includes/job/JobQueueFederated.php

   1 <?php
   2 /**
   3  * Job queue code for federated queues.
   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 enqueueing and running of background jobs for federated queues
  26  *
  27  * This class allows for queues to be partitioned into smaller queues.
  28  * A partition is defined by the configuration for a JobQueue instance.
  29  * For example, one can set $wgJobTypeConf['refreshLinks'] to point to a
  30  * JobQueueFederated instance, which itself would consist of three JobQueueRedis
  31  * instances, each using their own redis server. This would allow for the jobs
  32  * to be split (evenly or based on weights) accross multiple servers if a single
  33  * server becomes impractical or expensive. Different JobQueue classes can be mixed.
  34  *
  35  * The basic queue configuration (e.g. "order", "claimTTL") of a federated queue
  36  * is inherited by the partition queues. Additional configuration defines what
  37  * section each wiki is in, what partition queues each section uses (and their weight),
  38  * and the JobQueue configuration for each partition. Some sections might only need a
  39  * single queue partition, like the sections for groups of small wikis.
  40  *
  41  * If used for performance, then $wgMainCacheType should be set to memcached/redis.
  42  * Note that "fifo" cannot be used for the ordering, since the data is distributed.
  43  * One can still use "timestamp" instead, as in "roughly timestamp ordered".
  44  *
  45  * @ingroup JobQueue
  46  * @since 1.22
  47  */
  48 class JobQueueFederated extends JobQueue {
  49         /** @var Array (wiki ID => section name) */
  50         protected $sectionsByWiki = array();
  51         /** @var Array (section name => (partition name => weight)) */
  52         protected $partitionsBySection = array();
  53         /** @var Array (section name => config array) */
  54         protected $configByPartition = array();
  55         /** @var Array (partition names => integer) */
  56         protected $partitionsNoPush = array();
  57
  58         /** @var Array (partition name => JobQueue) */
  59         protected $partitionQueues = array();
  60         /** @var BagOStuff */
  61         protected $cache;
  62
  63         const CACHE_TTL_SHORT = 30; // integer; seconds to cache info without re-validating
  64         const CACHE_TTL_LONG = 300; // integer; seconds to cache info that is kept up to date
  65
  66         /**
  67          * @params include:
  68          *  - sectionsByWiki      : A map of wiki IDs to section names.
  69          *                          Wikis will default to using the section "default".
  70          *  - partitionsBySection : Map of section names to maps of (partition name => weight).
  71          *                          A section called 'default' must be defined if not all wikis
  72          *                          have explicitly defined sections.
  73          *  - configByPartition   : Map of queue partition names to configuration arrays.
  74          *                          These configuration arrays are passed to JobQueue::factory().
  75          *                          The options set here are overriden by those passed to this
  76          *                          the federated queue itself (e.g. 'order' and 'claimTTL').
  77          *  - partitionsNoPush    : List of partition names that can handle pop() but not push().
  78          *                          This can be used to migrate away from a certain partition.
  79          * @param array $params
  80          */
  81         protected function __construct( array $params ) {
  82                 parent::__construct( $params );
  83                 $this->sectionsByWiki = $params['sectionsByWiki'];
  84                 $this->partitionsBySection = $params['partitionsBySection'];
  85                 $this->configByPartition = $params['configByPartition'];
  86                 if ( isset( $params['partitionsNoPush'] ) ) {
  87                         $this->partitionsNoPush = array_flip( $params['partitionsNoPush'] );
  88                 }
  89                 $baseConfig = $params;
  90                 foreach ( array( 'class', 'sectionsByWiki',
  91                         'partitionsBySection', 'configByPartition', 'partitionsNoPush' ) as $o )
  92                 {
  93                         unset( $baseConfig[$o] );
  94                 }
  95                 foreach ( $this->getPartitionMap() as $partition => $w ) {
  96                         if ( !isset( $this->configByPartition[$partition] ) ) {
  97                                 throw new MWException( "No configuration for partition '$partition'." );
  98                         }
  99                         $this->partitionQueues[$partition] = JobQueue::factory(
 100                                 $baseConfig + $this->configByPartition[$partition]
 101                         );
 102                 }
 103                 // Aggregate cache some per-queue values if there are multiple partition queues
 104                 $this->cache = $this->isFederated() ? wfGetMainCache() : new EmptyBagOStuff();
 105         }
 106
 107         protected function supportedOrders() {
 108                 // No FIFO due to partitioning, though "rough timestamp order" is supported
 109                 return array( 'undefined', 'random', 'timestamp' );
 110         }
 111
 112         protected function optimalOrder() {
 113                 return 'undefined'; // defer to the partitions
 114         }
 115
 116         protected function supportsDelayedJobs() {
 117                 return true; // defer checks to the partitions
 118         }
 119
 120         protected function doIsEmpty() {
 121                 $key = $this->getCacheKey( 'empty' );
 122
 123                 $isEmpty = $this->cache->get( $key );
 124                 if ( $isEmpty === 'true' ) {
 125                         return true;
 126                 } elseif ( $isEmpty === 'false' ) {
 127                         return false;
 128                 }
 129
 130                 foreach ( $this->partitionQueues as $queue ) {
 131                         if ( !$queue->doIsEmpty() ) {
 132                                 $this->cache->add( $key, 'false', self::CACHE_TTL_LONG );
 133                                 return false;
 134                         }
 135                 }
 136
 137                 $this->cache->add( $key, 'true', self::CACHE_TTL_LONG );
 138                 return true;
 139         }
 140
 141         protected function doGetSize() {
 142                 return $this->getCrossPartitionSum( 'size', 'doGetSize' );
 143         }
 144
 145         protected function doGetAcquiredCount() {
 146                 return $this->getCrossPartitionSum( 'acquiredcount', 'doGetAcquiredCount' );
 147         }
 148
 149         protected function doGetDelayedCount() {
 150                 return $this->getCrossPartitionSum( 'delayedcount', 'doGetDelayedCount' );
 151         }
 152
 153         protected function doGetAbandonedCount() {
 154                 return $this->getCrossPartitionSum( 'abandonedcount', 'doGetAbandonedCount' );
 155         }
 156
 157         /**
 158          * @param string $type
 159          * @param string $method
 160          * @return integer
 161          */
 162         protected function getCrossPartitionSum( $type, $method ) {
 163                 $key = $this->getCacheKey( $type );
 164
 165                 $count = $this->cache->get( $key );
 166                 if ( is_int( $count ) ) {
 167                         return $count;
 168                 }
 169
 170                 $count = 0;
 171                 foreach ( $this->partitionQueues as $queue ) {
 172                         $count += $queue->$method();
 173                 }
 174
 175                 $this->cache->set( $key, $count, self::CACHE_TTL_SHORT );
 176                 return $count;
 177         }
 178
 179         protected function doBatchPush( array $jobs, $flags ) {
 180                 if ( !count( $jobs ) ) {
 181                         return true; // nothing to do
 182                 }
 183
 184                 $partitionsTry = array_diff_key(
 185                         $this->getPartitionMap(),
 186                         $this->partitionsNoPush
 187                 ); // (partition => weight)
 188
 189                 // Try to insert the jobs and update $partitionsTry on any failures
 190                 $jobsLeft = $this->tryJobInsertions( $jobs, $partitionsTry, $flags );
 191                 if ( count( $jobsLeft ) ) { // some jobs failed to insert?
 192                         // Try to insert the remaning jobs once more, ignoring the bad partitions
 193                         return !count( $this->tryJobInsertions( $jobsLeft, $partitionsTry, $flags ) );
 194                 } else {
 195                         return true;
 196                 }
 197         }
 198
 199         /**
 200          * @param array $jobs
 201          * @param array $partitionsTry
 202          * @param integer $flags
 203          * @return array List of Job object that could not be inserted
 204          */
 205         protected function tryJobInsertions( array $jobs, array &$partitionsTry, $flags ) {
 206                 if ( !count( $partitionsTry ) ) {
 207                         return $jobs; // can't insert anything
 208                 }
 209
 210                 $jobsLeft = array();
 211
 212                 $partitionRing = new HashRing( $partitionsTry );
 213                 // Because jobs are spread across partitions, per-job de-duplication needs
 214                 // to use a consistent hash to avoid allowing duplicate jobs per partition.
 215                 // When inserting a batch of de-duplicated jobs, QOS_ATOMIC is disregarded.
 216                 $uJobsByPartition = array(); // (partition name => job list)
 217                 foreach ( $jobs as $key => $job ) {
 218                         if ( $job->ignoreDuplicates() ) {
 219                                 $sha1 = sha1( serialize( $job->getDeduplicationInfo() ) );
 220                                 $uJobsByPartition[$partitionRing->getLocation( $sha1 )][] = $job;
 221                                 unset( $jobs[$key] );
 222                         }
 223                 }
 224                 // Get the batches of jobs that are not de-duplicated
 225                 if ( $flags & self::QOS_ATOMIC ) {
 226                         $nuJobBatches = array( $jobs ); // all or nothing
 227                 } else {
 228                         // Split the jobs into batches and spread them out over servers if there
 229                         // are many jobs. This helps keep the partitions even. Otherwise, send all
 230                         // the jobs to a single partition queue to avoids the extra connections.
 231                         $nuJobBatches = array_chunk( $jobs, 300 );
 232                 }
 233
 234                 // Insert the de-duplicated jobs into the queues...
 235                 foreach ( $uJobsByPartition as $partition => $jobBatch ) {
 236                         $queue = $this->partitionQueues[$partition];
 237                         if ( $queue->doBatchPush( $jobBatch, $flags ) ) {
 238                                 $key = $this->getCacheKey( 'empty' );
 239                                 $this->cache->set( $key, 'false', JobQueueDB::CACHE_TTL_LONG );
 240                         } else {
 241                                 unset( $partitionsTry[$partition] ); // blacklist partition
 242                                 $jobsLeft = array_merge( $jobsLeft, $jobBatch ); // not inserted
 243                         }
 244                 }
 245                 // Insert the jobs that are not de-duplicated into the queues...
 246                 foreach ( $nuJobBatches as $jobBatch ) {
 247                         $partition = ArrayUtils::pickRandom( $partitionsTry );
 248                         if ( $partition === false ) { // all partitions at 0 weight?
 249                                 $jobsLeft = array_merge( $jobsLeft, $jobBatch ); // not inserted
 250                         } else {
 251                                 $queue = $this->partitionQueues[$partition];
 252                                 if ( $queue->doBatchPush( $jobBatch, $flags ) ) {
 253                                         $key = $this->getCacheKey( 'empty' );
 254                                         $this->cache->set( $key, 'false', JobQueueDB::CACHE_TTL_LONG );
 255                                 } else {
 256                                         unset( $partitionsTry[$partition] ); // blacklist partition
 257                                         $jobsLeft = array_merge( $jobsLeft, $jobBatch ); // not inserted
 258                                 }
 259                         }
 260                 }
 261
 262                 return $jobsLeft;
 263         }
 264
 265         protected function doPop() {
 266                 $key = $this->getCacheKey( 'empty' );
 267
 268                 $isEmpty = $this->cache->get( $key );
 269                 if ( $isEmpty === 'true' ) {
 270                         return false;
 271                 }
 272
 273                 $partitionsTry = $this->getPartitionMap(); // (partition => weight)
 274
 275                 while ( count( $partitionsTry ) ) {
 276                         $partition = ArrayUtils::pickRandom( $partitionsTry );
 277                         if ( $partition === false ) {
 278                                 break; // all partitions at 0 weight
 279                         }
 280                         $queue = $this->partitionQueues[$partition];
 281                         $job = $queue->pop();
 282                         if ( $job ) {
 283                                 $job->metadata['QueuePartition'] = $partition;
 284                                 return $job;
 285                         } else {
 286                                 unset( $partitionsTry[$partition] ); // blacklist partition
 287                         }
 288                 }
 289
 290                 $this->cache->set( $key, 'true', JobQueueDB::CACHE_TTL_LONG );
 291                 return false;
 292         }
 293
 294         protected function doAck( Job $job ) {
 295                 if ( !isset( $job->metadata['QueuePartition'] ) ) {
 296                         throw new MWException( "The given job has no defined partition name." );
 297                 }
 298                 return $this->partitionQueues[$job->metadata['QueuePartition']]->ack( $job );
 299         }
 300
 301         protected function doDelete() {
 302                 foreach ( $this->partitionQueues as $queue ) {
 303                         $queue->doDelete();
 304                 }
 305         }
 306
 307         protected function doWaitForBackups() {
 308                 foreach ( $this->partitionQueues as $queue ) {
 309                         $queue->waitForBackups();
 310                 }
 311         }
 312
 313         protected function doGetPeriodicTasks() {
 314                 $tasks = array();
 315                 foreach ( $this->partitionQueues as $partition => $queue ) {
 316                         foreach ( $queue->getPeriodicTasks() as $task => $def ) {
 317                                 $tasks["{$partition}:{$task}"] = $def;
 318                         }
 319                 }
 320                 return $tasks;
 321         }
 322
 323         protected function doFlushCaches() {
 324                 static $types = array(
 325                         'empty',
 326                         'size',
 327                         'acquiredcount',
 328                         'delayedcount',
 329                         'abandonedcount'
 330                 );
 331                 foreach ( $types as $type ) {
 332                         $this->cache->delete( $this->getCacheKey( $type ) );
 333                 }
 334                 foreach ( $this->partitionQueues as $queue ) {
 335                         $queue->doFlushCaches();
 336                 }
 337         }
 338
 339         public function getAllQueuedJobs() {
 340                 $iterator = new AppendIterator();
 341                 foreach ( $this->partitionQueues as $queue ) {
 342                         $iterator->append( $queue->getAllQueuedJobs() );
 343                 }
 344                 return $iterator;
 345         }
 346
 347         public function getAllDelayedJobs() {
 348                 $iterator = new AppendIterator();
 349                 foreach ( $this->partitionQueues as $queue ) {
 350                         $iterator->append( $queue->getAllDelayedJobs() );
 351                 }
 352                 return $iterator;
 353         }
 354
 355         public function setTestingPrefix( $key ) {
 356                 foreach ( $this->partitionQueues as $queue ) {
 357                         $queue->setTestingPrefix( $key );
 358                 }
 359         }
 360
 361         /**
 362          * @return Array Map of (partition name => weight)
 363          */
 364         protected function getPartitionMap() {
 365                 $section = isset( $this->sectionsByWiki[$this->wiki] )
 366                         ? $this->sectionsByWiki[$this->wiki]
 367                         : 'default';
 368                 if ( !isset( $this->partitionsBySection[$section] ) ) {
 369                         throw new MWException( "No configuration for section '$section'." );
 370                 }
 371                 return $this->partitionsBySection[$section];
 372         }
 373
 374         /**
 375          * @return bool The queue is actually split up across multiple queue partitions
 376          */
 377         protected function isFederated() {
 378                 return ( count( $this->getPartitionMap() ) > 1 );
 379         }
 380
 381         /**
 382          * @return string
 383          */
 384         private function getCacheKey( $property ) {
 385                 list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 386                 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type, $property );
 387         }
 388 }