) {
+ if ( !$this->doWait( $i ) ) {
+ $this->mServers[$i]['slave pos'] = $this->mConnections[$i]->getSlavePos();
+ }
+ }
+ if ( $i !== false ) {
+ $this->mReadIndex = $i;
+ }
} else {
$i = false;
}
wfProfileOut( $fname );
return $i;
}
-
- # Set the master wait position
- # If a DB_SLAVE connection has been opened already, waits
- # Otherwise sets a variable telling it to wait if such a connection is opened
+
+ /**
+ * Get a random server to use in a query group
+ */
+ function getGroupIndex( $group ) {
+ if ( isset( $this->mGroupLoads[$group] ) ) {
+ $i = $this->pickRandom( $this->mGroupLoads[$group] );
+ } else {
+ $i = false;
+ }
+ wfDebug( "Query group $group => $i\n" );
+ return $i;
+ }
+
+ /**
+ * Set the master wait position
+ * If a DB_SLAVE connection has been opened already, waits
+ * Otherwise sets a variable telling it to wait if such a connection is opened
+ */
function waitFor( $file, $pos ) {
$fname = 'LoadBalancer::waitFor';
wfProfileIn( $fname );
if ( count( $this->mServers ) > 1 ) {
$this->mWaitForFile = $file;
$this->mWaitForPos = $pos;
+ $i = $this->mReadIndex;
- if ( $this->mReadIndex > 0 ) {
- if ( !$this->doWait( $this->mReadIndex ) ) {
- # Use master instead
- $this->mReadIndex = 0;
+ if ( $i > 0 ) {
+ if ( !$this->doWait( $i ) ) {
+ $this->mServers[$i]['slave pos'] = $this->mConnections[$i]->getSlavePos();
+ $this->mLaggedSlaveMode = true;
}
}
}
wfProfileOut( $fname );
}
- # Wait for a given slave to catch up to the master pos stored in $this
+ /**
+ * Wait for a given slave to catch up to the master pos stored in $this
+ */
function doWait( $index ) {
global $wgMemc;
$retVal = false;
+ # Debugging hacks
+ if ( isset( $this->mServers[$index]['lagged slave'] ) ) {
+ return false;
+ } elseif ( isset( $this->mServers[$index]['fake slave'] ) ) {
+ return true;
+ }
+
$key = 'masterpos:' . $index;
$memcPos = $wgMemc->get( $key );
if ( $memcPos ) {
}
if ( !$retVal && $this->isOpen( $index ) ) {
- $conn =& $this->mConnections( $index );
+ $conn =& $this->mConnections[$index];
wfDebug( "Waiting for slave #$index to catch up...\n" );
- $result = $conn->masterPosWait( $this->mWaitForFile, $this->mWaitForPos, MASTER_WAIT_TIMEOUT );
+ $result = $conn->masterPosWait( $this->mWaitForFile, $this->mWaitForPos, $this->mWaitTimeout );
if ( $result == -1 || is_null( $result ) ) {
# Timed out waiting for slave, use master instead
return $retVal;
}
- # Get a connection by index
- function &getConnection( $i, $fail = true )
+ /**
+ * Get a connection by index
+ */
+ function &getConnection( $i, $fail = true, $groups = array() )
{
$fname = 'LoadBalancer::getConnection';
wfProfileIn( $fname );
- /*
- # Task-based index
- if ( $i >= DB_TASK_FIRST && $i < DB_TASK_LAST ) {
- if ( $i % 2 ) {
- # Odd index use writer
- $i = DB_MASTER;
- } else {
- # Even index use reader
- $i = DB_SLAVE;
+
+ # Query groups
+ $groupIndex = false;
+ foreach ( $groups as $group ) {
+ $groupIndex = $this->getGroupIndex( $group );
+ if ( $groupIndex !== false ) {
+ $i = $groupIndex;
+ break;
}
- }*/
-
+ }
+
# Operation-based index
if ( $i == DB_SLAVE ) {
$i = $this->getReaderIndex();
}
# Now we have an explicit index into the servers array
$this->openConnection( $i, $fail );
+
wfProfileOut( $fname );
return $this->mConnections[$i];
}
- # Open a connection to the server given by the specified index
- # Index must be an actual index into the array
- /* private */ function openConnection( $i, $fail = false ) {
+ /**
+ * Open a connection to the server given by the specified index
+ * Index must be an actual index into the array
+ * Returns success
+ * @private
+ */
+ function openConnection( $i, $fail = false ) {
$fname = 'LoadBalancer::openConnection';
wfProfileIn( $fname );
+ $success = true;
if ( !$this->isOpen( $i ) ) {
$this->mConnections[$i] = $this->reallyOpenConnection( $this->mServers[$i] );
-
- if ( $i != 0 && $this->mWaitForFile ) {
- if ( !$this->doWait( $i ) ) {
- # Error waiting for this slave, use master instead
- $this->mReadIndex = 0;
- $i = 0;
- if ( !$this->isOpen( 0 ) ) {
- $this->mConnections[0] = $this->reallyOpenConnection( $this->mServers[0] );
- }
- wfDebug( "Failed over to {$this->mConnections[0]->mServer}\n" );
- }
- }
}
if ( !$this->isOpen( $i ) ) {
wfDebug( "Failed to connect to database $i at {$this->mServers[$i]['host']}\n" );
$this->reportConnectionError( $this->mConnections[$i] );
}
$this->mConnections[$i] = false;
+ $success = false;
}
$this->mLastIndex = $i;
wfProfileOut( $fname );
+ return $success;
}
- # Test if the specified index represents an open connection
- /* private */ function isOpen( $index ) {
+ /**
+ * Test if the specified index represents an open connection
+ * @private
+ */
+ function isOpen( $index ) {
if( !is_integer( $index ) ) {
return false;
}
}
}
- # Really opens a connection
- /* private */ function reallyOpenConnection( &$server ) {
- extract( $server );
- # Get class for this database type
- $class = 'Database' . ucfirst( $type );
- if ( !class_exists( $class ) ) {
- require_once( "$class.php" );
- }
+ /**
+ * Really opens a connection
+ * @private
+ */
+ function reallyOpenConnection( &$server ) {
+ if( !is_array( $server ) ) {
+ wfDebugDieBacktrace( 'You must update your load-balancing configuration. See DefaultSettings.php entry for $wgDBservers.' );
+ }
+
+ extract( $server );
+ # Get class for this database type
+ $class = 'Database' . ucfirst( $type );
+ if ( !class_exists( $class ) ) {
+ require_once( "$class.php" );
+ }
- # Create object
- return new $class( $host, $user, $password, $dbname, 1, $flags );
+ # Create object
+ return new $class( $host, $user, $password, $dbname, 1, $flags );
}
function reportConnectionError( &$conn )
return array_key_exists( $i, $this->mServers );
}
- # Get the number of defined servers (not the number of open connections)
+ /**
+ * Get the number of defined servers (not the number of open connections)
+ */
function getServerCount() {
return count( $this->mServers );
}
- # Save master pos to the session and to memcached, if the session exists
+ /**
+ * Save master pos to the session and to memcached, if the session exists
+ */
function saveMasterPos() {
global $wgSessionStarted;
if ( $wgSessionStarted && count( $this->mServers ) > 1 ) {
}
}
- # Loads the master pos from the session, waits for it if necessary
+ /**
+ * Loads the master pos from the session, waits for it if necessary
+ */
function loadMasterPos() {
if ( isset( $_SESSION['master_log_file'] ) && isset( $_SESSION['master_pos'] ) ) {
$this->waitFor( $_SESSION['master_log_file'], $_SESSION['master_pos'] );
}
}
- # Close all open connections
+ /**
+ * Close all open connections
+ */
function closeAll() {
foreach( $this->mConnections as $i => $conn ) {
if ( $this->isOpen( $i ) ) {
- $conn->close();
+ // Need to use this syntax because $conn is a copy not a reference
+ $this->mConnections[$i]->close();
}
}
}
function commitAll() {
foreach( $this->mConnections as $i => $conn ) {
if ( $this->isOpen( $i ) ) {
- $conn->immediateCommit();
+ // Need to use this syntax because $conn is a copy not a reference
+ $this->mConnections[$i]->immediateCommit();
+ }
+ }
+ }
+
+ function waitTimeout( $value = NULL ) {
+ return wfSetVar( $this->mWaitTimeout, $value );
+ }
+
+ function getLaggedSlaveMode() {
+ return $this->mLaggedSlaveMode;
+ }
+
+ function pingAll() {
+ $success = true;
+ foreach ( $this->mConnections as $i => $conn ) {
+ if ( $this->isOpen( $i ) ) {
+ if ( !$this->mConnections[$i]->ping() ) {
+ $success = false;
+ }
+ }
+ }
+ return $success;
+ }
+
+ /**
+ * Get the hostname and lag time of the most-lagged slave
+ * This is useful for maintenance scripts that need to throttle their updates
+ */
+ function getMaxLag() {
+ $maxLag = -1;
+ $host = '';
+ foreach ( $this->mServers as $i => $conn ) {
+ if ( $this->openConnection( $i ) ) {
+ $lag = $this->mConnections[$i]->getLag();
+ if ( $lag > $maxLag ) {
+ $maxLag = $lag;
+ $host = $this->mServers[$i]['host'];
+ }
+ }
+ }
+ return array( $host, $maxLag );
+ }
+
+ /**
+ * Get lag time for each DB
+ * Results are cached for a short time in memcached
+ */
+ function getLagTimes() {
+ $expiry = 5;
+ $requestRate = 10;
+
+ global $wgMemc;
+ $times = $wgMemc->get( 'lag_times' );
+ if ( $times ) {
+ # Randomly recache with probability rising over $expiry
+ $elapsed = time() - $times['timestamp'];
+ $chance = max( 0, ( $expiry - $elapsed ) * $requestRate );
+ if ( mt_rand( 0, $chance ) != 0 ) {
+ unset( $times['timestamp'] );
+ return $times;
}
}
+
+ # Cache key missing or expired
+
+ $times = array();
+ foreach ( $this->mServers as $i => $conn ) {
+ if ( $this->openConnection( $i ) ) {
+ $times[$i] = $this->mConnections[$i]->getLag();
+ }
+ }
+
+ # Add a timestamp key so we know when it was cached
+ $times['timestamp'] = time();
+ $wgMemc->set( 'lag_times', $times, $expiry );
+
+ # But don't give the timestamp to the caller
+ unset($times['timestamp']);
+ return $times;
}
}
+
+?>