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  * @author Aaron Schulz
  23  */
  24
  25 /**
  26  * Class to handle enqueueing and running of background jobs
  27  *
  28  * @ingroup JobQueue
  29  * @since 1.21
  30  */
  31 abstract class JobQueue {
  32         protected $wiki; // string; wiki ID
  33         protected $type; // string; job type
  34         protected $order; // string; job priority for pop()
  35         protected $claimTTL; // integer; seconds
  36
  37         const QoS_Atomic = 1; // integer; "all-or-nothing" job insertions
  38
  39         const MAX_ATTEMPTS = 3; // integer; number of times to try a job
  40
  41         /**
  42          * @param $params array
  43          */
  44         protected function __construct( array $params ) {
  45                 $this->wiki     = $params['wiki'];
  46                 $this->type     = $params['type'];
  47                 $this->order    = isset( $params['order'] ) ? $params['order'] : 'random';
  48                 $this->claimTTL = isset( $params['claimTTL'] ) ? $params['claimTTL'] : 0;
  49         }
  50
  51         /**
  52          * Get a job queue object of the specified type.
  53          * $params includes:
  54          *   - class    : What job class to use (determines job type)
  55          *   - wiki     : wiki ID of the wiki the jobs are for (defaults to current wiki)
  56          *   - type     : The name of the job types this queue handles
  57          *   - order    : Order that pop() selects jobs, one of "fifo", "timestamp" or "random".
  58          *                If "fifo" is used, the queue will effectively be FIFO. Note that
  59          *                job completion will not appear to be exactly FIFO if there are multiple
  60          *                job runners since jobs can take different times to finish once popped.
  61          *                If "timestamp" is used, the queue will at least be loosely ordered
  62          *                by timestamp, allowing for some jobs to be popped off out of order.
  63          *                If "random" is used, pop() will pick jobs in random order. This might be
  64          *                useful for improving concurrency depending on the queue storage medium.
  65          *   - claimTTL : If supported, the queue will recycle jobs that have been popped
  66          *                but not acknowledged as completed after this many seconds. Recycling
  67          *                of jobs simple means re-inserting them into the queue. Jobs can be
  68          *                attempted up to three times before being discarded.
  69          *
  70          * Queue classes should throw an exception if they do not support the options given.
  71          *
  72          * @param $params array
  73          * @return JobQueue
  74          * @throws MWException
  75          */
  76         final public static function factory( array $params ) {
  77                 $class = $params['class'];
  78                 if ( !MWInit::classExists( $class ) ) {
  79                         throw new MWException( "Invalid job queue class '$class'." );
  80                 }
  81                 $obj = new $class( $params );
  82                 if ( !( $obj instanceof self ) ) {
  83                         throw new MWException( "Class '$class' is not a " . __CLASS__ . " class." );
  84                 }
  85                 return $obj;
  86         }
  87
  88         /**
  89          * @return string Wiki ID
  90          */
  91         final public function getWiki() {
  92                 return $this->wiki;
  93         }
  94
  95         /**
  96          * @return string Job type that this queue handles
  97          */
  98         final public function getType() {
  99                 return $this->type;
 100         }
 101
 102         /**
 103          * Quickly check if the queue is empty (has no available jobs).
 104          * Queue classes should use caching if they are any slower without memcached.
 105          *
 106          * @return bool
 107          */
 108         final public function isEmpty() {
 109                 wfProfileIn( __METHOD__ );
 110                 $res = $this->doIsEmpty();
 111                 wfProfileOut( __METHOD__ );
 112                 return $res;
 113         }
 114
 115         /**
 116          * @see JobQueue::isEmpty()
 117          * @return bool
 118          */
 119         abstract protected function doIsEmpty();
 120
 121         /**
 122          * Get the number of available jobs in the queue.
 123          * Queue classes should use caching if they are any slower without memcached.
 124          *
 125          * @return integer
 126          */
 127         final public function getSize() {
 128                 wfProfileIn( __METHOD__ );
 129                 $res = $this->doGetSize();
 130                 wfProfileOut( __METHOD__ );
 131                 return $res;
 132         }
 133
 134         /**
 135          * @see JobQueue::getSize()
 136          * @return integer
 137          */
 138         abstract protected function doGetSize();
 139
 140         /**
 141          * Get the number of acquired jobs (these are temporarily out of the queue).
 142          * Queue classes should use caching if they are any slower without memcached.
 143          *
 144          * @return integer
 145          */
 146         final public function getAcquiredCount() {
 147                 wfProfileIn( __METHOD__ );
 148                 $res = $this->doGetAcquiredCount();
 149                 wfProfileOut( __METHOD__ );
 150                 return $res;
 151         }
 152
 153         /**
 154          * @see JobQueue::getAcquiredCount()
 155          * @return integer
 156          */
 157         abstract protected function doGetAcquiredCount();
 158
 159         /**
 160          * Push a single jobs into the queue.
 161          * This does not require $wgJobClasses to be set for the given job type.
 162          *
 163          * @param $jobs Job|Array
 164          * @param $flags integer Bitfield (supports JobQueue::QoS_Atomic)
 165          * @throws MWException
 166          * @return bool Returns false on failure
 167          */
 168         final public function push( $jobs, $flags = 0 ) {
 169                 $jobs = is_array( $jobs ) ? $jobs : array( $jobs );
 170
 171                 return $this->batchPush( $jobs, $flags );
 172         }
 173
 174         /**
 175          * Push a batch of jobs into the queue.
 176          * This does not require $wgJobClasses to be set for the given job type.
 177          *
 178          * @param $jobs array List of Jobs
 179          * @param $flags integer Bitfield (supports JobQueue::QoS_Atomic)
 180          * @throws MWException
 181          * @return bool Returns false on failure
 182          */
 183         final public function batchPush( array $jobs, $flags = 0 ) {
 184                 foreach ( $jobs as $job ) {
 185                         if ( $job->getType() !== $this->type ) {
 186                                 throw new MWException( "Got '{$job->getType()}' job; expected '{$this->type}'." );
 187                         }
 188                 }
 189                 wfProfileIn( __METHOD__ );
 190                 $ok = $this->doBatchPush( $jobs, $flags );
 191                 wfProfileOut( __METHOD__ );
 192                 return $ok;
 193         }
 194
 195         /**
 196          * @see JobQueue::batchPush()
 197          * @return bool
 198          */
 199         abstract protected function doBatchPush( array $jobs, $flags );
 200
 201         /**
 202          * Pop a job off of the queue.
 203          * This requires $wgJobClasses to be set for the given job type.
 204          *
 205          * @return Job|bool Returns false on failure
 206          */
 207         final public function pop() {
 208                 wfProfileIn( __METHOD__ );
 209                 $job = $this->doPop();
 210                 wfProfileOut( __METHOD__ );
 211                 return $job;
 212         }
 213
 214         /**
 215          * @see JobQueue::pop()
 216          * @return Job
 217          */
 218         abstract protected function doPop();
 219
 220         /**
 221          * Acknowledge that a job was completed.
 222          *
 223          * This does nothing for certain queue classes or if "claimTTL" is not set.
 224          *
 225          * @param $job Job
 226          * @throws MWException
 227          * @return bool
 228          */
 229         final public function ack( Job $job ) {
 230                 if ( $job->getType() !== $this->type ) {
 231                         throw new MWException( "Got '{$job->getType()}' job; expected '{$this->type}'." );
 232                 }
 233                 wfProfileIn( __METHOD__ );
 234                 $ok = $this->doAck( $job );
 235                 wfProfileOut( __METHOD__ );
 236                 return $ok;
 237         }
 238
 239         /**
 240          * @see JobQueue::ack()
 241          * @return bool
 242          */
 243         abstract protected function doAck( Job $job );
 244
 245         /**
 246          * Register the "root job" of a given job into the queue for de-duplication.
 247          * This should only be called right *after* all the new jobs have been inserted.
 248          * This is used to turn older, duplicate, job entries into no-ops. The root job
 249          * information will remain in the registry until it simply falls out of cache.
 250          *
 251          * This requires that $job has two special fields in the "params" array:
 252          *   - rootJobSignature : hash (e.g. SHA1) that identifies the task
 253          *   - rootJobTimestamp : TS_MW timestamp of this instance of the task
 254          *
 255          * A "root job" is a conceptual job that consist of potentially many smaller jobs
 256          * that are actually inserted into the queue. For example, "refreshLinks" jobs are
 257          * spawned when a template is edited. One can think of the task as "update links
 258          * of pages that use template X" and an instance of that task as a "root job".
 259          * However, what actually goes into the queue are potentially many refreshLinks2 jobs.
 260          * Since these jobs include things like page ID ranges and DB master positions, and morph
 261          * into smaller refreshLinks2 jobs recursively, simple duplicate detection (like job_sha1)
 262          * for individual jobs being identical is not useful.
 263          *
 264          * In the case of "refreshLinks", if these jobs are still in the queue when the template
 265          * is edited again, we want all of these old refreshLinks jobs for that template to become
 266          * no-ops. This can greatly reduce server load, since refreshLinks jobs involves parsing.
 267          * Essentially, the new batch of jobs belong to a new "root job" and the older ones to a
 268          * previous "root job" for the same task of "update links of pages that use template X".
 269          *
 270          * This does nothing for certain queue classes.
 271          *
 272          * @param $job Job
 273          * @throws MWException
 274          * @return bool
 275          */
 276         final public function deduplicateRootJob( Job $job ) {
 277                 if ( $job->getType() !== $this->type ) {
 278                         throw new MWException( "Got '{$job->getType()}' job; expected '{$this->type}'." );
 279                 }
 280                 wfProfileIn( __METHOD__ );
 281                 $ok = $this->doDeduplicateRootJob( $job );
 282                 wfProfileOut( __METHOD__ );
 283                 return $ok;
 284         }
 285
 286         /**
 287          * @see JobQueue::deduplicateRootJob()
 288          * @param $job Job
 289          * @return bool
 290          */
 291         protected function doDeduplicateRootJob( Job $job ) {
 292                 return true;
 293         }
 294
 295         /**
 296          * Wait for any slaves or backup servers to catch up.
 297          *
 298          * This does nothing for certain queue classes.
 299          *
 300          * @return void
 301          */
 302         final public function waitForBackups() {
 303                 wfProfileIn( __METHOD__ );
 304                 $this->doWaitForBackups();
 305                 wfProfileOut( __METHOD__ );
 306         }
 307
 308         /**
 309          * @see JobQueue::waitForBackups()
 310          * @return void
 311          */
 312         protected function doWaitForBackups() {}
 313 }