includes/job/Job.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                 wfProfileIn( __METHOD__ );
 117
 118                 $dbr = wfGetDB( DB_SLAVE );
 119
 120                 /* Get a job from the slave, start with an offset,
 121                         scan full set afterwards, avoid hitting purged rows
 122
 123                         NB: If random fetch previously was used, offset
 124                                 will always be ahead of few entries
 125                 */
 126
 127                 $conditions = self::defaultQueueConditions();
 128
 129                 $offset = intval( $offset );
 130                 $options = array( 'ORDER BY' => 'job_id', 'USE INDEX' => 'PRIMARY' );
 131
 132                 $row = $dbr->selectRow( 'job', '*',
 133                         array_merge( $conditions, array( "job_id >= $offset" ) ),
 134                         __METHOD__,
 135                         $options
 136                 );
 137
 138                 // Refetching without offset is needed as some of job IDs could have had delayed commits
 139                 // and have lower IDs than jobs already executed, blame concurrency :)
 140                 //
 141                 if ( $row === false ) {
 142                         if ( $offset != 0 ) {
 143                                 $row = $dbr->selectRow( 'job', '*', $conditions, __METHOD__, $options );
 144                         }
 145
 146                         if ( $row === false ) {
 147                                 wfProfileOut( __METHOD__ );
 148                                 return false;
 149                         }
 150                 }
 151
 152                 // Try to delete it from the master
 153                 $dbw = wfGetDB( DB_MASTER );
 154                 $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 155                 $affected = $dbw->affectedRows();
 156
 157                 if ( !$affected ) {
 158                         // Failed, someone else beat us to it
 159                         // Try getting a random row
 160                         $row = $dbw->selectRow( 'job', array( 'minjob' => 'MIN(job_id)',
 161                                 'maxjob' => 'MAX(job_id)' ), '1=1', __METHOD__ );
 162                         if ( $row === false || is_null( $row->minjob ) || is_null( $row->maxjob ) ) {
 163                                 // No jobs to get
 164                                 wfProfileOut( __METHOD__ );
 165                                 return false;
 166                         }
 167                         // Get the random row
 168                         $row = $dbw->selectRow( 'job', '*',
 169                                 'job_id >= ' . mt_rand( $row->minjob, $row->maxjob ), __METHOD__ );
 170                         if ( $row === false ) {
 171                                 // Random job gone before we got the chance to select it
 172                                 // Give up
 173                                 wfProfileOut( __METHOD__ );
 174                                 return false;
 175                         }
 176                         // Delete the random row
 177                         $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 178                         $affected = $dbw->affectedRows();
 179
 180                         if ( !$affected ) {
 181                                 // Random job gone before we exclusively deleted it
 182                                 // Give up
 183                                 wfProfileOut( __METHOD__ );
 184                                 return false;
 185                         }
 186                 }
 187
 188                 // If execution got to here, there's a row in $row that has been deleted from the database
 189                 // by this thread. Hence the concurrent pop was successful.
 190                 wfIncrStats( 'job-pop' );
 191                 $namespace = $row->job_namespace;
 192                 $dbkey = $row->job_title;
 193                 $title = Title::makeTitleSafe( $namespace, $dbkey );
 194
 195                 if ( is_null( $title ) ) {
 196                         wfProfileOut( __METHOD__ );
 197                         return false;
 198                 }
 199
 200                 $job = Job::factory( $row->job_cmd, $title, Job::extractBlob( $row->job_params ), $row->job_id );
 201
 202                 // Remove any duplicates it may have later in the queue
 203                 $job->removeDuplicates();
 204
 205                 wfProfileOut( __METHOD__ );
 206                 return $job;
 207         }
 208
 209         /**
 210          * Create the appropriate object to handle a specific job
 211          *
 212          * @param $command String: Job command
 213          * @param $title Title: Associated title
 214          * @param $params Array|bool: Job parameters
 215          * @param $id Int: Job identifier
 216          * @throws MWException
 217          * @return Job
 218          */
 219         static function factory( $command, Title $title, $params = false, $id = 0 ) {
 220                 global $wgJobClasses;
 221                 if( isset( $wgJobClasses[$command] ) ) {
 222                         $class = $wgJobClasses[$command];
 223                         return new $class( $title, $params, $id );
 224                 }
 225                 throw new MWException( "Invalid job command `{$command}`" );
 226         }
 227
 228         /**
 229          * @param $params
 230          * @return string
 231          */
 232         static function makeBlob( $params ) {
 233                 if ( $params !== false ) {
 234                         return serialize( $params );
 235                 } else {
 236                         return '';
 237                 }
 238         }
 239
 240         /**
 241          * @param $blob
 242          * @return bool|mixed
 243          */
 244         static function extractBlob( $blob ) {
 245                 if ( (string)$blob !== '' ) {
 246                         return unserialize( $blob );
 247                 } else {
 248                         return false;
 249                 }
 250         }
 251
 252         /**
 253          * Batch-insert a group of jobs into the queue.
 254          * This will be wrapped in a transaction with a forced commit.
 255          *
 256          * This may add duplicate at insert time, but they will be
 257          * removed later on, when the first one is popped.
 258          *
 259          * @param $jobs array of Job objects
 260          */
 261         static function batchInsert( $jobs ) {
 262                 if ( !count( $jobs ) ) {
 263                         return;
 264                 }
 265                 $dbw = wfGetDB( DB_MASTER );
 266                 $rows = array();
 267
 268                 /**
 269                  * @var $job Job
 270                  */
 271                 foreach ( $jobs as $job ) {
 272                         $rows[] = $job->insertFields();
 273                         if ( count( $rows ) >= 50 ) {
 274                                 # Do a small transaction to avoid slave lag
 275                                 $dbw->begin( __METHOD__ );
 276                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 277                                 $dbw->commit( __METHOD__ );
 278                                 $rows = array();
 279                         }
 280                 }
 281                 if ( $rows ) { // last chunk
 282                         $dbw->begin( __METHOD__ );
 283                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 284                         $dbw->commit( __METHOD__ );
 285                 }
 286                 wfIncrStats( 'job-insert', count( $jobs ) );
 287         }
 288
 289         /**
 290          * Insert a group of jobs into the queue.
 291          *
 292          * Same as batchInsert() but does not commit and can thus
 293          * be rolled-back as part of a larger transaction. However,
 294          * large batches of jobs can cause slave lag.
 295          *
 296          * @param $jobs array of Job objects
 297          */
 298         static function safeBatchInsert( $jobs ) {
 299                 if ( !count( $jobs ) ) {
 300                         return;
 301                 }
 302                 $dbw = wfGetDB( DB_MASTER );
 303                 $rows = array();
 304                 foreach ( $jobs as $job ) {
 305                         $rows[] = $job->insertFields();
 306                         if ( count( $rows ) >= 500 ) {
 307                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 308                                 $rows = array();
 309                         }
 310                 }
 311                 if ( $rows ) { // last chunk
 312                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 313                 }
 314                 wfIncrStats( 'job-insert', count( $jobs ) );
 315         }
 316
 317
 318         /**
 319          * SQL conditions to apply on most JobQueue queries
 320          *
 321          * Whenever we exclude jobs types from the default queue, we want to make
 322          * sure that queries to the job queue actually ignore them.
 323          *
 324          * @return array SQL conditions suitable for Database:: methods
 325          */
 326         static function defaultQueueConditions( ) {
 327                 global $wgJobTypesExcludedFromDefaultQueue;
 328                 $conditions = array();
 329                 if ( count( $wgJobTypesExcludedFromDefaultQueue ) > 0 ) {
 330                         $dbr = wfGetDB( DB_SLAVE );
 331                         foreach ( $wgJobTypesExcludedFromDefaultQueue as $cmdType ) {
 332                                 $conditions[] = "job_cmd != " . $dbr->addQuotes( $cmdType );
 333                         }
 334                 }
 335                 return $conditions;
 336         }
 337
 338         /*-------------------------------------------------------------------------
 339          * Non-static functions
 340          *------------------------------------------------------------------------*/
 341
 342         /**
 343          * @param $command
 344          * @param $title
 345          * @param $params array|bool
 346          * @param $id int
 347          */
 348         function __construct( $command, $title, $params = false, $id = 0 ) {
 349                 $this->command = $command;
 350                 $this->title = $title;
 351                 $this->params = $params;
 352                 $this->id = $id;
 353
 354                 // A bit of premature generalisation
 355                 // Oh well, the whole class is premature generalisation really
 356                 $this->removeDuplicates = true;
 357         }
 358
 359         /**
 360          * Insert a single job into the queue.
 361          * @return bool true on success
 362          */
 363         function insert() {
 364                 $fields = $this->insertFields();
 365
 366                 $dbw = wfGetDB( DB_MASTER );
 367
 368                 if ( $this->removeDuplicates ) {
 369                         $res = $dbw->select( 'job', array( '1' ), $fields, __METHOD__ );
 370                         if ( $dbw->numRows( $res ) ) {
 371                                 return true;
 372                         }
 373                 }
 374                 wfIncrStats( 'job-insert' );
 375                 return $dbw->insert( 'job', $fields, __METHOD__ );
 376         }
 377
 378         /**
 379          * @return array
 380          */
 381         protected function insertFields() {
 382                 $dbw = wfGetDB( DB_MASTER );
 383                 return array(
 384                         'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
 385                         'job_cmd' => $this->command,
 386                         'job_namespace' => $this->title->getNamespace(),
 387                         'job_title' => $this->title->getDBkey(),
 388                         'job_timestamp' => $dbw->timestamp(),
 389                         'job_params' => Job::makeBlob( $this->params )
 390                 );
 391         }
 392
 393         /**
 394          * Remove jobs in the job queue which are duplicates of this job.
 395          * This is deadlock-prone and so starts its own transaction.
 396          */
 397         function removeDuplicates() {
 398                 if ( !$this->removeDuplicates ) {
 399                         return;
 400                 }
 401
 402                 $fields = $this->insertFields();
 403                 unset( $fields['job_id'] );
 404                 $dbw = wfGetDB( DB_MASTER );
 405                 $dbw->begin( __METHOD__ );
 406                 $dbw->delete( 'job', $fields, __METHOD__ );
 407                 $affected = $dbw->affectedRows();
 408                 $dbw->commit( __METHOD__ );
 409                 if ( $affected ) {
 410                         wfIncrStats( 'job-dup-delete', $affected );
 411                 }
 412         }
 413
 414         /**
 415          * @return string
 416          */
 417         function toString() {
 418                 $paramString = '';
 419                 if ( $this->params ) {
 420                         foreach ( $this->params as $key => $value ) {
 421                                 if ( $paramString != '' ) {
 422                                         $paramString .= ' ';
 423                                 }
 424                                 $paramString .= "$key=$value";
 425                         }
 426                 }
 427
 428                 if ( is_object( $this->title ) ) {
 429                         $s = "{$this->command} " . $this->title->getPrefixedDBkey();
 430                         if ( $paramString !== '' ) {
 431                                 $s .= ' ' . $paramString;
 432                         }
 433                         return $s;
 434                 } else {
 435                         return "{$this->command} $paramString";
 436                 }
 437         }
 438
 439         protected function setLastError( $error ) {
 440                 $this->error = $error;
 441         }
 442
 443         function getLastError() {
 444                 return $this->error;
 445         }
 446 }