includes/job/JobQueue.php

   1 <?php
   2 /**
   3  * Job queue base code.
   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  * @defgroup JobQueue JobQueue
  22  */
  23
  24 /**
  25  * Class to both describe a background job and handle jobs.
  26  *
  27  * @ingroup JobQueue
  28  */
  29 abstract class Job {
  30
  31         /**
  32          * @var Title
  33          */
  34         var $title;
  35
  36         var $command,
  37                 $params,
  38                 $id,
  39                 $removeDuplicates,
  40                 $error;
  41
  42         /*-------------------------------------------------------------------------
  43          * Abstract functions
  44          *------------------------------------------------------------------------*/
  45
  46         /**
  47          * Run the job
  48          * @return boolean success
  49          */
  50         abstract function run();
  51
  52         /*-------------------------------------------------------------------------
  53          * Static functions
  54          *------------------------------------------------------------------------*/
  55
  56         /**
  57          * Pop a job of a certain type.  This tries less hard than pop() to
  58          * actually find a job; it may be adversely affected by concurrent job
  59          * runners.
  60          *
  61          * @param $type string
  62          *
  63          * @return Job
  64          */
  65         static function pop_type( $type ) {
  66                 wfProfilein( __METHOD__ );
  67
  68                 $dbw = wfGetDB( DB_MASTER );
  69
  70                 $dbw->begin( __METHOD__ );
  71
  72                 $row = $dbw->selectRow(
  73                         'job',
  74                         '*',
  75                         array( 'job_cmd' => $type ),
  76                         __METHOD__,
  77                         array( 'LIMIT' => 1, 'FOR UPDATE' )
  78                 );
  79
  80                 if ( $row === false ) {
  81                         $dbw->commit( __METHOD__ );
  82                         wfProfileOut( __METHOD__ );
  83                         return false;
  84                 }
  85
  86                 /* Ensure we "own" this row */
  87                 $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
  88                 $affected = $dbw->affectedRows();
  89                 $dbw->commit( __METHOD__ );
  90
  91                 if ( $affected == 0 ) {
  92                         wfProfileOut( __METHOD__ );
  93                         return false;
  94                 }
  95
  96                 wfIncrStats( 'job-pop' );
  97                 $namespace = $row->job_namespace;
  98                 $dbkey = $row->job_title;
  99                 $title = Title::makeTitleSafe( $namespace, $dbkey );
 100                 $job = Job::factory( $row->job_cmd, $title, Job::extractBlob( $row->job_params ),
 101                         $row->job_id );
 102
 103                 $job->removeDuplicates();
 104
 105                 wfProfileOut( __METHOD__ );
 106                 return $job;
 107         }
 108
 109         /**
 110          * Pop a job off the front of the queue
 111          *
 112          * @param $offset Integer: Number of jobs to skip
 113          * @return Job or false if there's no jobs
 114          */
 115         static function pop( $offset = 0 ) {
 116                 global $wgJobTypesExcludedFromDefaultQueue;
 117                 wfProfileIn( __METHOD__ );
 118
 119                 $dbr = wfGetDB( DB_SLAVE );
 120
 121                 /* Get a job from the slave, start with an offset,
 122                         scan full set afterwards, avoid hitting purged rows
 123
 124                         NB: If random fetch previously was used, offset
 125                                 will always be ahead of few entries
 126                 */
 127                 $conditions = array();
 128                 if ( count( $wgJobTypesExcludedFromDefaultQueue ) != 0 ) {
 129                         foreach ( $wgJobTypesExcludedFromDefaultQueue as $cmdType ) {
 130                                 $conditions[] = "job_cmd != " . $dbr->addQuotes( $cmdType );
 131                         }
 132                 }
 133                 $offset = intval( $offset );
 134                 $options = array( 'ORDER BY' => 'job_id', 'USE INDEX' => 'PRIMARY' );
 135
 136                 $row = $dbr->selectRow( 'job', '*',
 137                         array_merge( $conditions, array( "job_id >= $offset" ) ),
 138                         __METHOD__,
 139                         $options
 140                 );
 141
 142                 // Refetching without offset is needed as some of job IDs could have had delayed commits
 143                 // and have lower IDs than jobs already executed, blame concurrency :)
 144                 //
 145                 if ( $row === false ) {
 146                         if ( $offset != 0 ) {
 147                                 $row = $dbr->selectRow( 'job', '*', $conditions, __METHOD__, $options );
 148                         }
 149
 150                         if ( $row === false ) {
 151                                 wfProfileOut( __METHOD__ );
 152                                 return false;
 153                         }
 154                 }
 155
 156                 // Try to delete it from the master
 157                 $dbw = wfGetDB( DB_MASTER );
 158                 $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 159                 $affected = $dbw->affectedRows();
 160                 $dbw->commit( __METHOD__ );
 161
 162                 if ( !$affected ) {
 163                         // Failed, someone else beat us to it
 164                         // Try getting a random row
 165                         $row = $dbw->selectRow( 'job', array( 'MIN(job_id) as minjob',
 166                                 'MAX(job_id) as maxjob' ), '1=1', __METHOD__ );
 167                         if ( $row === false || is_null( $row->minjob ) || is_null( $row->maxjob ) ) {
 168                                 // No jobs to get
 169                                 wfProfileOut( __METHOD__ );
 170                                 return false;
 171                         }
 172                         // Get the random row
 173                         $row = $dbw->selectRow( 'job', '*',
 174                                 'job_id >= ' . mt_rand( $row->minjob, $row->maxjob ), __METHOD__ );
 175                         if ( $row === false ) {
 176                                 // Random job gone before we got the chance to select it
 177                                 // Give up
 178                                 wfProfileOut( __METHOD__ );
 179                                 return false;
 180                         }
 181                         // Delete the random row
 182                         $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 183                         $affected = $dbw->affectedRows();
 184                         $dbw->commit( __METHOD__ );
 185
 186                         if ( !$affected ) {
 187                                 // Random job gone before we exclusively deleted it
 188                                 // Give up
 189                                 wfProfileOut( __METHOD__ );
 190                                 return false;
 191                         }
 192                 }
 193
 194                 // If execution got to here, there's a row in $row that has been deleted from the database
 195                 // by this thread. Hence the concurrent pop was successful.
 196                 wfIncrStats( 'job-pop' );
 197                 $namespace = $row->job_namespace;
 198                 $dbkey = $row->job_title;
 199                 $title = Title::makeTitleSafe( $namespace, $dbkey );
 200
 201                 if ( is_null( $title ) ) {
 202                         return false;
 203                 }
 204
 205                 $job = Job::factory( $row->job_cmd, $title, Job::extractBlob( $row->job_params ), $row->job_id );
 206
 207                 // Remove any duplicates it may have later in the queue
 208                 $job->removeDuplicates();
 209
 210                 wfProfileOut( __METHOD__ );
 211                 return $job;
 212         }
 213
 214         /**
 215          * Create the appropriate object to handle a specific job
 216          *
 217          * @param $command String: Job command
 218          * @param $title Title: Associated title
 219          * @param $params Array: Job parameters
 220          * @param $id Int: Job identifier
 221          * @return Job
 222          */
 223         static function factory( $command, Title $title, $params = false, $id = 0 ) {
 224                 global $wgJobClasses;
 225                 if( isset( $wgJobClasses[$command] ) ) {
 226                         $class = $wgJobClasses[$command];
 227                         return new $class( $title, $params, $id );
 228                 }
 229                 throw new MWException( "Invalid job command `{$command}`" );
 230         }
 231
 232         /**
 233          * @param $params
 234          * @return string
 235          */
 236         static function makeBlob( $params ) {
 237                 if ( $params !== false ) {
 238                         return serialize( $params );
 239                 } else {
 240                         return '';
 241                 }
 242         }
 243
 244         /**
 245          * @param $blob
 246          * @return bool|mixed
 247          */
 248         static function extractBlob( $blob ) {
 249                 if ( (string)$blob !== '' ) {
 250                         return unserialize( $blob );
 251                 } else {
 252                         return false;
 253                 }
 254         }
 255
 256         /**
 257          * Batch-insert a group of jobs into the queue.
 258          * This will be wrapped in a transaction with a forced commit.
 259          *
 260          * This may add duplicate at insert time, but they will be
 261          * removed later on, when the first one is popped.
 262          *
 263          * @param $jobs array of Job objects
 264          */
 265         static function batchInsert( $jobs ) {
 266                 if ( !count( $jobs ) ) {
 267                         return;
 268                 }
 269                 $dbw = wfGetDB( DB_MASTER );
 270                 $rows = array();
 271
 272                 /**
 273                  * @var $job Job
 274                  */
 275                 foreach ( $jobs as $job ) {
 276                         $rows[] = $job->insertFields();
 277                         if ( count( $rows ) >= 50 ) {
 278                                 # Do a small transaction to avoid slave lag
 279                                 $dbw->begin( __METHOD__ );
 280                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 281                                 $dbw->commit( __METHOD__ );
 282                                 $rows = array();
 283                         }
 284                 }
 285                 if ( $rows ) { // last chunk
 286                         $dbw->begin( __METHOD__ );
 287                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 288                         $dbw->commit( __METHOD__ );
 289                 }
 290                 wfIncrStats( 'job-insert', count( $jobs ) );
 291         }
 292
 293         /**
 294          * Insert a group of jobs into the queue.
 295          *
 296          * Same as batchInsert() but does not commit and can thus
 297          * be rolled-back as part of a larger transaction. However,
 298          * large batches of jobs can cause slave lag.
 299          *
 300          * @param $jobs array of Job objects
 301          */
 302         static function safeBatchInsert( $jobs ) {
 303                 if ( !count( $jobs ) ) {
 304                         return;
 305                 }
 306                 $dbw = wfGetDB( DB_MASTER );
 307                 $rows = array();
 308                 foreach ( $jobs as $job ) {
 309                         $rows[] = $job->insertFields();
 310                         if ( count( $rows ) >= 500 ) {
 311                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 312                                 $rows = array();
 313                         }
 314                 }
 315                 if ( $rows ) { // last chunk
 316                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 317                 }
 318                 wfIncrStats( 'job-insert', count( $jobs ) );
 319         }
 320
 321         /*-------------------------------------------------------------------------
 322          * Non-static functions
 323          *------------------------------------------------------------------------*/
 324
 325         /**
 326          * @param $command
 327          * @param $title
 328          * @param $params array
 329          * @param int $id
 330          */
 331         function __construct( $command, $title, $params = false, $id = 0 ) {
 332                 $this->command = $command;
 333                 $this->title = $title;
 334                 $this->params = $params;
 335                 $this->id = $id;
 336
 337                 // A bit of premature generalisation
 338                 // Oh well, the whole class is premature generalisation really
 339                 $this->removeDuplicates = true;
 340         }
 341
 342         /**
 343          * Insert a single job into the queue.
 344          * @return bool true on success
 345          */
 346         function insert() {
 347                 $fields = $this->insertFields();
 348
 349                 $dbw = wfGetDB( DB_MASTER );
 350
 351                 if ( $this->removeDuplicates ) {
 352                         $res = $dbw->select( 'job', array( '1' ), $fields, __METHOD__ );
 353                         if ( $dbw->numRows( $res ) ) {
 354                                 return true;
 355                         }
 356                 }
 357                 wfIncrStats( 'job-insert' );
 358                 return $dbw->insert( 'job', $fields, __METHOD__ );
 359         }
 360
 361         /**
 362          * @return array
 363          */
 364         protected function insertFields() {
 365                 $dbw = wfGetDB( DB_MASTER );
 366                 return array(
 367                         'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
 368                         'job_cmd' => $this->command,
 369                         'job_namespace' => $this->title->getNamespace(),
 370                         'job_title' => $this->title->getDBkey(),
 371                         'job_timestamp' => $dbw->timestamp(),
 372                         'job_params' => Job::makeBlob( $this->params )
 373                 );
 374         }
 375
 376         /**
 377          * Remove jobs in the job queue which are duplicates of this job.
 378          * This is deadlock-prone and so starts its own transaction.
 379          */
 380         function removeDuplicates() {
 381                 if ( !$this->removeDuplicates ) {
 382                         return;
 383                 }
 384
 385                 $fields = $this->insertFields();
 386                 unset( $fields['job_id'] );
 387                 $dbw = wfGetDB( DB_MASTER );
 388                 $dbw->begin( __METHOD__ );
 389                 $dbw->delete( 'job', $fields, __METHOD__ );
 390                 $affected = $dbw->affectedRows();
 391                 $dbw->commit( __METHOD__ );
 392                 if ( $affected ) {
 393                         wfIncrStats( 'job-dup-delete', $affected );
 394                 }
 395         }
 396
 397         /**
 398          * @return string
 399          */
 400         function toString() {
 401                 $paramString = '';
 402                 if ( $this->params ) {
 403                         foreach ( $this->params as $key => $value ) {
 404                                 if ( $paramString != '' ) {
 405                                         $paramString .= ' ';
 406                                 }
 407                                 $paramString .= "$key=$value";
 408                         }
 409                 }
 410
 411                 if ( is_object( $this->title ) ) {
 412                         $s = "{$this->command} " . $this->title->getPrefixedDBkey();
 413                         if ( $paramString !== '' ) {
 414                                 $s .= ' ' . $paramString;
 415                         }
 416                         return $s;
 417                 } else {
 418                         return "{$this->command} $paramString";
 419                 }
 420         }
 421
 422         protected function setLastError( $error ) {
 423                 $this->error = $error;
 424         }
 425
 426         function getLastError() {
 427                 return $this->error;
 428         }
 429 }