MediaWiki  1.28.0
ExternalStoreDB.php
Go to the documentation of this file.
1 <?php
40  public function fetchFromURL( $url ) {
41  list( $cluster, $id, $itemID ) = $this->parseURL( $url );
42  $ret = $this->fetchBlob( $cluster, $id, $itemID );
43 
44  if ( $itemID !== false && $ret !== false ) {
45  return $ret->getItem( $itemID );
46  }
47 
48  return $ret;
49  }
50 
60  public function batchFetchFromURLs( array $urls ) {
61  $batched = $inverseUrlMap = [];
62  foreach ( $urls as $url ) {
63  list( $cluster, $id, $itemID ) = $this->parseURL( $url );
64  $batched[$cluster][$id][] = $itemID;
65  // false $itemID gets cast to int, but should be ok
66  // since we do === from the $itemID in $batched
67  $inverseUrlMap[$cluster][$id][$itemID] = $url;
68  }
69  $ret = [];
70  foreach ( $batched as $cluster => $batchByCluster ) {
71  $res = $this->batchFetchBlobs( $cluster, $batchByCluster );
73  foreach ( $res as $id => $blob ) {
74  foreach ( $batchByCluster[$id] as $itemID ) {
75  $url = $inverseUrlMap[$cluster][$id][$itemID];
76  if ( $itemID === false ) {
77  $ret[$url] = $blob;
78  } else {
79  $ret[$url] = $blob->getItem( $itemID );
80  }
81  }
82  }
83  }
84 
85  return $ret;
86  }
87 
88  public function store( $location, $data ) {
89  $dbw = $this->getMaster( $location );
90  $id = $dbw->nextSequenceValue( 'blob_blob_id_seq' );
91  $dbw->insert( $this->getTable( $dbw ),
92  [ 'blob_id' => $id, 'blob_text' => $data ],
93  __METHOD__ );
94  $id = $dbw->insertId();
95  if ( !$id ) {
96  throw new MWException( __METHOD__ . ': no insert ID' );
97  }
98 
99  return "DB://$location/$id";
100  }
101 
108  function getLoadBalancer( $cluster ) {
109  $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
110 
111  return wfGetLBFactory()->getExternalLB( $cluster, $wiki );
112  }
113 
120  function getSlave( $cluster ) {
122 
123  $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
124  $lb = $this->getLoadBalancer( $cluster );
125 
126  if ( !in_array( "DB://" . $cluster, (array)$wgDefaultExternalStore ) ) {
127  wfDebug( "read only external store\n" );
128  $lb->allowLagged( true );
129  } else {
130  wfDebug( "writable external store\n" );
131  }
132 
133  $db = $lb->getConnectionRef( DB_REPLICA, [], $wiki );
134  $db->clearFlag( DBO_TRX ); // sanity
135 
136  return $db;
137  }
138 
145  function getMaster( $cluster ) {
146  $wiki = isset( $this->params['wiki'] ) ? $this->params['wiki'] : false;
147  $lb = $this->getLoadBalancer( $cluster );
148 
149  $db = $lb->getConnectionRef( DB_MASTER, [], $wiki );
150  $db->clearFlag( DBO_TRX ); // sanity
151 
152  return $db;
153  }
154 
161  function getTable( $db ) {
162  $table = $db->getLBInfo( 'blobs table' );
163  if ( is_null( $table ) ) {
164  $table = 'blobs';
165  }
166 
167  return $table;
168  }
169 
180  function fetchBlob( $cluster, $id, $itemID ) {
187  static $externalBlobCache = [];
188 
189  $cacheID = ( $itemID === false ) ? "$cluster/$id" : "$cluster/$id/";
190  if ( isset( $externalBlobCache[$cacheID] ) ) {
191  wfDebugLog( 'ExternalStoreDB-cache',
192  "ExternalStoreDB::fetchBlob cache hit on $cacheID" );
193 
194  return $externalBlobCache[$cacheID];
195  }
196 
197  wfDebugLog( 'ExternalStoreDB-cache',
198  "ExternalStoreDB::fetchBlob cache miss on $cacheID" );
199 
200  $dbr = $this->getSlave( $cluster );
201  $ret = $dbr->selectField( $this->getTable( $dbr ),
202  'blob_text', [ 'blob_id' => $id ], __METHOD__ );
203  if ( $ret === false ) {
204  wfDebugLog( 'ExternalStoreDB',
205  "ExternalStoreDB::fetchBlob master fallback on $cacheID" );
206  // Try the master
207  $dbw = $this->getMaster( $cluster );
208  $ret = $dbw->selectField( $this->getTable( $dbw ),
209  'blob_text', [ 'blob_id' => $id ], __METHOD__ );
210  if ( $ret === false ) {
211  wfDebugLog( 'ExternalStoreDB',
212  "ExternalStoreDB::fetchBlob master failed to find $cacheID" );
213  }
214  }
215  if ( $itemID !== false && $ret !== false ) {
216  // Unserialise object; caller extracts item
217  $ret = unserialize( $ret );
218  }
219 
220  $externalBlobCache = [ $cacheID => $ret ];
221 
222  return $ret;
223  }
224 
233  function batchFetchBlobs( $cluster, array $ids ) {
234  $dbr = $this->getSlave( $cluster );
235  $res = $dbr->select( $this->getTable( $dbr ),
236  [ 'blob_id', 'blob_text' ], [ 'blob_id' => array_keys( $ids ) ], __METHOD__ );
237  $ret = [];
238  if ( $res !== false ) {
239  $this->mergeBatchResult( $ret, $ids, $res );
240  }
241  if ( $ids ) {
242  wfDebugLog( __CLASS__, __METHOD__ .
243  " master fallback on '$cluster' for: " .
244  implode( ',', array_keys( $ids ) ) );
245  // Try the master
246  $dbw = $this->getMaster( $cluster );
247  $res = $dbw->select( $this->getTable( $dbr ),
248  [ 'blob_id', 'blob_text' ],
249  [ 'blob_id' => array_keys( $ids ) ],
250  __METHOD__ );
251  if ( $res === false ) {
252  wfDebugLog( __CLASS__, __METHOD__ . " master failed on '$cluster'" );
253  } else {
254  $this->mergeBatchResult( $ret, $ids, $res );
255  }
256  }
257  if ( $ids ) {
258  wfDebugLog( __CLASS__, __METHOD__ .
259  " master on '$cluster' failed locating items: " .
260  implode( ',', array_keys( $ids ) ) );
261  }
262 
263  return $ret;
264  }
265 
272  private function mergeBatchResult( array &$ret, array &$ids, $res ) {
273  foreach ( $res as $row ) {
274  $id = $row->blob_id;
275  $itemIDs = $ids[$id];
276  unset( $ids[$id] ); // to track if everything is found
277  if ( count( $itemIDs ) === 1 && reset( $itemIDs ) === false ) {
278  // single result stored per blob
279  $ret[$id] = $row->blob_text;
280  } else {
281  // multi result stored per blob
282  $ret[$id] = unserialize( $row->blob_text );
283  }
284  }
285  }
286 
291  protected function parseURL( $url ) {
292  $path = explode( '/', $url );
293 
294  return [
295  $path[2], // cluster
296  $path[3], // id
297  isset( $path[4] ) ? $path[4] : false // itemID
298  ];
299  }
300 }
fetchBlob($cluster, $id, $itemID)
Fetch a blob item out of the database; a cache of the last-loaded blob will be kept so that multiple ...
deferred txt A few of the database updates required by various functions here can be deferred until after the result page is displayed to the user For updating the view updating the linked to tables after a etc PHP does not yet have any way to tell the server to actually return and disconnect while still running these but it might have such a feature in the future We handle these by creating a deferred update object and putting those objects on a global list
Definition: deferred.txt:11
fetchFromURL($url)
The provided URL is in the form of DB://cluster/id or DB://cluster/id/itemid for concatened storage...
the array() calling protocol came about after MediaWiki 1.4rc1.
store($location, $data)
array $wgDefaultExternalStore
The place to put new revisions, false to put them in the local text table.
Accessable external objects in a particular storage medium.
processing should stop and the error should be shown to the user * false
Definition: hooks.txt:189
null means default in associative array with keys and values unescaped Should be merged with default with a value of false meaning to suppress the attribute in associative array with keys and values unescaped noclasses & $ret
Definition: hooks.txt:1936
when a variable name is used in a it is silently declared as a new local masking the global
Definition: design.txt:93
const DB_MASTER
Definition: defines.php:23
batchFetchFromURLs(array $urls)
Fetch data from given external store URLs.
wfDebug($text, $dest= 'all', array $context=[])
Sends a line to the debug log if enabled or, optionally, to a comment in output.
wfDebugLog($logGroup, $text, $dest= 'all', array $context=[])
Send a line to a supplementary debug log file, if configured, or main debug log if not...
unserialize($serialized)
Definition: ApiMessage.php:102
$res
Definition: database.txt:21
mergeBatchResult(array &$ret, array &$ids, $res)
Helper function for self::batchFetchBlobs for merging master/replica DB results.
This document is intended to provide useful advice for parties seeking to redistribute MediaWiki to end users It s targeted particularly at maintainers for Linux since it s been observed that distribution packages of MediaWiki often break We ve consistently had to recommend that users seeking support use official tarballs instead of their distribution s and this often solves whatever problem the user is having It would be nice if this could such as
Definition: distributors.txt:9
const DBO_TRX
Definition: defines.php:9
getTable($db)
Get the 'blobs' table name for this database.
injection txt This is an overview of how MediaWiki makes use of dependency injection The design described here grew from the discussion of RFC T384 The term dependency this means that anything an object needs to operate should be injected from the the object itself should only know narrow no concrete implementation of the logic it relies on The requirement to inject everything typically results in an architecture that based on two main types of and essentially stateless service objects that use other service objects to operate on the value objects As of the beginning MediaWiki is only starting to use the DI approach Much of the code still relies on global state or direct resulting in a highly cyclical dependency which acts as the top level factory for services in MediaWiki which can be used to gain access to default instances of various services MediaWikiServices however also allows new services to be defined and default services to be redefined Services are defined or redefined by providing a callback the instantiator that will return a new instance of the service When it will create an instance of MediaWikiServices and populate it with the services defined in the files listed by thereby bootstrapping the DI framework Per $wgServiceWiringFiles lists includes ServiceWiring php
Definition: injection.txt:35
wfGetLBFactory()
Get the load balancer factory object.
const DB_REPLICA
Definition: defines.php:22
batchFetchBlobs($cluster, array $ids)
Fetch multiple blob items out of the database.
getSlave($cluster)
Get a replica DB connection for the specified cluster.
DB accessable external objects.
getLoadBalancer($cluster)
Get a LoadBalancer for the specified cluster.
getMaster($cluster)
Get a master database connection for the specified cluster.