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                 $dbw->commit( __METHOD__ );
 157
 158                 if ( !$affected ) {
 159                         // Failed, someone else beat us to it
 160                         // Try getting a random row
 161                         $row = $dbw->selectRow( 'job', array( 'MIN(job_id) as minjob',
 162                                 'MAX(job_id) as maxjob' ), '1=1', __METHOD__ );
 163                         if ( $row === false || is_null( $row->minjob ) || is_null( $row->maxjob ) ) {
 164                                 // No jobs to get
 165                                 wfProfileOut( __METHOD__ );
 166                                 return false;
 167                         }
 168                         // Get the random row
 169                         $row = $dbw->selectRow( 'job', '*',
 170                                 'job_id >= ' . mt_rand( $row->minjob, $row->maxjob ), __METHOD__ );
 171                         if ( $row === false ) {
 172                                 // Random job gone before we got the chance to select it
 173                                 // Give up
 174                                 wfProfileOut( __METHOD__ );
 175                                 return false;
 176                         }
 177                         // Delete the random row
 178                         $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 179                         $affected = $dbw->affectedRows();
 180                         $dbw->commit( __METHOD__ );
 181
 182                         if ( !$affected ) {
 183                                 // Random job gone before we exclusively deleted it
 184                                 // Give up
 185                                 wfProfileOut( __METHOD__ );
 186                                 return false;
 187                         }
 188                 }
 189
 190                 // If execution got to here, there's a row in $row that has been deleted from the database
 191                 // by this thread. Hence the concurrent pop was successful.
 192                 wfIncrStats( 'job-pop' );
 193                 $namespace = $row->job_namespace;
 194                 $dbkey = $row->job_title;
 195                 $title = Title::makeTitleSafe( $namespace, $dbkey );
 196
 197                 if ( is_null( $title ) ) {
 198                         wfProfileOut( __METHOD__ );
 199                         return false;
 200                 }
 201
 202                 $job = Job::factory( $row->job_cmd, $title, Job::extractBlob( $row->job_params ), $row->job_id );
 203
 204                 // Remove any duplicates it may have later in the queue
 205                 $job->removeDuplicates();
 206
 207                 wfProfileOut( __METHOD__ );
 208                 return $job;
 209         }
 210
 211         /**
 212          * Create the appropriate object to handle a specific job
 213          *
 214          * @param $command String: Job command
 215          * @param $title Title: Associated title
 216          * @param $params Array|bool: Job parameters
 217          * @param $id Int: Job identifier
 218          * @throws MWException
 219          * @return Job
 220          */
 221         static function factory( $command, Title $title, $params = false, $id = 0 ) {
 222                 global $wgJobClasses;
 223                 if( isset( $wgJobClasses[$command] ) ) {
 224                         $class = $wgJobClasses[$command];
 225                         return new $class( $title, $params, $id );
 226                 }
 227                 throw new MWException( "Invalid job command `{$command}`" );
 228         }
 229
 230         /**
 231          * @param $params
 232          * @return string
 233          */
 234         static function makeBlob( $params ) {
 235                 if ( $params !== false ) {
 236                         return serialize( $params );
 237                 } else {
 238                         return '';
 239                 }
 240         }
 241
 242         /**
 243          * @param $blob
 244          * @return bool|mixed
 245          */
 246         static function extractBlob( $blob ) {
 247                 if ( (string)$blob !== '' ) {
 248                         return unserialize( $blob );
 249                 } else {
 250                         return false;
 251                 }
 252         }
 253
 254         /**
 255          * Batch-insert a group of jobs into the queue.
 256          * This will be wrapped in a transaction with a forced commit.
 257          *
 258          * This may add duplicate at insert time, but they will be
 259          * removed later on, when the first one is popped.
 260          *
 261          * @param $jobs array of Job objects
 262          */
 263         static function batchInsert( $jobs ) {
 264                 if ( !count( $jobs ) ) {
 265                         return;
 266                 }
 267                 $dbw = wfGetDB( DB_MASTER );
 268                 $rows = array();
 269
 270                 /**
 271                  * @var $job Job
 272                  */
 273                 foreach ( $jobs as $job ) {
 274                         $rows[] = $job->insertFields();
 275                         if ( count( $rows ) >= 50 ) {
 276                                 # Do a small transaction to avoid slave lag
 277                                 $dbw->begin( __METHOD__ );
 278                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 279                                 $dbw->commit( __METHOD__ );
 280                                 $rows = array();
 281                         }
 282                 }
 283                 if ( $rows ) { // last chunk
 284                         $dbw->begin( __METHOD__ );
 285                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 286                         $dbw->commit( __METHOD__ );
 287                 }
 288                 wfIncrStats( 'job-insert', count( $jobs ) );
 289         }
 290
 291         /**
 292          * Insert a group of jobs into the queue.
 293          *
 294          * Same as batchInsert() but does not commit and can thus
 295          * be rolled-back as part of a larger transaction. However,
 296          * large batches of jobs can cause slave lag.
 297          *
 298          * @param $jobs array of Job objects
 299          */
 300         static function safeBatchInsert( $jobs ) {
 301                 if ( !count( $jobs ) ) {
 302                         return;
 303                 }
 304                 $dbw = wfGetDB( DB_MASTER );
 305                 $rows = array();
 306                 foreach ( $jobs as $job ) {
 307                         $rows[] = $job->insertFields();
 308                         if ( count( $rows ) >= 500 ) {
 309                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 310                                 $rows = array();
 311                         }
 312                 }
 313                 if ( $rows ) { // last chunk
 314                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 315                 }
 316                 wfIncrStats( 'job-insert', count( $jobs ) );
 317         }
 318
 319
 320         /**
 321          * SQL conditions to apply on most JobQueue queries
 322          *
 323          * Whenever we exclude jobs types from the default queue, we want to make
 324          * sure that queries to the job queue actually ignore them.
 325          *
 326          * @return array SQL conditions suitable for Database:: methods
 327          */
 328         static function defaultQueueConditions( ) {
 329                 global $wgJobTypesExcludedFromDefaultQueue;
 330                 $conditions = array();
 331                 if ( count( $wgJobTypesExcludedFromDefaultQueue ) > 0 ) {
 332                         $dbr = wfGetDB( DB_SLAVE );
 333                         foreach ( $wgJobTypesExcludedFromDefaultQueue as $cmdType ) {
 334                                 $conditions[] = "job_cmd != " . $dbr->addQuotes( $cmdType );
 335                         }
 336                 }
 337                 return $conditions;
 338         }
 339
 340         /*-------------------------------------------------------------------------
 341          * Non-static functions
 342          *------------------------------------------------------------------------*/
 343
 344         /**
 345          * @param $command
 346          * @param $title
 347          * @param $params array|bool
 348          * @param $id int
 349          */
 350         function __construct( $command, $title, $params = false, $id = 0 ) {
 351                 $this->command = $command;
 352                 $this->title = $title;
 353                 $this->params = $params;
 354                 $this->id = $id;
 355
 356                 // A bit of premature generalisation
 357                 // Oh well, the whole class is premature generalisation really
 358                 $this->removeDuplicates = true;
 359         }
 360
 361         /**
 362          * Insert a single job into the queue.
 363          * @return bool true on success
 364          */
 365         function insert() {
 366                 $fields = $this->insertFields();
 367
 368                 $dbw = wfGetDB( DB_MASTER );
 369
 370                 if ( $this->removeDuplicates ) {
 371                         $res = $dbw->select( 'job', array( '1' ), $fields, __METHOD__ );
 372                         if ( $dbw->numRows( $res ) ) {
 373                                 return true;
 374                         }
 375                 }
 376                 wfIncrStats( 'job-insert' );
 377                 return $dbw->insert( 'job', $fields, __METHOD__ );
 378         }
 379
 380         /**
 381          * @return array
 382          */
 383         protected function insertFields() {
 384                 $dbw = wfGetDB( DB_MASTER );
 385                 return array(
 386                         'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
 387                         'job_cmd' => $this->command,
 388                         'job_namespace' => $this->title->getNamespace(),
 389                         'job_title' => $this->title->getDBkey(),
 390                         'job_timestamp' => $dbw->timestamp(),
 391                         'job_params' => Job::makeBlob( $this->params )
 392                 );
 393         }
 394
 395         /**
 396          * Remove jobs in the job queue which are duplicates of this job.
 397          * This is deadlock-prone and so starts its own transaction.
 398          */
 399         function removeDuplicates() {
 400                 if ( !$this->removeDuplicates ) {
 401                         return;
 402                 }
 403
 404                 $fields = $this->insertFields();
 405                 unset( $fields['job_id'] );
 406                 $dbw = wfGetDB( DB_MASTER );
 407                 $dbw->begin( __METHOD__ );
 408                 $dbw->delete( 'job', $fields, __METHOD__ );
 409                 $affected = $dbw->affectedRows();
 410                 $dbw->commit( __METHOD__ );
 411                 if ( $affected ) {
 412                         wfIncrStats( 'job-dup-delete', $affected );
 413                 }
 414         }
 415
 416         /**
 417          * @return string
 418          */
 419         function toString() {
 420                 $paramString = '';
 421                 if ( $this->params ) {
 422                         foreach ( $this->params as $key => $value ) {
 423                                 if ( $paramString != '' ) {
 424                                         $paramString .= ' ';
 425                                 }
 426                                 $paramString .= "$key=$value";
 427                         }
 428                 }
 429
 430                 if ( is_object( $this->title ) ) {
 431                         $s = "{$this->command} " . $this->title->getPrefixedDBkey();
 432                         if ( $paramString !== '' ) {
 433                                 $s .= ' ' . $paramString;
 434                         }
 435                         return $s;
 436                 } else {
 437                         return "{$this->command} $paramString";
 438                 }
 439         }
 440
 441         protected function setLastError( $error ) {
 442                 $this->error = $error;
 443         }
 444
 445         function getLastError() {
 446                 return $this->error;
 447         }
 448 }