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
);
60 * @see JobQueue::doBatchPush()
63 protected function doBatchPush( array $jobs, $flags ) {
64 if ( count( $jobs ) ) {
65 $dbw = $this->getMasterDB();
67 $rowSet = array(); // (sha1 => job) map for jobs that are de-duplicated
68 $rowList = array(); // list of jobs for jobs that are are not de-duplicated
70 foreach ( $jobs as $job ) {
71 $row = $this->insertFields( $job );
72 if ( $job->ignoreDuplicates() ) {
73 $rowSet[$row['job_sha1']] = $row;
79 $atomic = ( $flags & self
::QoS_Atomic
);
80 $key = $this->getEmptinessCacheKey();
81 $ttl = self
::CACHE_TTL
;
83 $dbw->onTransactionIdle(
84 function() use ( $dbw, $rowSet, $rowList, $atomic, $key, $ttl
89 $dbw->begin( __METHOD__
); // wrap all the job additions in one transaction
92 // Strip out any duplicate jobs that are already in the queue...
93 if ( count( $rowSet ) ) {
94 $res = $dbw->select( 'job', 'job_sha1',
96 // No job_type condition since it's part of the job_sha1 hash
97 'job_sha1' => array_keys( $rowSet ),
98 'job_token' => '' // unclaimed
102 foreach ( $res as $row ) {
103 wfDebug( "Job with hash '{$row->job_sha1}' is a duplicate." );
104 unset( $rowSet[$row->job_sha1
] ); // already enqueued
107 // Build the full list of job rows to insert
108 $rows = array_merge( $rowList, array_values( $rowSet ) );
109 // Insert the job rows in chunks to avoid slave lag...
110 foreach ( array_chunk( $rows, 50 ) as $rowBatch ) {
111 $dbw->insert( 'job', $rowBatch, __METHOD__
);
113 } catch ( DBError
$e ) {
115 $dbw->rollback( __METHOD__
);
120 $dbw->commit( __METHOD__
);
123 $wgMemc->set( $key, 'false', $ttl ); // queue is not empty
131 * @see JobQueue::doPop()
134 protected function doPop() {
137 if ( $wgMemc->get( $this->getEmptinessCacheKey() ) === 'true' ) {
138 return false; // queue is empty
141 $dbw = $this->getMasterDB();
142 $dbw->commit( __METHOD__
, 'flush' ); // flush existing transaction
144 $uuid = wfRandomString( 32 ); // pop attempt
145 $job = false; // job popped off
146 // Occasionally recycle jobs back into the queue that have been claimed too long
147 if ( mt_rand( 0, 99 ) == 0 ) {
148 $this->recycleStaleJobs();
150 do { // retry when our row is invalid or deleted as a duplicate
151 // Try to reserve a row in the DB...
152 if ( in_array( $this->order
, array( 'fifo', 'timestamp' ) ) ) {
153 $row = $this->claimOldest( $uuid );
154 } else { // random first
155 $rand = mt_rand( 0, self
::MAX_JOB_RANDOM
); // encourage concurrent UPDATEs
156 $gte = (bool)mt_rand( 0, 1 ); // find rows with rand before/after $rand
157 $row = $this->claimRandom( $uuid, $rand, $gte );
158 if ( !$row ) { // need to try the other direction
159 $row = $this->claimRandom( $uuid, $rand, !$gte );
162 // Check if we found a row to reserve...
164 $wgMemc->set( $this->getEmptinessCacheKey(), 'true', self
::CACHE_TTL
);
165 break; // nothing to do
167 // Get the job object from the row...
168 $title = Title
::makeTitleSafe( $row->job_namespace
, $row->job_title
);
170 $dbw->delete( 'job', array( 'job_id' => $row->job_id
), __METHOD__
);
171 wfIncrStats( 'job-pop' );
172 wfDebugLog( 'JobQueueDB', "Row has invalid title '{$row->job_title}'." );
173 continue; // try again
175 $job = Job
::factory( $row->job_cmd
, $title,
176 self
::extractBlob( $row->job_params
), $row->job_id
);
177 $job->id
= $row->job_id
; // XXX: work around broken subclasses
178 // Flag this job as an old duplicate based on its "root" job...
179 if ( $this->isRootJobOldDuplicate( $job ) ) {
180 $job = DuplicateJob
::newFromJob( $job ); // convert to a no-op
189 * Reserve a row with a single UPDATE without holding row locks over RTTs...
191 * @param $uuid string 32 char hex string
192 * @param $rand integer Random unsigned integer (31 bits)
193 * @param $gte bool Search for job_random >= $random (otherwise job_random <= $random)
196 protected function claimRandom( $uuid, $rand, $gte ) {
197 $dbw = $this->getMasterDB();
198 $dir = $gte ?
'ASC' : 'DESC';
199 $ineq = $gte ?
'>=' : '<=';
201 $row = false; // the row acquired
202 // This uses a replication safe method for acquiring jobs. One could use UPDATE+LIMIT
203 // instead, but that either uses ORDER BY (in which case it deadlocks in MySQL) or is
204 // not replication safe. Due to http://bugs.mysql.com/bug.php?id=6980, subqueries cannot
205 // be used here with MySQL.
207 $row = $dbw->selectRow( 'job', '*', // find a random job
209 'job_cmd' => $this->type
,
211 "job_random {$ineq} {$dbw->addQuotes( $rand )}" ),
213 array( 'ORDER BY' => "job_random {$dir}" )
215 if ( $row ) { // claim the job
216 $dbw->update( 'job', // update by PK
218 'job_token' => $uuid,
219 'job_token_timestamp' => $dbw->timestamp(),
220 'job_attempts = job_attempts+1' ),
221 array( 'job_cmd' => $this->type
, 'job_id' => $row->job_id
, 'job_token' => '' ),
224 // This might get raced out by another runner when claiming the previously
225 // selected row. The use of job_random should minimize this problem, however.
226 if ( !$dbw->affectedRows() ) {
227 $row = false; // raced out
230 break; // nothing to do
238 * Reserve a row with a single UPDATE without holding row locks over RTTs...
240 * @param $uuid string 32 char hex string
243 protected function claimOldest( $uuid ) {
244 $dbw = $this->getMasterDB();
246 $row = false; // the row acquired
248 if ( $dbw->getType() === 'mysql' ) {
249 // Per http://bugs.mysql.com/bug.php?id=6980, we can't use subqueries on the
250 // same table being changed in an UPDATE query in MySQL (gives Error: 1093).
251 // Oracle and Postgre have no such limitation. However, MySQL offers an
252 // alternative here by supporting ORDER BY + LIMIT for UPDATE queries.
253 $dbw->query( "UPDATE {$dbw->tableName( 'job' )} " .
255 "job_token = {$dbw->addQuotes( $uuid ) }, " .
256 "job_token_timestamp = {$dbw->addQuotes( $dbw->timestamp() )}, " .
257 "job_attempts = job_attempts+1 " .
259 "job_cmd = {$dbw->addQuotes( $this->type )} " .
260 "AND job_token = {$dbw->addQuotes( '' )} " .
261 ") ORDER BY job_id ASC LIMIT 1",
265 // Use a subquery to find the job, within an UPDATE to claim it.
266 // This uses as much of the DB wrapper functions as possible.
269 'job_token' => $uuid,
270 'job_token_timestamp' => $dbw->timestamp(),
271 'job_attempts = job_attempts+1' ),
272 array( 'job_id = (' .
273 $dbw->selectSQLText( 'job', 'job_id',
274 array( 'job_cmd' => $this->type
, 'job_token' => '' ),
276 array( 'ORDER BY' => 'job_id ASC', 'LIMIT' => 1 ) ) .
282 // Fetch any row that we just reserved...
283 if ( $dbw->affectedRows() ) {
284 $row = $dbw->selectRow( 'job', '*',
285 array( 'job_cmd' => $this->type
, 'job_token' => $uuid ), __METHOD__
287 if ( !$row ) { // raced out by duplicate job removal
288 wfDebugLog( 'JobQueueDB', "Row deleted as duplicate by another process." );
291 break; // nothing to do
299 * Recycle or destroy any jobs that have been claimed for too long
301 * @return integer Number of jobs recycled/deleted
303 protected function recycleStaleJobs() {
305 $dbw = $this->getMasterDB();
306 $count = 0; // affected rows
308 if ( !$dbw->lock( "jobqueue-recycle-{$this->type}", __METHOD__
, 1 ) ) {
309 return $count; // already in progress
312 // Remove claims on jobs acquired for too long if enabled...
313 if ( $this->claimTTL
> 0 ) {
314 $claimCutoff = $dbw->timestamp( $now - $this->claimTTL
);
315 // Get the IDs of jobs that have be claimed but not finished after too long.
316 // These jobs can be recycled into the queue by expiring the claim. Selecting
317 // the IDs first means that the UPDATE can be done by primary key (less deadlocks).
318 $res = $dbw->select( 'job', 'job_id',
320 'job_cmd' => $this->type
,
321 "job_token != {$dbw->addQuotes( '' )}", // was acquired
322 "job_token_timestamp < {$dbw->addQuotes( $claimCutoff )}", // stale
323 "job_attempts < {$dbw->addQuotes( self::MAX_ATTEMPTS )}" ), // retries left
326 $ids = array_map( function( $o ) { return $o->job_id
; }, iterator_to_array( $res ) );
327 if ( count( $ids ) ) {
328 // Reset job_token for these jobs so that other runners will pick them up.
329 // Set the timestamp to the current time, as it is useful to now that the job
330 // was already tried before (the timestamp becomes the "released" time).
334 'job_token_timestamp' => $dbw->timestamp( $now ) ), // time of release
339 $count +
= $dbw->affectedRows();
343 // Just destroy any stale jobs...
344 $pruneCutoff = $dbw->timestamp( $now - self
::MAX_AGE_PRUNE
);
346 'job_cmd' => $this->type
,
347 "job_token != {$dbw->addQuotes( '' )}", // was acquired
348 "job_token_timestamp < {$dbw->addQuotes( $pruneCutoff )}" // stale
350 if ( $this->claimTTL
> 0 ) { // only prune jobs attempted too many times...
351 $conds[] = "job_attempts >= {$dbw->addQuotes( self::MAX_ATTEMPTS )}";
353 // Get the IDs of jobs that are considered stale and should be removed. Selecting
354 // the IDs first means that the UPDATE can be done by primary key (less deadlocks).
355 $res = $dbw->select( 'job', 'job_id', $conds, __METHOD__
);
356 $ids = array_map( function( $o ) { return $o->job_id
; }, iterator_to_array( $res ) );
357 if ( count( $ids ) ) {
358 $dbw->delete( 'job', array( 'job_id' => $ids ), __METHOD__
);
359 $count +
= $dbw->affectedRows();
362 $dbw->unlock( "jobqueue-recycle-{$this->type}", __METHOD__
);
368 * @see JobQueue::doAck()
371 protected function doAck( Job
$job ) {
372 if ( !$job->getId() ) {
373 throw new MWException( "Job of type '{$job->getType()}' has no ID." );
376 $dbw = $this->getMasterDB();
377 $dbw->commit( __METHOD__
, 'flush' ); // flush existing transaction
379 // Delete a row with a single DELETE without holding row locks over RTTs...
381 array( 'job_cmd' => $this->type
, 'job_id' => $job->getId() ), __METHOD__
);
387 * @see JobQueue::doDeduplicateRootJob()
390 protected function doDeduplicateRootJob( Job
$job ) {
391 $params = $job->getParams();
392 if ( !isset( $params['rootJobSignature'] ) ) {
393 throw new MWException( "Cannot register root job; missing 'rootJobSignature'." );
394 } elseif ( !isset( $params['rootJobTimestamp'] ) ) {
395 throw new MWException( "Cannot register root job; missing 'rootJobTimestamp'." );
397 $key = $this->getRootJobCacheKey( $params['rootJobSignature'] );
398 // Callers should call batchInsert() and then this function so that if the insert
399 // fails, the de-duplication registration will be aborted. Since the insert is
400 // deferred till "transaction idle", do that same here, so that the ordering is
401 // maintained. Having only the de-duplication registration succeed would cause
402 // jobs to become no-ops without any actual jobs that made them redundant.
403 $this->getMasterDB()->onTransactionIdle( function() use ( $params, $key ) {
406 $timestamp = $wgMemc->get( $key ); // current last timestamp of this job
407 if ( $timestamp && $timestamp >= $params['rootJobTimestamp'] ) {
408 return true; // a newer version of this root job was enqueued
411 // Update the timestamp of the last root job started at the location...
412 return $wgMemc->set( $key, $params['rootJobTimestamp'], 14*86400 ); // 2 weeks
419 * Check if the "root" job of a given job has been superseded by a newer one
424 protected function isRootJobOldDuplicate( Job
$job ) {
427 $params = $job->getParams();
428 if ( !isset( $params['rootJobSignature'] ) ) {
429 return false; // job has no de-deplication info
430 } elseif ( !isset( $params['rootJobTimestamp'] ) ) {
431 trigger_error( "Cannot check root job; missing 'rootJobTimestamp'." );
435 // Get the last time this root job was enqueued
436 $timestamp = $wgMemc->get( $this->getRootJobCacheKey( $params['rootJobSignature'] ) );
438 // Check if a new root job was started at the location after this one's...
439 return ( $timestamp && $timestamp > $params['rootJobTimestamp'] );
443 * @see JobQueue::doWaitForBackups()
446 protected function doWaitForBackups() {
451 * @return DatabaseBase
453 protected function getSlaveDB() {
454 return wfGetDB( DB_SLAVE
, array(), $this->wiki
);
458 * @return DatabaseBase
460 protected function getMasterDB() {
461 return wfGetDB( DB_MASTER
, array(), $this->wiki
);
468 protected function insertFields( Job
$job ) {
469 $dbw = $this->getMasterDB();
471 // Fields that describe the nature of the job
472 'job_cmd' => $job->getType(),
473 'job_namespace' => $job->getTitle()->getNamespace(),
474 'job_title' => $job->getTitle()->getDBkey(),
475 'job_params' => self
::makeBlob( $job->getParams() ),
476 // Additional job metadata
477 'job_id' => $dbw->nextSequenceValue( 'job_job_id_seq' ),
478 'job_timestamp' => $dbw->timestamp(),
479 'job_sha1' => wfBaseConvert(
480 sha1( serialize( $job->getDeduplicationInfo() ) ),
483 'job_random' => mt_rand( 0, self
::MAX_JOB_RANDOM
)
490 private function getEmptinessCacheKey() {
491 list( $db, $prefix ) = wfSplitWikiID( $this->wiki
);
492 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type
, 'isempty' );
496 * @param string $signature Hash identifier of the root job
499 private function getRootJobCacheKey( $signature ) {
500 list( $db, $prefix ) = wfSplitWikiID( $this->wiki
);
501 return wfForeignMemcKey( $db, $prefix, 'jobqueue', $this->type
, 'rootjob', $signature );
508 protected static function makeBlob( $params ) {
509 if ( $params !== false ) {
510 return serialize( $params );
520 protected static function extractBlob( $blob ) {
521 if ( (string)$blob !== '' ) {
522 return unserialize( $blob );