3 * Database-backed job queue code.
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.
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.
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
21 * @author Aaron Schulz
25 * Class to handle job queues stored in the DB
30 class JobQueueDB
extends JobQueue
{
31 const CACHE_TTL
= 300; // integer; seconds to cache queue information
32 const MAX_AGE_PRUNE
= 604800; // integer; seconds a job can live once claimed
33 const MAX_ATTEMPTS
= 3; // integer; number of times to try a job
34 const MAX_JOB_RANDOM
= 2147483647; // integer; 2^31 - 1, used for job_random
37 * @see JobQueue::doIsEmpty()
40 protected function doIsEmpty() {
43 $key = $this->getEmptinessCacheKey();
45 $isEmpty = $wgMemc->get( $key );
46 if ( $isEmpty === 'true' ) {
48 } elseif ( $isEmpty === 'false' ) {
52 $found = $this->getSlaveDB()->selectField(
53 'job', '1', array( 'job_cmd' => $this->type
), __METHOD__
56 $wgMemc->add( $key, $found ?
'false' : 'true', self
::CACHE_TTL
);
61 * @see JobQueue::doBatchPush()
64 protected function doBatchPush( array $jobs, $flags ) {
65 if ( count( $jobs ) ) {
66 $dbw = $this->getMasterDB();
68 $rowSet = array(); // (sha1 => job) map for jobs that are de-duplicated
69 $rowList = array(); // list of jobs for jobs that are are not de-duplicated
71 foreach ( $jobs as $job ) {
72 $row = $this->insertFields( $job );
73 if ( $job->ignoreDuplicates() ) {
74 $rowSet[$row['job_sha1']] = $row;
80 $atomic = ( $flags & self
::QoS_Atomic
);
81 $key = $this->getEmptinessCacheKey();
82 $ttl = self
::CACHE_TTL
;
84 $dbw->onTransactionIdle(
85 function() use ( $dbw, $rowSet, $rowList, $atomic, $key, $ttl
90 $dbw->begin( __METHOD__
); // wrap all the job additions in one transaction
93 // Strip out any duplicate jobs that are already in the queue...
94 if ( count( $rowSet ) ) {
95 $res = $dbw->select( 'job', 'job_sha1',
97 // No job_type condition since it's part of the job_sha1 hash
98 'job_sha1' => array_keys( $rowSet ),
99 'job_token' => '' // unclaimed
103 foreach ( $res as $row ) {
104 wfDebug( "Job with hash '{$row->job_sha1}' is a duplicate." );
105 unset( $rowSet[$row->job_sha1
] ); // already enqueued
108 // Build the full list of job rows to insert
109 $rows = array_merge( $rowList, array_values( $rowSet ) );
110 // Insert the job rows in chunks to avoid slave lag...
111 foreach ( array_chunk( $rows, 50 ) as $rowBatch ) {
112 $dbw->insert( 'job', $rowBatch, __METHOD__
);
114 } catch ( DBError
$e ) {
116 $dbw->rollback( __METHOD__
);
121 $dbw->commit( __METHOD__
);
124 $wgMemc->set( $key, 'false', $ttl ); // queue is not empty
132 * @see JobQueue::doPop()
135 protected function doPop() {
138 if ( $wgMemc->get( $this->getEmptinessCacheKey() ) === 'true' ) {
139 return false; // queue is empty
142 $dbw = $this->getMasterDB();
143 $dbw->commit( __METHOD__
, 'flush' ); // flush existing transaction
145 $uuid = wfRandomString( 32 ); // pop attempt
146 $job = false; // job popped off
147 // Occasionally recycle jobs back into the queue that have been claimed too long
148 if ( mt_rand( 0, 99 ) == 0 ) {
149 $this->recycleStaleJobs();
151 do { // retry when our row is invalid or deleted as a duplicate
152 // Try to reserve a row in the DB...
153 if ( in_array( $this->order
, array( 'fifo', 'timestamp' ) ) ) {
154 $row = $this->claimOldest( $uuid );
155 } else { // random first
156 $rand = mt_rand( 0, self
::MAX_JOB_RANDOM
); // encourage concurrent UPDATEs
157 $gte = (bool)mt_rand( 0, 1 ); // find rows with rand before/after $rand
158 $row = $this->claimRandom( $uuid, $rand, $gte );
159 if ( !$row ) { // need to try the other direction
160 $row = $this->claimRandom( $uuid, $rand, !$gte );
163 // Check if we found a row to reserve...
165 $wgMemc->set( $this->getEmptinessCacheKey(), 'true', self
::CACHE_TTL
);
166 break; // nothing to do
168 // Get the job object from the row...
169 $title = Title
::makeTitleSafe( $row->job_namespace
, $row->job_title
);
171 $dbw->delete( 'job', array( 'job_id' => $row->job_id
), __METHOD__
);
172 wfIncrStats( 'job-pop' );
173 wfDebugLog( 'JobQueueDB', "Row has invalid title '{$row->job_title}'." );
174 continue; // try again
176 $job = Job
::factory( $row->job_cmd
, $title,
177 self
::extractBlob( $row->job_params
), $row->job_id
);
178 $job->id
= $row->job_id
; // XXX: work around broken subclasses
179 // Flag this job as an old duplicate based on its "root" job...
180 if ( $this->isRootJobOldDuplicate( $job ) ) {
181 $job = DuplicateJob
::newFromJob( $job ); // convert to a no-op
190 * Reserve a row with a single UPDATE without holding row locks over RTTs...
192 * @param $uuid string 32 char hex string
193 * @param $rand integer Random unsigned integer (31 bits)
194 * @param $gte bool Search for job_random >= $random (otherwise job_random <= $random)
197 protected function claimRandom( $uuid, $rand, $gte ) {
198 $dbw = $this->getMasterDB();
199 $dir = $gte ?
'ASC' : 'DESC';
200 $ineq = $gte ?
'>=' : '<=';
202 $row = false; // the row acquired
203 // This uses a replication safe method for acquiring jobs. One could use UPDATE+LIMIT
204 // instead, but that either uses ORDER BY (in which case it deadlocks in MySQL) or is
205 // not replication safe. Due to http://bugs.mysql.com/bug.php?id=6980, subqueries cannot
206 // be used here with MySQL.
208 $row = $dbw->selectRow( 'job', '*', // find a random job
210 'job_cmd' => $this->type
,
212 "job_random {$ineq} {$dbw->addQuotes( $rand )}" ),
214 array( 'ORDER BY' => "job_random {$dir}" )
216 if ( $row ) { // claim the job
217 $dbw->update( 'job', // update by PK
219 'job_token' => $uuid,
220 'job_token_timestamp' => $dbw->timestamp(),
221 'job_attempts = job_attempts+1' ),
222 array( 'job_cmd' => $this->type
, 'job_id' => $row->job_id
, 'job_token' => '' ),
225 // This might get raced out by another runner when claiming the previously
226 // selected row. The use of job_random should minimize this problem, however.
227 if ( !$dbw->affectedRows() ) {
228 $row = false; // raced out
231 break; // nothing to do
239 * Reserve a row with a single UPDATE without holding row locks over RTTs...
241 * @param $uuid string 32 char hex string
244 protected function claimOldest( $uuid ) {
245 $dbw = $this->getMasterDB();
247 $row = false; // the row acquired
249 if ( $dbw->getType() === 'mysql' ) {
250 // Per http://bugs.mysql.com/bug.php?id=6980, we can't use subqueries on the
251 // same table being changed in an UPDATE query in MySQL (gives Error: 1093).
252 // Oracle and Postgre have no such limitation. However, MySQL offers an
253 // alternative here by supporting ORDER BY + LIMIT for UPDATE queries.
254 $dbw->query( "UPDATE {$dbw->tableName( 'job' )} " .
256 "job_token = {$dbw->addQuotes( $uuid ) }, " .
257 "job_token_timestamp = {$dbw->addQuotes( $dbw->timestamp() )}, " .
258 "job_attempts = job_attempts+1 " .
260 "job_cmd = {$dbw->addQuotes( $this->type )} " .
261 "AND job_token = {$dbw->addQuotes( '' )} " .
262 ") ORDER BY job_id ASC LIMIT 1",
266 // Use a subquery to find the job, within an UPDATE to claim it.
267 // This uses as much of the DB wrapper functions as possible.
270 'job_token' => $uuid,
271 'job_token_timestamp' => $dbw->timestamp(),
272 'job_attempts = job_attempts+1' ),
273 array( 'job_id = (' .
274 $dbw->selectSQLText( 'job', 'job_id',
275 array( 'job_cmd' => $this->type
, 'job_token' => '' ),
277 array( 'ORDER BY' => 'job_id ASC', 'LIMIT' => 1 ) ) .
283 // Fetch any row that we just reserved...
284 if ( $dbw->affectedRows() ) {
285 $row = $dbw->selectRow( 'job', '*',
286 array( 'job_cmd' => $this->type
, 'job_token' => $uuid ), __METHOD__
288 if ( !$row ) { // raced out by duplicate job removal
289 wfDebugLog( 'JobQueueDB', "Row deleted as duplicate by another process." );
292 break; // nothing to do
300 * Recycle or destroy any jobs that have been claimed for too long
302 * @return integer Number of jobs recycled/deleted
304 protected function recycleStaleJobs() {
306 $dbw = $this->getMasterDB();
307 $count = 0; // affected rows
309 if ( !$dbw->lock( "jobqueue-recycle-{$this->type}", __METHOD__
, 1 ) ) {
310 return $count; // already in progress
313 // Remove claims on jobs acquired for too long if enabled...
314 if ( $this->claimTTL
> 0 ) {
315 $claimCutoff = $dbw->timestamp( $now - $this->claimTTL
);
316 // Get the IDs of jobs that have be claimed but not finished after too long.
317 // These jobs can be recycled into the queue by expiring the claim. Selecting
318 // the IDs first means that the UPDATE can be done by primary key (less deadlocks).
319 $res = $dbw->select( 'job', 'job_id',
321 'job_cmd' => $this->type
,
322 "job_token != {$dbw->addQuotes( '' )}", // was acquired
323 "job_token_timestamp < {$dbw->addQuotes( $claimCutoff )}", // stale
324 "job_attempts < {$dbw->addQuotes( self::MAX_ATTEMPTS )}" ), // retries left
327 $ids = array_map( function( $o ) { return $o->job_id
; }, iterator_to_array( $res ) );
328 if ( count( $ids ) ) {
329 // Reset job_token for these jobs so that other runners will pick them up.
330 // Set the timestamp to the current time, as it is useful to now that the job
331 // was already tried before (the timestamp becomes the "released" time).
335 'job_token_timestamp' => $dbw->timestamp( $now ) ), // time of release
340 $count +
= $dbw->affectedRows();
344 // Just destroy any stale jobs...
345 $pruneCutoff = $dbw->timestamp( $now - self
::MAX_AGE_PRUNE
);
347 'job_cmd' => $this->type
,
348 "job_token != {$dbw->addQuotes( '' )}", // was acquired
349 "job_token_timestamp < {$dbw->addQuotes( $pruneCutoff )}" // stale
351 if ( $this->claimTTL
> 0 ) { // only prune jobs attempted too many times...
352 $conds[] = "job_attempts >= {$dbw->addQuotes( self::MAX_ATTEMPTS )}";
354 // Get the IDs of jobs that are considered stale and should be removed. Selecting
355 // the IDs first means that the UPDATE can be done by primary key (less deadlocks).
356 $res = $dbw->select( 'job', 'job_id', $conds, __METHOD__
);
357 $ids = array_map( function( $o ) { return $o->job_id
; }, iterator_to_array( $res ) );
358 if ( count( $ids ) ) {
359 $dbw->delete( 'job', array( 'job_id' => $ids ), __METHOD__
);
360 $count +
= $dbw->affectedRows();
363 $dbw->unlock( "jobqueue-recycle-{$this->type}", __METHOD__
);
369 * @see JobQueue::doAck()
372 protected function doAck( Job
$job ) {
373 if ( !$job->getId() ) {
374 throw new MWException( "Job of type '{$job->getType()}' has no ID." );
377 $dbw = $this->getMasterDB();
378 $dbw->commit( __METHOD__
, 'flush' ); // flush existing transaction
380 // Delete a row with a single DELETE without holding row locks over RTTs...
382 array( 'job_cmd' => $this->type
, 'job_id' => $job->getId() ), __METHOD__
);
388 * @see JobQueue::doDeduplicateRootJob()
391 protected function doDeduplicateRootJob( Job
$job ) {
392 $params = $job->getParams();
393 if ( !isset( $params['rootJobSignature'] ) ) {
394 throw new MWException( "Cannot register root job; missing 'rootJobSignature'." );
395 } elseif ( !isset( $params['rootJobTimestamp'] ) ) {
396 throw new MWException( "Cannot register root job; missing 'rootJobTimestamp'." );
398 $key = $this->getRootJobCacheKey( $params['rootJobSignature'] );
399 // Callers should call batchInsert() and then this function so that if the insert
400 // fails, the de-duplication registration will be aborted. Since the insert is
401 // deferred till "transaction idle", do that same here, so that the ordering is
402 // maintained. Having only the de-duplication registration succeed would cause
403 // jobs to become no-ops without any actual jobs that made them redundant.
404 $this->getMasterDB()->onTransactionIdle( function() use ( $params, $key ) {
407 $timestamp = $wgMemc->get( $key ); // current last timestamp of this job
408 if ( $timestamp && $timestamp >= $params['rootJobTimestamp'] ) {
409 return true; // a newer version of this root job was enqueued
412 // Update the timestamp of the last root job started at the location...
413 return $wgMemc->set( $key, $params['rootJobTimestamp'], 14*86400 ); // 2 weeks
420 * Check if the "root" job of a given job has been superseded by a newer one
425 protected function isRootJobOldDuplicate( Job
$job ) {
428 $params = $job->getParams();
429 if ( !isset( $params['rootJobSignature'] ) ) {
430 return false; // job has no de-deplication info
431 } elseif ( !isset( $params['rootJobTimestamp'] ) ) {
432 trigger_error( "Cannot check root job; missing 'rootJobTimestamp'." );
436 // Get the last time this root job was enqueued
437 $timestamp = $wgMemc->get( $this->getRootJobCacheKey( $params['rootJobSignature'] ) );
439 // Check if a new root job was started at the location after this one's...
440 return ( $timestamp && $timestamp > $params['rootJobTimestamp'] );
444 * @see JobQueue::doWaitForBackups()
447 protected function doWaitForBackups() {
452 * @return DatabaseBase
454 protected function getSlaveDB() {
455 return wfGetDB( DB_SLAVE
, array(), $this->wiki
);
459 * @return DatabaseBase
461 protected function getMasterDB() {
462 return wfGetDB( DB_MASTER
, array(), $this->wiki
);
469 protected function insertFields( Job
$job ) {
470 $dbw = $this->getMasterDB();
472 // Fields that describe the nature of the job
473 'job_cmd' => $job->getType(),
474 'job_namespace' => $job->getTitle()->getNamespace(),
475 'job_title' => $job->getTitle()->getDBkey(),
476 'job_params' => self
::makeBlob( $job->getParams() ),
477 // Additional job metadata
478 'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
479 'job_timestamp' => $dbw->timestamp(),
480 'job_sha1' => wfBaseConvert(
481 sha1( serialize( $job->getDeduplicationInfo() ) ),
484 'job_random' => mt_rand( 0, self
::MAX_JOB_RANDOM
)
491 private function getEmptinessCacheKey() {
492 list( $db, $prefix ) = wfSplitWikiID( $this->wiki
);
493 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type
, 'isempty' );
497 * @param string $signature Hash identifier of the root job
500 private function getRootJobCacheKey( $signature ) {
501 list( $db, $prefix ) = wfSplitWikiID( $this->wiki
);
502 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type
, 'rootjob', $signature );
509 protected static function makeBlob( $params ) {
510 if ( $params !== false ) {
511 return serialize( $params );
521 protected static function extractBlob( $blob ) {
522 if ( (string)$blob !== '' ) {
523 return unserialize( $blob );