includes/job/JobQueueDB.php

   1 <?php
   2 /**
   3  * Database-backed job queue 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  * @author Aaron Schulz
  22  */
  23
  24 /**
  25  * Class to handle job queues stored in the DB
  26  *
  27  * @ingroup JobQueue
  28  * @since 1.20
  29  */
  30 class JobQueueDB extends JobQueue {
  31         const CACHE_TTL      = 30; // integer; seconds
  32         const MAX_JOB_RANDOM = 2147483647; // 2^31 - 1; used for job_random
  33
  34         /**
  35          * @see JobQueue::doIsEmpty()
  36          * @return bool
  37          */
  38         protected function doIsEmpty() {
  39                 global $wgMemc;
  40
  41                 $key = $this->getEmptinessCacheKey();
  42
  43                 $isEmpty = $wgMemc->get( $key );
  44                 if ( $isEmpty === 'true' ) {
  45                         return true;
  46                 } elseif ( $isEmpty === 'false' ) {
  47                         return false;
  48                 }
  49
  50                 $found = $this->getSlaveDB()->selectField(
  51                         'job', '1', array( 'job_cmd' => $this->type ), __METHOD__
  52                 );
  53
  54                 $wgMemc->add( $key, $found ? 'false' : 'true', self::CACHE_TTL );
  55         }
  56
  57         /**
  58          * @see JobQueue::doBatchPush()
  59          * @return bool
  60          */
  61         protected function doBatchPush( array $jobs, $flags ) {
  62                 if ( count( $jobs ) ) {
  63                         $dbw = $this->getMasterDB();
  64
  65                         $rows = array();
  66                         foreach ( $jobs as $job ) {
  67                                 $rows[] = $this->insertFields( $job );
  68                         }
  69                         $atomic = ( $flags & self::QoS_Atomic );
  70                         $key    = $this->getEmptinessCacheKey();
  71                         $ttl    = self::CACHE_TTL;
  72
  73                         $dbw->onTransactionIdle( function() use ( $dbw, $rows, $atomic, $key, $ttl ) {
  74                                 global $wgMemc;
  75
  76                                 $autoTrx = $dbw->getFlag( DBO_TRX ); // automatic begin() enabled?
  77                                 if ( $atomic ) {
  78                                         $dbw->begin(); // wrap all the job additions in one transaction
  79                                 } else {
  80                                         $dbw->clearFlag( DBO_TRX ); // make each query its own transaction
  81                                 }
  82                                 try {
  83                                         foreach ( array_chunk( $rows, 50 ) as $rowBatch ) { // avoid slave lag
  84                                                 $dbw->insert( 'job', $rowBatch, __METHOD__ );
  85                                         }
  86                                 } catch ( DBError $e ) {
  87                                         if ( $atomic ) {
  88                                                 $dbw->rollback();
  89                                         } else {
  90                                                 $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
  91                                         }
  92                                         throw $e;
  93                                 }
  94                                 if ( $atomic ) {
  95                                         $dbw->commit();
  96                                 } else {
  97                                         $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
  98                                 }
  99
 100                                 $wgMemc->set( $key, 'false', $ttl );
 101                         } );
 102                 }
 103
 104                 return true;
 105         }
 106
 107         /**
 108          * @see JobQueue::doPop()
 109          * @return Job|bool
 110          */
 111         protected function doPop() {
 112                 global $wgMemc;
 113
 114                 if ( $wgMemc->get( $this->getEmptinessCacheKey() ) === 'true' ) {
 115                         return false; // queue is empty
 116                 }
 117
 118                 $dbw = $this->getMasterDB();
 119                 $dbw->commit( __METHOD__, 'flush' ); // flush existing transaction
 120
 121                 $uuid = wfRandomString( 32 ); // pop attempt
 122                 $job = false; // job popped off
 123                 $autoTrx = $dbw->getFlag( DBO_TRX ); // automatic begin() enabled?
 124                 $dbw->clearFlag( DBO_TRX ); // make each query its own transaction
 125                 try {
 126                         do { // retry when our row is invalid or deleted as a duplicate
 127                                 // Try to reserve a row in the DB...
 128                                 if ( $this->order === 'timestamp' ) { // oldest first
 129                                         $row = $this->claimOldest( $uuid );
 130                                 } else { // random first
 131                                         $rand = mt_rand( 0, self::MAX_JOB_RANDOM ); // encourage concurrent UPDATEs
 132                                         $gte  = (bool)mt_rand( 0, 1 ); // find rows with rand before/after $rand
 133                                         $row  = $this->claimRandom( $uuid, $rand, $gte );
 134                                         if ( !$row ) { // need to try the other direction
 135                                                 $row = $this->claimRandom( $uuid, $rand, !$gte );
 136                                         }
 137                                 }
 138                                 // Check if we found a row to reserve...
 139                                 if ( !$row ) {
 140                                         $wgMemc->set( $this->getEmptinessCacheKey(), 'true', self::CACHE_TTL );
 141                                         break; // nothing to do
 142                                 }
 143                                 // Get the job object from the row...
 144                                 $title = Title::makeTitleSafe( $row->job_namespace, $row->job_title );
 145                                 if ( !$title ) {
 146                                         $dbw->delete( 'job', array( 'job_id' => $row->job_id ), __METHOD__ );
 147                                         wfDebugLog( 'JobQueueDB', "Row has invalid title '{$row->job_title}'." );
 148                                         continue; // try again
 149                                 }
 150                                 $job = Job::factory( $row->job_cmd, $title,
 151                                         self::extractBlob( $row->job_params ), $row->job_id );
 152                                 // Delete any *other* duplicate jobs in the queue...
 153                                 if ( $job->ignoreDuplicates() && strlen( $row->job_sha1 ) ) {
 154                                         $dbw->delete( 'job',
 155                                                 array( 'job_sha1' => $row->job_sha1,
 156                                                         "job_id != {$dbw->addQuotes( $row->job_id )}" ),
 157                                                 __METHOD__
 158                                         );
 159                                 }
 160                                 break; // done
 161                         } while( true );
 162                 } catch ( DBError $e ) {
 163                         $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
 164                         throw $e;
 165                 }
 166                 $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
 167
 168                 return $job;
 169         }
 170
 171         /**
 172          * Reserve a row with a single UPDATE without holding row locks over RTTs...
 173          *
 174          * @param $uuid string 32 char hex string
 175          * @param $rand integer Random unsigned integer (31 bits)
 176          * @param $gte bool Search for job_random >= $random (otherwise job_random <= $random)
 177          * @return Row|false
 178          */
 179         protected function claimRandom( $uuid, $rand, $gte ) {
 180                 $dbw  = $this->getMasterDB();
 181                 $dir  = $gte ? 'ASC' : 'DESC';
 182                 $ineq = $gte ? '>=' : '<=';
 183
 184                 $row = false; // the row acquired
 185                 // This uses a replication safe method for acquiring jobs. One could use UPDATE+LIMIT
 186                 // instead, but that either uses ORDER BY (in which case it deadlocks in MySQL) or is
 187                 // not replication safe. Due to http://bugs.mysql.com/bug.php?id=6980, subqueries cannot
 188                 // be used here with MySQL.
 189                 do {
 190                         $row = $dbw->selectRow( 'job', '*', // find a random job
 191                                 array(
 192                                         'job_cmd'   => $this->type,
 193                                         'job_token' => '',
 194                                         "job_random {$ineq} {$dbw->addQuotes( $rand )}" ),
 195                                 __METHOD__,
 196                                 array( 'ORDER BY' => "job_random {$dir}" )
 197                         );
 198                         if ( $row ) { // claim the job
 199                                 $dbw->update( 'job', // update by PK
 200                                         array( 'job_token' => $uuid, 'job_token_timestamp' => $dbw->timestamp() ),
 201                                         array( 'job_cmd' => $this->type, 'job_id' => $row->job_id, 'job_token' => '' ),
 202                                         __METHOD__
 203                                 );
 204                                 // This might get raced out by another runner when claiming the previously
 205                                 // selected row. The use of job_random should minimize this problem, however.
 206                                 if ( !$dbw->affectedRows() ) {
 207                                         $row = false; // raced out
 208                                 }
 209                         } else {
 210                                 break; // nothing to do
 211                         }
 212                 } while ( !$row );
 213
 214                 return $row;
 215         }
 216
 217         /**
 218          * Reserve a row with a single UPDATE without holding row locks over RTTs...
 219          *
 220          * @param $uuid string 32 char hex string
 221          * @return Row|false
 222          */
 223         protected function claimOldest( $uuid ) {
 224                 $dbw  = $this->getMasterDB();
 225
 226                 $row = false; // the row acquired
 227                 do {
 228                         if ( $dbw->getType() === 'mysql' ) {
 229                                 // Per http://bugs.mysql.com/bug.php?id=6980, we can't use subqueries on the
 230                                 // same table being changed in an UPDATE query in MySQL (gives Error: 1093).
 231                                 // Oracle and Postgre have no such limitation. However, MySQL offers an
 232                                 // alternative here by supporting ORDER BY + LIMIT for UPDATE queries.
 233                                 $dbw->query( "UPDATE {$dbw->tableName( 'job' )}
 234                                         SET
 235                                                 job_token = {$dbw->addQuotes( $uuid ) },
 236                                                 job_token_timestamp = {$dbw->addQuotes( $dbw->timestamp() )}
 237                                         WHERE (
 238                                                 job_cmd = {$dbw->addQuotes( $this->type )}
 239                                                 AND job_token = {$dbw->addQuotes( '' )}
 240                                         ) ORDER BY job_random ASC LIMIT 1",
 241                                         __METHOD__
 242                                 );
 243                         } else {
 244                                 // Use a subquery to find the job, within an UPDATE to claim it.
 245                                 // This uses as much of the DB wrapper functions as possible.
 246                                 $dbw->update( 'job',
 247                                         array( 'job_token' => $uuid, 'job_token_timestamp' => $dbw->timestamp() ),
 248                                         array( 'job_id = (' .
 249                                                 $dbw->selectSQLText( 'job', 'job_id',
 250                                                         array( 'job_cmd' => $this->type, 'job_token' => '' ),
 251                                                         __METHOD__,
 252                                                         array( 'ORDER BY' => 'job_random ASC', 'LIMIT' => 1 ) ) .
 253                                                 ')'
 254                                         ),
 255                                         __METHOD__
 256                                 );
 257                         }
 258                         // Fetch any row that we just reserved...
 259                         if ( $dbw->affectedRows() ) {
 260                                 $row = $dbw->selectRow( 'job', '*',
 261                                         array( 'job_cmd' => $this->type, 'job_token' => $uuid ), __METHOD__
 262                                 );
 263                                 if ( !$row ) { // raced out by duplicate job removal
 264                                         wfDebugLog( 'JobQueueDB', "Row deleted as duplicate by another process." );
 265                                 }
 266                         } else {
 267                                 break; // nothing to do
 268                         }
 269                 } while ( !$row );
 270
 271                 return $row;
 272         }
 273
 274         /**
 275          * @see JobQueue::doAck()
 276          * @return Job|bool
 277          */
 278         protected function doAck( Job $job ) {
 279                 $dbw = $this->getMasterDB();
 280                 if ( $dbw->trxLevel() ) {
 281                         wfWarn( "Attempted to ack a job in a transaction; committing first." );
 282                         $dbw->commit(); // push existing transaction
 283                 }
 284
 285                 $autoTrx = $dbw->getFlag( DBO_TRX ); // automatic begin() enabled?
 286                 $dbw->clearFlag( DBO_TRX ); // make each query its own transaction
 287                 try {
 288                         // Delete a row with a single DELETE without holding row locks over RTTs...
 289                         $dbw->delete( 'job', array( 'job_cmd' => $this->type, 'job_id' => $job->getId() ) );
 290                 } catch ( Exception $e ) {
 291                         $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
 292                         throw $e;
 293                 }
 294                 $dbw->setFlag( $autoTrx ? DBO_TRX : 0 ); // restore automatic begin()
 295
 296                 return true;
 297         }
 298
 299         /**
 300          * @see JobQueue::doWaitForBackups()
 301          * @return void
 302          */
 303         protected function doWaitForBackups() {
 304                 wfWaitForSlaves();
 305         }
 306
 307         /**
 308          * @return DatabaseBase
 309          */
 310         protected function getSlaveDB() {
 311                 return wfGetDB( DB_SLAVE, array(), $this->wiki );
 312         }
 313
 314         /**
 315          * @return DatabaseBase
 316          */
 317         protected function getMasterDB() {
 318                 return wfGetDB( DB_MASTER, array(), $this->wiki );
 319         }
 320
 321         /**
 322          * @param $job Job
 323          * @return array
 324          */
 325         protected function insertFields( Job $job ) {
 326                 // Rows that describe the nature of the job
 327                 $descFields = array(
 328                         'job_cmd'       => $job->getType(),
 329                         'job_namespace' => $job->getTitle()->getNamespace(),
 330                         'job_title'     => $job->getTitle()->getDBkey(),
 331                         'job_params'    => self::makeBlob( $job->getParams() ),
 332                 );
 333                 // Additional job metadata
 334                 if ( $this->order === 'timestamp' ) { // oldest first
 335                         $random = time() - 1325376000; // seconds since "January 1, 2012"
 336                 } else { // random first
 337                         $random = mt_rand( 0, self::MAX_JOB_RANDOM );
 338                 }
 339                 $dbw = $this->getMasterDB();
 340                 $metaFields = array(
 341                         'job_id'        => $dbw->nextSequenceValue( 'job_job_id_seq' ),
 342                         'job_timestamp' => $dbw->timestamp(),
 343                         'job_sha1'      => wfBaseConvert( sha1( serialize( $descFields ) ), 16, 36, 32 ),
 344                         'job_random'    => mt_rand( 0, self::MAX_JOB_RANDOM )
 345                 );
 346                 return ( $descFields + $metaFields );
 347         }
 348
 349         /**
 350          * @return string
 351          */
 352         private function getEmptinessCacheKey() {
 353                 list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 354                 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type, 'isempty' );
 355         }
 356
 357         /**
 358          * @param $params
 359          * @return string
 360          */
 361         protected static function makeBlob( $params ) {
 362                 if ( $params !== false ) {
 363                         return serialize( $params );
 364                 } else {
 365                         return '';
 366                 }
 367         }
 368
 369         /**
 370          * @param $blob
 371          * @return bool|mixed
 372          */
 373         protected static function extractBlob( $blob ) {
 374                 if ( (string)$blob !== '' ) {
 375                         return unserialize( $blob );
 376                 } else {
 377                         return false;
 378                 }
 379         }
 380 }