includes/jobqueue/JobRunner.php

   1 <?php
   2 /**
   3  * Job queue runner utility methods
   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 JobQueue
  22  */
  23
  24 /**
  25  * Job queue runner utility methods
  26  *
  27  * @ingroup JobQueue
  28  * @since 1.24
  29  */
  30 class JobRunner {
  31         /** @var callable|null Debug output handler */
  32         protected $debug;
  33
  34         /**
  35          * @param callable $debug Optional debug output handler
  36          */
  37         public function setDebugHandler( $debug ) {
  38                 $this->debug = $debug;
  39         }
  40
  41         /**
  42          * Run jobs of the specified number/type for the specified time
  43          *
  44          * The response map has a 'job' field that lists status of each job, including:
  45          *   - type   : the job type
  46          *   - status : ok/failed
  47          *   - error  : any error message string
  48          *   - time   : the job run time in ms
  49          * The response map also has:
  50          *   - backoffs : the (job type => seconds) map of backoff times
  51          *   - elapsed  : the total time spent running tasks in ms
  52          *   - reached  : the reason the script finished, one of (none-ready, job-limit, time-limit)
  53          *
  54          * This method outputs status information only if a debug handler was set.
  55          * Any exceptions are caught and logged, but are not reported as output.
  56          *
  57          * @param array $options Map of parameters:
  58          *    - type     : the job type (or false for the default types)
  59          *    - maxJobs  : maximum number of jobs to run
  60          *    - maxTime  : maximum time in seconds before stopping
  61          *    - throttle : whether to respect job backoff configuration
  62          * @return array Summary response that can easily be JSON serialized
  63          */
  64         public function run( array $options ) {
  65                 $response = array( 'jobs' => array(), 'reached' => 'none-ready' );
  66
  67                 $type = isset( $options['type'] ) ? $options['type'] : false;
  68                 $maxJobs = isset( $options['maxJobs'] ) ? $options['maxJobs'] : false;
  69                 $maxTime = isset( $options['maxTime'] ) ? $options['maxTime'] : false;
  70                 $noThrottle = isset( $options['throttle'] ) && !$options['throttle'];
  71
  72                 $group = JobQueueGroup::singleton();
  73                 // Handle any required periodic queue maintenance
  74                 $count = $group->executeReadyPeriodicTasks();
  75                 if ( $count > 0 ) {
  76                         $this->runJobsLog( "Executed $count periodic queue task(s)." );
  77                 }
  78
  79                 // Flush any pending DB writes for sanity
  80                 wfGetLBFactory()->commitMasterChanges();
  81
  82                 $backoffs = $this->loadBackoffs(); // map of (type => UNIX expiry)
  83                 $startingBackoffs = $backoffs; // avoid unnecessary writes
  84                 $backoffExpireFunc = function ( $t ) {
  85                         return $t > time();
  86                 };
  87
  88                 $jobsRun = 0; // counter
  89                 $timeMsTotal = 0;
  90                 $flags = JobQueueGroup::USE_CACHE;
  91                 $startTime = microtime( true ); // time since jobs started running
  92                 $lastTime = microtime( true ); // time since last slave check
  93                 do {
  94                         $backoffs = array_filter( $backoffs, $backoffExpireFunc );
  95                         $blacklist = $noThrottle ? array() : array_keys( $backoffs );
  96                         if ( $type === false ) {
  97                                 $job = $group->pop( JobQueueGroup::TYPE_DEFAULT, $flags, $blacklist );
  98                         } elseif ( in_array( $type, $blacklist ) ) {
  99                                 $job = false; // requested queue in backoff state
 100                         } else {
 101                                 $job = $group->pop( $type ); // job from a single queue
 102                         }
 103                         if ( $job ) { // found a job
 104                                 $jType = $job->getType();
 105
 106                                 $this->runJobsLog( $job->toString() . " STARTING" );
 107
 108                                 // Run the job...
 109                                 wfProfileIn( __METHOD__ . '-' . get_class( $job ) );
 110                                 $t = microtime( true );
 111                                 try {
 112                                         ++$jobsRun;
 113                                         $status = $job->run();
 114                                         $error = $job->getLastError();
 115                                         wfGetLBFactory()->commitMasterChanges();
 116                                 } catch ( MWException $e ) {
 117                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 118                                         $status = false;
 119                                         $error = get_class( $e ) . ': ' . $e->getMessage();
 120                                         MWExceptionHandler::logException( $e );
 121                                 }
 122                                 $timeMs = intval( ( microtime( true ) - $t ) * 1000 );
 123                                 wfProfileOut( __METHOD__ . '-' . get_class( $job ) );
 124                                 $timeMsTotal += $timeMs;
 125
 126                                 // Mark the job as done on success or when the job cannot be retried
 127                                 if ( $status !== false || !$job->allowRetries() ) {
 128                                         $group->ack( $job ); // done
 129                                 }
 130
 131                                 if ( $status === false ) {
 132                                         $this->runJobsLog( $job->toString() . " t=$timeMs error={$error}" );
 133                                 } else {
 134                                         $this->runJobsLog( $job->toString() . " t=$timeMs good" );
 135                                 }
 136
 137                                 $response['jobs'][] = array(
 138                                         'type'   => $jType,
 139                                         'status' => ( $status === false ) ? 'failed' : 'ok',
 140                                         'error'  => $error,
 141                                         'time'   => $timeMs
 142                                 );
 143
 144                                 // Back off of certain jobs for a while (for throttling and for errors)
 145                                 $ttw = $this->getBackoffTimeToWait( $job );
 146                                 if ( $status === false && mt_rand( 0, 49 ) == 0 ) {
 147                                         $ttw = max( $ttw, 30 );
 148                                 }
 149                                 if ( $ttw > 0 ) {
 150                                         $backoffs[$jType] = isset( $backoffs[$jType] ) ? $backoffs[$jType] : 0;
 151                                         $backoffs[$jType] = max( $backoffs[$jType], time() + $ttw );
 152                                 }
 153
 154                                 // Break out if we hit the job count or wall time limits...
 155                                 if ( $maxJobs && $jobsRun >= $maxJobs ) {
 156                                         $response['reached'] = 'job-limit';
 157                                         break;
 158                                 } elseif ( $maxTime && ( microtime( true ) - $startTime ) > $maxTime ) {
 159                                         $response['reached'] = 'time-limit';
 160                                         break;
 161                                 }
 162
 163                                 // Don't let any of the main DB slaves get backed up
 164                                 $timePassed = microtime( true ) - $lastTime;
 165                                 if ( $timePassed >= 5 || $timePassed < 0 ) {
 166                                         wfWaitForSlaves( $lastTime );
 167                                         $lastTime = microtime( true );
 168                                 }
 169                                 // Don't let any queue slaves/backups fall behind
 170                                 if ( $jobsRun > 0 && ( $jobsRun % 100 ) == 0 ) {
 171                                         $group->waitForBackups();
 172                                 }
 173
 174                                 // Bail if near-OOM instead of in a job
 175                                 $this->assertMemoryOK();
 176                         }
 177                 } while ( $job ); // stop when there are no jobs
 178
 179                 // Sync the persistent backoffs for the next runJobs.php pass
 180                 $backoffs = array_filter( $backoffs, $backoffExpireFunc );
 181                 if ( $backoffs !== $startingBackoffs ) {
 182                         $this->syncBackoffs( $backoffs );
 183                 }
 184
 185                 $response['backoffs'] = $backoffs;
 186                 $response['elapsed'] = $timeMsTotal;
 187
 188                 return $response;
 189         }
 190
 191         /**
 192          * @param Job $job
 193          * @return int Seconds for this runner to avoid doing more jobs of this type
 194          * @see $wgJobBackoffThrottling
 195          */
 196         private function getBackoffTimeToWait( Job $job ) {
 197                 global $wgJobBackoffThrottling;
 198
 199                 if ( !isset( $wgJobBackoffThrottling[$job->getType()] ) ||
 200                         $job instanceof DuplicateJob // no work was done
 201                 ) {
 202                         return 0; // not throttled
 203                 }
 204
 205                 $itemsPerSecond = $wgJobBackoffThrottling[$job->getType()];
 206                 if ( $itemsPerSecond <= 0 ) {
 207                         return 0; // not throttled
 208                 }
 209
 210                 $seconds = 0;
 211                 if ( $job->workItemCount() > 0 ) {
 212                         $exactSeconds = $job->workItemCount() / $itemsPerSecond;
 213                         // use randomized rounding
 214                         $seconds = floor( $exactSeconds );
 215                         $remainder = $exactSeconds - $seconds;
 216                         $seconds += ( mt_rand() / mt_getrandmax() < $remainder ) ? 1 : 0;
 217                 }
 218
 219                 return (int)$seconds;
 220         }
 221
 222         /**
 223          * Get the previous backoff expiries from persistent storage
 224          *
 225          * @return array Map of (job type => backoff expiry timestamp)
 226          */
 227         private function loadBackoffs() {
 228                 $section = new ProfileSection( __METHOD__ );
 229
 230                 $backoffs = array();
 231                 $file = wfTempDir() . '/mw-runJobs-backoffs.json';
 232                 if ( is_file( $file ) ) {
 233                         $handle = fopen( $file, 'rb' );
 234                         flock( $handle, LOCK_SH );
 235                         $content = stream_get_contents( $handle );
 236                         flock( $handle, LOCK_UN );
 237                         fclose( $handle );
 238                         $backoffs = json_decode( $content, true ) ?: array();
 239                 }
 240
 241                 return $backoffs;
 242         }
 243
 244         /**
 245          * Merge the current backoff expiries from persistent storage
 246          *
 247          * @param array $backoffs Map of (job type => backoff expiry timestamp)
 248          */
 249         private function syncBackoffs( array $backoffs ) {
 250                 $section = new ProfileSection( __METHOD__ );
 251
 252                 $file = wfTempDir() . '/mw-runJobs-backoffs.json';
 253                 $handle = fopen( $file, 'wb+' );
 254                 flock( $handle, LOCK_EX );
 255                 $content = stream_get_contents( $handle );
 256                 $cBackoffs = json_decode( $content, true ) ?: array();
 257                 foreach ( $backoffs as $type => $timestamp ) {
 258                         $cBackoffs[$type] = isset( $cBackoffs[$type] ) ? $cBackoffs[$type] : 0;
 259                         $cBackoffs[$type] = max( $cBackoffs[$type], $backoffs[$type] );
 260                 }
 261                 ftruncate( $handle, 0 );
 262                 fwrite( $handle, json_encode( $backoffs ) );
 263                 flock( $handle, LOCK_UN );
 264                 fclose( $handle );
 265         }
 266
 267         /**
 268          * Make sure that this script is not too close to the memory usage limit.
 269          * It is better to die in between jobs than OOM right in the middle of one.
 270          * @throws MWException
 271          */
 272         private function assertMemoryOK() {
 273                 static $maxBytes = null;
 274                 if ( $maxBytes === null ) {
 275                         $m = array();
 276                         if ( preg_match( '!^(\d+)(k|m|g|)$!i', ini_get( 'memory_limit' ), $m ) ) {
 277                                 list( , $num, $unit ) = $m;
 278                                 $conv = array( 'g' => 1073741824, 'm' => 1048576, 'k' => 1024, '' => 1 );
 279                                 $maxBytes = $num * $conv[strtolower( $unit )];
 280                         } else {
 281                                 $maxBytes = 0;
 282                         }
 283                 }
 284                 $usedBytes = memory_get_usage();
 285                 if ( $maxBytes && $usedBytes >= 0.95 * $maxBytes ) {
 286                         throw new MWException( "Detected excessive memory usage ($usedBytes/$maxBytes)." );
 287                 }
 288         }
 289
 290         /**
 291          * Log the job message
 292          * @param string $msg The message to log
 293          */
 294         private function runJobsLog( $msg ) {
 295                 if ( $this->debug ) {
 296                         call_user_func_array( $this->debug, array( wfTimestamp( TS_DB ) . " $msg\n" ) );
 297                 }
 298                 wfDebugLog( 'runJobs', $msg );
 299         }
 300 }