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                         wfProfileOut( __METHOD__ );
 203                         return false;
 204                 }
 205
 206                 $job = Job::factory( $row->job_cmd, $title, Job::extractBlob( $row->job_params ), $row->job_id );
 207
 208                 // Remove any duplicates it may have later in the queue
 209                 $job->removeDuplicates();
 210
 211                 wfProfileOut( __METHOD__ );
 212                 return $job;
 213         }
 214
 215         /**
 216          * Create the appropriate object to handle a specific job
 217          *
 218          * @param $command String: Job command
 219          * @param $title Title: Associated title
 220          * @param $params Array: Job parameters
 221          * @param $id Int: Job identifier
 222          * @return Job
 223          */
 224         static function factory( $command, Title $title, $params = false, $id = 0 ) {
 225                 global $wgJobClasses;
 226                 if( isset( $wgJobClasses[$command] ) ) {
 227                         $class = $wgJobClasses[$command];
 228                         return new $class( $title, $params, $id );
 229                 }
 230                 throw new MWException( "Invalid job command `{$command}`" );
 231         }
 232
 233         /**
 234          * @param $params
 235          * @return string
 236          */
 237         static function makeBlob( $params ) {
 238                 if ( $params !== false ) {
 239                         return serialize( $params );
 240                 } else {
 241                         return '';
 242                 }
 243         }
 244
 245         /**
 246          * @param $blob
 247          * @return bool|mixed
 248          */
 249         static function extractBlob( $blob ) {
 250                 if ( (string)$blob !== '' ) {
 251                         return unserialize( $blob );
 252                 } else {
 253                         return false;
 254                 }
 255         }
 256
 257         /**
 258          * Batch-insert a group of jobs into the queue.
 259          * This will be wrapped in a transaction with a forced commit.
 260          *
 261          * This may add duplicate at insert time, but they will be
 262          * removed later on, when the first one is popped.
 263          *
 264          * @param $jobs array of Job objects
 265          */
 266         static function batchInsert( $jobs ) {
 267                 if ( !count( $jobs ) ) {
 268                         return;
 269                 }
 270                 $dbw = wfGetDB( DB_MASTER );
 271                 $rows = array();
 272
 273                 /**
 274                  * @var $job Job
 275                  */
 276                 foreach ( $jobs as $job ) {
 277                         $rows[] = $job->insertFields();
 278                         if ( count( $rows ) >= 50 ) {
 279                                 # Do a small transaction to avoid slave lag
 280                                 $dbw->begin( __METHOD__ );
 281                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 282                                 $dbw->commit( __METHOD__ );
 283                                 $rows = array();
 284                         }
 285                 }
 286                 if ( $rows ) { // last chunk
 287                         $dbw->begin( __METHOD__ );
 288                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 289                         $dbw->commit( __METHOD__ );
 290                 }
 291                 wfIncrStats( 'job-insert', count( $jobs ) );
 292         }
 293
 294         /**
 295          * Insert a group of jobs into the queue.
 296          *
 297          * Same as batchInsert() but does not commit and can thus
 298          * be rolled-back as part of a larger transaction. However,
 299          * large batches of jobs can cause slave lag.
 300          *
 301          * @param $jobs array of Job objects
 302          */
 303         static function safeBatchInsert( $jobs ) {
 304                 if ( !count( $jobs ) ) {
 305                         return;
 306                 }
 307                 $dbw = wfGetDB( DB_MASTER );
 308                 $rows = array();
 309                 foreach ( $jobs as $job ) {
 310                         $rows[] = $job->insertFields();
 311                         if ( count( $rows ) >= 500 ) {
 312                                 $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 313                                 $rows = array();
 314                         }
 315                 }
 316                 if ( $rows ) { // last chunk
 317                         $dbw->insert( 'job', $rows, __METHOD__, 'IGNORE' );
 318                 }
 319                 wfIncrStats( 'job-insert', count( $jobs ) );
 320         }
 321
 322         /*-------------------------------------------------------------------------
 323          * Non-static functions
 324          *------------------------------------------------------------------------*/
 325
 326         /**
 327          * @param $command
 328          * @param $title
 329          * @param $params array
 330          * @param int $id
 331          */
 332         function __construct( $command, $title, $params = false, $id = 0 ) {
 333                 $this->command = $command;
 334                 $this->title = $title;
 335                 $this->params = $params;
 336                 $this->id = $id;
 337
 338                 // A bit of premature generalisation
 339                 // Oh well, the whole class is premature generalisation really
 340                 $this->removeDuplicates = true;
 341         }
 342
 343         /**
 344          * Insert a single job into the queue.
 345          * @return bool true on success
 346          */
 347         function insert() {
 348                 $fields = $this->insertFields();
 349
 350                 $dbw = wfGetDB( DB_MASTER );
 351
 352                 if ( $this->removeDuplicates ) {
 353                         $res = $dbw->select( 'job', array( '1' ), $fields, __METHOD__ );
 354                         if ( $dbw->numRows( $res ) ) {
 355                                 return true;
 356                         }
 357                 }
 358                 wfIncrStats( 'job-insert' );
 359                 return $dbw->insert( 'job', $fields, __METHOD__ );
 360         }
 361
 362         /**
 363          * @return array
 364          */
 365         protected function insertFields() {
 366                 $dbw = wfGetDB( DB_MASTER );
 367                 return array(
 368                         'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
 369                         'job_cmd' => $this->command,
 370                         'job_namespace' => $this->title->getNamespace(),
 371                         'job_title' => $this->title->getDBkey(),
 372                         'job_timestamp' => $dbw->timestamp(),
 373                         'job_params' => Job::makeBlob( $this->params )
 374                 );
 375         }
 376
 377         /**
 378          * Remove jobs in the job queue which are duplicates of this job.
 379          * This is deadlock-prone and so starts its own transaction.
 380          */
 381         function removeDuplicates() {
 382                 if ( !$this->removeDuplicates ) {
 383                         return;
 384                 }
 385
 386                 $fields = $this->insertFields();
 387                 unset( $fields['job_id'] );
 388                 $dbw = wfGetDB( DB_MASTER );
 389                 $dbw->begin( __METHOD__ );
 390                 $dbw->delete( 'job', $fields, __METHOD__ );
 391                 $affected = $dbw->affectedRows();
 392                 $dbw->commit( __METHOD__ );
 393                 if ( $affected ) {
 394                         wfIncrStats( 'job-dup-delete', $affected );
 395                 }
 396         }
 397
 398         /**
 399          * @return string
 400          */
 401         function toString() {
 402                 $paramString = '';
 403                 if ( $this->params ) {
 404                         foreach ( $this->params as $key => $value ) {
 405                                 if ( $paramString != '' ) {
 406                                         $paramString .= ' ';
 407                                 }
 408                                 $paramString .= "$key=$value";
 409                         }
 410                 }
 411
 412                 if ( is_object( $this->title ) ) {
 413                         $s = "{$this->command} " . $this->title->getPrefixedDBkey();
 414                         if ( $paramString !== '' ) {
 415                                 $s .= ' ' . $paramString;
 416                         }
 417                         return $s;
 418                 } else {
 419                         return "{$this->command} $paramString";
 420                 }
 421         }
 422
 423         protected function setLastError( $error ) {
 424                 $this->error = $error;
 425         }
 426
 427         function getLastError() {
 428                 return $this->error;
 429         }
 430 }