/** @var callable|null Debug output handler */
protected $debug;
+ /**
+ * @var LoggerInterface $logger
+ */
+ protected $logger;
+
/**
* @param callable $debug Optional debug output handler
*/
$this->debug = $debug;
}
- /**
- * @var LoggerInterface $logger
- */
- protected $logger;
-
/**
* @param LoggerInterface $logger
+ * @return void
*/
public function setLogger( LoggerInterface $logger ) {
$this->logger = $logger;
* @return array Summary response that can easily be JSON serialized
*/
public function run( array $options ) {
- global $wgJobClasses;
+ global $wgJobClasses, $wgTrxProfilerLimits;
$response = array( 'jobs' => array(), 'reached' => 'none-ready' );
return $response;
}
- $group = JobQueueGroup::singleton();
- // Handle any required periodic queue maintenance
- $count = $group->executeReadyPeriodicTasks();
- if ( $count > 0 ) {
- $msg = "Executed $count periodic queue task(s).";
- $this->logger->debug( $msg );
- $this->debugCallback( $msg );
- }
-
// Bail out if in read-only mode
if ( wfReadOnly() ) {
$response['reached'] = 'read-only';
return $response;
}
- // Bail out if there is too much DB lag
- list( , $maxLag ) = wfGetLBFactory()->getMainLB( wfWikiID() )->getMaxLag();
- if ( $maxLag >= 5 ) {
+ $profiler = Profiler::instance();
+
+ // Catch huge single updates that lead to slave lag
+ $trxProfiler = $profiler->getTransactionProfiler();
+ $trxProfiler->setLogger( LoggerFactory::getInstance( 'DBPerformance' ) );
+ $trxProfiler->setExpectations( $wgTrxProfilerLimits['JobRunner'], __METHOD__ );
+
+ // Bail out if there is too much DB lag.
+ // This check should not block as we want to try other wiki queues.
+ $maxAllowedLag = 3;
+ list( , $maxLag ) = wfGetLB( wfWikiID() )->getMaxLag();
+ if ( $maxLag >= $maxAllowedLag ) {
$response['reached'] = 'slave-lag-limit';
return $response;
}
+ $group = JobQueueGroup::singleton();
+
// Flush any pending DB writes for sanity
- wfGetLBFactory()->commitMasterChanges();
+ wfGetLBFactory()->commitAll();
// Some jobs types should not run until a certain timestamp
$backoffs = array(); // map of (type => UNIX expiry)
$backoffDeltas = array(); // map of (type => seconds)
$wait = 'wait'; // block to read backoffs the first time
- $jobsRun = 0;
+ $stats = RequestContext::getMain()->getStats();
+ $jobsPopped = 0;
$timeMsTotal = 0;
$flags = JobQueueGroup::USE_CACHE;
- $checkPeriod = 5.0; // seconds
- $checkPhase = mt_rand( 0, 1000 * $checkPeriod ) / 1000; // avoid stampedes
$startTime = microtime( true ); // time since jobs started running
- $lastTime = microtime( true ) - $checkPhase; // time since last slave check
+ $checkLagPeriod = 1.0; // check slave lag this many seconds
+ $lastCheckTime = 1; // timestamp of last slave check
do {
// Sync the persistent backoffs with concurrent runners
$backoffs = $this->syncBackoffDeltas( $backoffs, $backoffDeltas, $wait );
}
if ( $job ) { // found a job
+ $popTime = time();
$jType = $job->getType();
// Back off of certain jobs for a while (for throttling and for errors)
}
$msg = $job->toString() . " STARTING";
- $this->logger->info( $msg );
+ $this->logger->debug( $msg );
$this->debugCallback( $msg );
// Run the job...
+ $psection = $profiler->scopedProfileIn( __METHOD__ . '-' . $jType );
$jobStartTime = microtime( true );
try {
- ++$jobsRun;
+ ++$jobsPopped;
$status = $job->run();
$error = $job->getLastError();
- wfGetLBFactory()->commitMasterChanges();
+ $this->commitMasterChanges( $job );
+
+ DeferredUpdates::doUpdates();
+ $this->commitMasterChanges( $job );
} catch ( Exception $e ) {
MWExceptionHandler::rollbackMasterChangesAndLog( $e );
$status = false;
$error = get_class( $e ) . ': ' . $e->getMessage();
MWExceptionHandler::logException( $e );
}
+ // Commit all outstanding connections that are in a transaction
+ // to get a fresh repeatable read snapshot on every connection.
+ wfGetLBFactory()->commitAll();
$timeMs = intval( ( microtime( true ) - $jobStartTime ) * 1000 );
$timeMsTotal += $timeMs;
+ $profiler->scopedProfileOut( $psection );
+
+ if ( $job->getQueuedTimestamp() ) {
+ // Record time to run for the job type
+ $stats->timing( "job-pickuptime-$jType",
+ $popTime - $job->getQueuedTimestamp() );
+ }
// Mark the job as done on success or when the job cannot be retried
if ( $status !== false || !$job->allowRetries() ) {
);
// Break out if we hit the job count or wall time limits...
- if ( $maxJobs && $jobsRun >= $maxJobs ) {
+ if ( $maxJobs && $jobsPopped >= $maxJobs ) {
$response['reached'] = 'job-limit';
break;
} elseif ( $maxTime && ( microtime( true ) - $startTime ) > $maxTime ) {
// Don't let any of the main DB slaves get backed up.
// This only waits for so long before exiting and letting
// other wikis in the farm (on different masters) get a chance.
- $timePassed = microtime( true ) - $lastTime;
- if ( $timePassed >= 5 || $timePassed < 0 ) {
- if ( !wfWaitForSlaves( $lastTime, false, '*', 5 ) ) {
+ $timePassed = microtime( true ) - $lastCheckTime;
+ if ( $timePassed >= $checkLagPeriod || $timePassed < 0 ) {
+ if ( !wfWaitForSlaves( $lastCheckTime, false, '*', $maxAllowedLag ) ) {
$response['reached'] = 'slave-lag-limit';
break;
}
- $lastTime = microtime( true );
+ $lastCheckTime = microtime( true );
}
// Don't let any queue slaves/backups fall behind
- if ( $jobsRun > 0 && ( $jobsRun % 100 ) == 0 ) {
+ if ( $jobsPopped > 0 && ( $jobsPopped % 100 ) == 0 ) {
$group->waitForBackups();
}
* @return array Map of (job type => backoff expiry timestamp)
*/
private function loadBackoffs( array $backoffs, $mode = 'wait' ) {
-
$file = wfTempDir() . '/mw-runJobs-backoffs.json';
if ( is_file( $file ) ) {
$noblock = ( $mode === 'nowait' ) ? LOCK_NB : 0;
* @return array The new backoffs account for $backoffs and the latest file data
*/
private function syncBackoffDeltas( array $backoffs, array &$deltas, $mode = 'wait' ) {
-
if ( !$deltas ) {
return $this->loadBackoffs( $backoffs, $mode );
}
call_user_func_array( $this->debug, array( wfTimestamp( TS_DB ) . " $msg\n" ) );
}
}
+
+ /**
+ * Issue a commit on all masters who are currently in a transaction and have
+ * made changes to the database. It also supports sometimes waiting for the
+ * local wiki's slaves to catch up. See the documentation for
+ * $wgJobSerialCommitThreshold for more.
+ *
+ * @param Job $job
+ * @throws DBError
+ */
+ private function commitMasterChanges( Job $job ) {
+ global $wgJobSerialCommitThreshold;
+
+ $lb = wfGetLB( wfWikiID() );
+ if ( $wgJobSerialCommitThreshold !== false ) {
+ // Generally, there is one master connection to the local DB
+ $dbwSerial = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
+ } else {
+ $dbwSerial = false;
+ }
+
+ if ( !$dbwSerial
+ || !$dbwSerial->namedLocksEnqueue()
+ || $dbwSerial->pendingWriteQueryDuration() < $wgJobSerialCommitThreshold
+ ) {
+ // Writes are all to foreign DBs, named locks don't form queues,
+ // or $wgJobSerialCommitThreshold is not reached; commit changes now
+ wfGetLBFactory()->commitMasterChanges();
+ return;
+ }
+
+ $ms = intval( 1000 * $dbwSerial->pendingWriteQueryDuration() );
+ $msg = $job->toString() . " COMMIT ENQUEUED [{$ms}ms of writes]";
+ $this->logger->info( $msg );
+ $this->debugCallback( $msg );
+
+ // Wait for an exclusive lock to commit
+ if ( !$dbwSerial->lock( 'jobrunner-serial-commit', __METHOD__, 30 ) ) {
+ // This will trigger a rollback in the main loop
+ throw new DBError( $dbwSerial, "Timed out waiting on commit queue." );
+ }
+ // Wait for the generic slave to catch up
+ $pos = $lb->getMasterPos();
+ if ( $pos ) {
+ $lb->waitForOne( $pos );
+ }
+
+ $fname = __METHOD__;
+ // Re-ping all masters with transactions. This throws DBError if some
+ // connection died while waiting on locks/slaves, triggering a rollback.
+ wfGetLBFactory()->forEachLB( function( LoadBalancer $lb ) use ( $fname ) {
+ $lb->forEachOpenConnection( function( DatabaseBase $conn ) use ( $fname ) {
+ if ( $conn->writesOrCallbacksPending() ) {
+ $conn->query( "SELECT 1", $fname );
+ }
+ } );
+ } );
+
+ // Actually commit the DB master changes
+ wfGetLBFactory()->commitMasterChanges();
+
+ // Release the lock
+ $dbwSerial->unlock( 'jobrunner-serial-commit', __METHOD__ );
+ }
}
dépôts / lhc / web / wiklou.git / blobdiff