MediaWiki  master
syncFileBackend.php
Go to the documentation of this file.
1 <?php
24 require_once __DIR__ . '/Maintenance.php';
25 
33  public function __construct() {
34  parent::__construct();
35  $this->addDescription( 'Sync one file backend with another using the journal' );
36  $this->addOption( 'src', 'Name of backend to sync from', true, true );
37  $this->addOption( 'dst', 'Name of destination backend to sync', false, true );
38  $this->addOption( 'start', 'Starting journal ID', false, true );
39  $this->addOption( 'end', 'Ending journal ID', false, true );
40  $this->addOption( 'posdir', 'Directory to read/record journal positions', false, true );
41  $this->addOption( 'posdump', 'Just dump current journal position into the position dir.' );
42  $this->addOption( 'postime', 'For position dumps, get the ID at this time', false, true );
43  $this->addOption( 'backoff', 'Stop at entries younger than this age (sec).', false, true );
44  $this->addOption( 'verbose', 'Verbose mode', false, false, 'v' );
45  $this->setBatchSize( 50 );
46  }
47 
48  public function execute() {
49  $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
50 
51  $posDir = $this->getOption( 'posdir' );
52  $posFile = $posDir ? $posDir . '/' . wfWikiID() : false;
53 
54  if ( $this->hasOption( 'posdump' ) ) {
55  // Just dump the current position into the specified position dir
56  if ( !$this->hasOption( 'posdir' ) ) {
57  $this->fatalError( "Param posdir required!" );
58  }
59  if ( $this->hasOption( 'postime' ) ) {
60  $id = (int)$src->getJournal()->getPositionAtTime( $this->getOption( 'postime' ) );
61  $this->output( "Requested journal position is $id.\n" );
62  } else {
63  $id = (int)$src->getJournal()->getCurrentPosition();
64  $this->output( "Current journal position is $id.\n" );
65  }
66  if ( file_put_contents( $posFile, $id, LOCK_EX ) !== false ) {
67  $this->output( "Saved journal position file.\n" );
68  } else {
69  $this->output( "Could not save journal position file.\n" );
70  }
71  if ( $this->isQuiet() ) {
72  print $id; // give a single machine-readable number
73  }
74 
75  return;
76  }
77 
78  if ( !$this->hasOption( 'dst' ) ) {
79  $this->fatalError( "Param dst required!" );
80  }
81  $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
82 
83  $start = $this->getOption( 'start', 0 );
84  if ( !$start && $posFile && is_dir( $posDir ) ) {
85  $start = is_file( $posFile )
86  ? (int)trim( file_get_contents( $posFile ) )
87  : 0;
88  ++$start; // we already did this ID, start with the next one
89  $startFromPosFile = true;
90  } else {
91  $startFromPosFile = false;
92  }
93 
94  if ( $this->hasOption( 'backoff' ) ) {
95  $time = time() - $this->getOption( 'backoff', 0 );
96  $end = (int)$src->getJournal()->getPositionAtTime( $time );
97  } else {
98  $end = $this->getOption( 'end', INF );
99  }
100 
101  $this->output( "Synchronizing backend '{$dst->getName()}' to '{$src->getName()}'...\n" );
102  $this->output( "Starting journal position is $start.\n" );
103  if ( is_finite( $end ) ) {
104  $this->output( "Ending journal position is $end.\n" );
105  }
106 
107  // Periodically update the position file
108  $callback = function ( $pos ) use ( $startFromPosFile, $posFile, $start ) {
109  if ( $startFromPosFile && $pos >= $start ) { // successfully advanced
110  file_put_contents( $posFile, $pos, LOCK_EX );
111  }
112  };
113 
114  // Actually sync the dest backend with the reference backend
115  $lastOKPos = $this->syncBackends( $src, $dst, $start, $end, $callback );
116 
117  // Update the sync position file
118  if ( $startFromPosFile && $lastOKPos >= $start ) { // successfully advanced
119  if ( file_put_contents( $posFile, $lastOKPos, LOCK_EX ) !== false ) {
120  $this->output( "Updated journal position file.\n" );
121  } else {
122  $this->output( "Could not update journal position file.\n" );
123  }
124  }
125 
126  if ( $lastOKPos === false ) {
127  if ( !$start ) {
128  $this->output( "No journal entries found.\n" );
129  } else {
130  $this->output( "No new journal entries found.\n" );
131  }
132  } else {
133  $this->output( "Stopped synchronization at journal position $lastOKPos.\n" );
134  }
135 
136  if ( $this->isQuiet() ) {
137  print $lastOKPos; // give a single machine-readable number
138  }
139  }
140 
152  protected function syncBackends(
153  FileBackend $src, FileBackend $dst, $start, $end, Closure $callback
154  ) {
155  $lastOKPos = 0; // failed
156  $first = true; // first batch
157 
158  if ( $start > $end ) { // sanity
159  $this->fatalError( "Error: given starting ID greater than ending ID." );
160  }
161 
162  $next = null;
163  do {
164  $limit = min( $this->getBatchSize(), $end - $start + 1 ); // don't go pass ending ID
165  $this->output( "Doing id $start to " . ( $start + $limit - 1 ) . "...\n" );
166 
167  $entries = $src->getJournal()->getChangeEntries( $start, $limit, $next );
168  $start = $next; // start where we left off next time
169  if ( $first && !count( $entries ) ) {
170  return false; // nothing to do
171  }
172  $first = false;
173 
174  $lastPosInBatch = 0;
175  $pathsInBatch = []; // changed paths
176  foreach ( $entries as $entry ) {
177  if ( $entry['op'] !== 'null' ) { // null ops are just for reference
178  $pathsInBatch[$entry['path']] = 1; // remove duplicates
179  }
180  $lastPosInBatch = $entry['id'];
181  }
182 
183  $status = $this->syncFileBatch( array_keys( $pathsInBatch ), $src, $dst );
184  if ( $status->isOK() ) {
185  $lastOKPos = max( $lastOKPos, $lastPosInBatch );
186  $callback( $lastOKPos ); // update position file
187  } else {
188  $this->error( print_r( $status->getErrorsArray(), true ) );
189  break; // no gaps; everything up to $lastPos must be OK
190  }
191 
192  if ( !$start ) {
193  $this->output( "End of journal entries.\n" );
194  }
195  } while ( $start && $start <= $end );
196 
197  return $lastOKPos;
198  }
199 
208  protected function syncFileBatch( array $paths, FileBackend $src, FileBackend $dst ) {
210  if ( !count( $paths ) ) {
211  return $status; // nothing to do
212  }
213 
214  // Source: convert internal backend names (FileBackendMultiWrite) to the public one
215  $sPaths = $this->replaceNamePaths( $paths, $src );
216  // Destination: get corresponding path name
217  $dPaths = $this->replaceNamePaths( $paths, $dst );
218 
219  // Lock the live backend paths from modification
220  $sLock = $src->getScopedFileLocks( $sPaths, LockManager::LOCK_UW, $status );
221  $eLock = $dst->getScopedFileLocks( $dPaths, LockManager::LOCK_EX, $status );
222  if ( !$status->isOK() ) {
223  return $status;
224  }
225 
226  $src->preloadFileStat( [ 'srcs' => $sPaths, 'latest' => 1 ] );
227  $dst->preloadFileStat( [ 'srcs' => $dPaths, 'latest' => 1 ] );
228 
229  $ops = [];
230  $fsFiles = [];
231  foreach ( $sPaths as $i => $sPath ) {
232  $dPath = $dPaths[$i]; // destination
233  $sExists = $src->fileExists( [ 'src' => $sPath, 'latest' => 1 ] );
234  if ( $sExists === true ) { // exists in source
235  if ( $this->filesAreSame( $src, $dst, $sPath, $dPath ) ) {
236  continue; // avoid local copies for non-FS backends
237  }
238  // Note: getLocalReference() is fast for FS backends
239  $fsFile = $src->getLocalReference( [ 'src' => $sPath, 'latest' => 1 ] );
240  if ( !$fsFile ) {
241  $this->error( "Unable to sync '$dPath': could not get local copy." );
242  $status->fatal( 'backend-fail-internal', $src->getName() );
243 
244  return $status;
245  }
246  $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
247  // Note: prepare() is usually fast for key/value backends
248  $status->merge( $dst->prepare( [
249  'dir' => dirname( $dPath ), 'bypassReadOnly' => 1 ] ) );
250  if ( !$status->isOK() ) {
251  return $status;
252  }
253  $ops[] = [ 'op' => 'store',
254  'src' => $fsFile->getPath(), 'dst' => $dPath, 'overwrite' => 1 ];
255  } elseif ( $sExists === false ) { // does not exist in source
256  $ops[] = [ 'op' => 'delete', 'src' => $dPath, 'ignoreMissingSource' => 1 ];
257  } else {
258  $this->error( "Unable to sync '$dPath': could not stat file." );
259  $status->fatal( 'backend-fail-internal', $src->getName() );
260 
261  return $status;
262  }
263  }
264 
265  $t_start = microtime( true );
266  $status = $dst->doQuickOperations( $ops, [ 'bypassReadOnly' => 1 ] );
267  if ( !$status->isOK() ) {
268  sleep( 10 ); // wait and retry copy again
269  $status = $dst->doQuickOperations( $ops, [ 'bypassReadOnly' => 1 ] );
270  }
271  $elapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
272  if ( $status->isOK() && $this->getOption( 'verbose' ) ) {
273  $this->output( "Synchronized these file(s) [{$elapsed_ms}ms]:\n" .
274  implode( "\n", $dPaths ) . "\n" );
275  }
276 
277  return $status;
278  }
279 
287  protected function replaceNamePaths( $paths, FileBackend $backend ) {
288  return preg_replace(
289  '!^mwstore://([^/]+)!',
290  StringUtils::escapeRegexReplacement( "mwstore://" . $backend->getName() ),
291  $paths // string or array
292  );
293  }
294 
295  protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
296  return (
297  ( $src->getFileSize( [ 'src' => $sPath ] )
298  === $dst->getFileSize( [ 'src' => $dPath ] ) // short-circuit
299  ) && ( $src->getFileSha1Base36( [ 'src' => $sPath ] )
300  === $dst->getFileSha1Base36( [ 'src' => $dPath ] )
301  )
302  );
303  }
304 }
305 
307 require_once RUN_MAINTENANCE_IF_MAIN;
The wiki should then use memcached to cache various data To use multiple just add more items to the array To increase the weight of a make its entry a array("192.168.0.1:11211", 2))
error( $err, $die=0)
Throw an error to the user.
Apache License January AND DISTRIBUTION Definitions License shall mean the terms and conditions for use
getOption( $name, $default=null)
Get an option, or return the default.
Maintenance script that syncs one file backend to another based on the journal of later...
static escapeRegexReplacement( $string)
Escape a string to make it suitable for inclusion in a preg_replace() replacement parameter...
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
Definition: maintenance.txt:39
getName()
Get the unique backend name.
getScopedFileLocks(array $paths, $type, StatusValue $status, $timeout=0)
Lock the files at the given storage paths in the backend.
setBatchSize( $s=0)
Set the batch size.
hasOption( $name)
Checks to see if a particular option exists.
require_once RUN_MAINTENANCE_IF_MAIN
Definition: maintenance.txt:50
see documentation in includes Linker php for Linker::makeImageLink & $time
Definition: hooks.txt:1799
syncBackends(FileBackend $src, FileBackend $dst, $start, $end, Closure $callback)
Sync $dst backend to $src backend based on the $src logs given after $start.
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 just before the function returns a value If you return true
Definition: hooks.txt:1982
Status::newGood()` to allow deletion, and then `return false` from the hook function. Ensure you consume the 'ChangeTagAfterDelete' hook to carry out custom deletion actions. $tag:name of the tag $user:user initiating the action & $status:Status object. See above. 'ChangeTagsListActive':Allows you to nominate which of the tags your extension uses are in active use. & $tags:list of all active tags. Append to this array. 'ChangeTagsAfterUpdateTags':Called after tags have been updated with the ChangeTags::updateTags function. Params:$addedTags:tags effectively added in the update $removedTags:tags effectively removed in the update $prevTags:tags that were present prior to the update $rc_id:recentchanges table id $rev_id:revision table id $log_id:logging table id $params:tag params $rc:RecentChange being tagged when the tagging accompanies the action, or null $user:User who performed the tagging when the tagging is subsequent to the action, or null 'ChangeTagsAllowedAdd':Called when checking if a user can add tags to a change. & $allowedTags:List of all the tags the user is allowed to add. Any tags the user wants to add( $addTags) that are not in this array will cause it to fail. You may add or remove tags to this array as required. $addTags:List of tags user intends to add. $user:User who is adding the tags. 'ChangeUserGroups':Called before user groups are changed. $performer:The User who will perform the change $user:The User whose groups will be changed & $add:The groups that will be added & $remove:The groups that will be removed 'Collation::factory':Called if $wgCategoryCollation is an unknown collation. $collationName:Name of the collation in question & $collationObject:Null. Replace with a subclass of the Collation class that implements the collation given in $collationName. 'ConfirmEmailComplete':Called after a user 's email has been confirmed successfully. $user:user(object) whose email is being confirmed 'ContentAlterParserOutput':Modify parser output for a given content object. Called by Content::getParserOutput after parsing has finished. Can be used for changes that depend on the result of the parsing but have to be done before LinksUpdate is called(such as adding tracking categories based on the rendered HTML). $content:The Content to render $title:Title of the page, as context $parserOutput:ParserOutput to manipulate 'ContentGetParserOutput':Customize parser output for a given content object, called by AbstractContent::getParserOutput. May be used to override the normal model-specific rendering of page content. $content:The Content to render $title:Title of the page, as context $revId:The revision ID, as context $options:ParserOptions for rendering. To avoid confusing the parser cache, the output can only depend on parameters provided to this hook function, not on global state. $generateHtml:boolean, indicating whether full HTML should be generated. If false, generation of HTML may be skipped, but other information should still be present in the ParserOutput object. & $output:ParserOutput, to manipulate or replace 'ContentHandlerDefaultModelFor':Called when the default content model is determined for a given title. May be used to assign a different model for that title. $title:the Title in question & $model:the model name. Use with CONTENT_MODEL_XXX constants. 'ContentHandlerForModelID':Called when a ContentHandler is requested for a given content model name, but no entry for that model exists in $wgContentHandlers. Note:if your extension implements additional models via this hook, please use GetContentModels hook to make them known to core. $modeName:the requested content model name & $handler:set this to a ContentHandler object, if desired. 'ContentModelCanBeUsedOn':Called to determine whether that content model can be used on a given page. This is especially useful to prevent some content models to be used in some special location. $contentModel:ID of the content model in question $title:the Title in question. & $ok:Output parameter, whether it is OK to use $contentModel on $title. Handler functions that modify $ok should generally return false to prevent further hooks from further modifying $ok. 'ContribsPager::getQueryInfo':Before the contributions query is about to run & $pager:Pager object for contributions & $queryInfo:The query for the contribs Pager 'ContribsPager::reallyDoQuery':Called before really executing the query for My Contributions & $data:an array of results of all contribs queries $pager:The ContribsPager object hooked into $offset:Index offset, inclusive $limit:Exact query limit $descending:Query direction, false for ascending, true for descending 'ContributionsLineEnding':Called before a contributions HTML line is finished $page:SpecialPage object for contributions & $ret:the HTML line $row:the DB row for this line & $classes:the classes to add to the surrounding< li > & $attribs:associative array of other HTML attributes for the< li > element. Currently only data attributes reserved to MediaWiki are allowed(see Sanitizer::isReservedDataAttribute). 'ContributionsToolLinks':Change tool links above Special:Contributions $id:User identifier $title:User page title & $tools:Array of tool links $specialPage:SpecialPage instance for context and services. Can be either SpecialContributions or DeletedContributionsPage. Extensions should type hint against a generic SpecialPage though. 'ConvertContent':Called by AbstractContent::convert when a conversion to another content model is requested. Handler functions that modify $result should generally return false to disable further attempts at conversion. $content:The Content object to be converted. $toModel:The ID of the content model to convert to. $lossy:boolean indicating whether lossy conversion is allowed. & $result:Output parameter, in case the handler function wants to provide a converted Content object. Note that $result->getContentModel() must return $toModel. 'ContentSecurityPolicyDefaultSource':Modify the allowed CSP load sources. This affects all directives except for the script directive. If you want to add a script source, see ContentSecurityPolicyScriptSource hook. & $defaultSrc:Array of Content-Security-Policy allowed sources $policyConfig:Current configuration for the Content-Security-Policy header $mode:ContentSecurityPolicy::REPORT_ONLY_MODE or ContentSecurityPolicy::FULL_MODE depending on type of header 'ContentSecurityPolicyDirectives':Modify the content security policy directives. Use this only if ContentSecurityPolicyDefaultSource and ContentSecurityPolicyScriptSource do not meet your needs. & $directives:Array of CSP directives $policyConfig:Current configuration for the CSP header $mode:ContentSecurityPolicy::REPORT_ONLY_MODE or ContentSecurityPolicy::FULL_MODE depending on type of header 'ContentSecurityPolicyScriptSource':Modify the allowed CSP script sources. Note that you also have to use ContentSecurityPolicyDefaultSource if you want non-script sources to be loaded from whatever you add. & $scriptSrc:Array of CSP directives $policyConfig:Current configuration for the CSP header $mode:ContentSecurityPolicy::REPORT_ONLY_MODE or ContentSecurityPolicy::FULL_MODE depending on type of header 'CustomEditor':When invoking the page editor Return true to allow the normal editor to be used, or false if implementing a custom editor, e.g. for a special namespace, etc. $article:Article being edited $user:User performing the edit 'DeletedContribsPager::reallyDoQuery':Called before really executing the query for Special:DeletedContributions Similar to ContribsPager::reallyDoQuery & $data:an array of results of all contribs queries $pager:The DeletedContribsPager object hooked into $offset:Index offset, inclusive $limit:Exact query limit $descending:Query direction, false for ascending, true for descending 'DeletedContributionsLineEnding':Called before a DeletedContributions HTML line is finished. Similar to ContributionsLineEnding $page:SpecialPage object for DeletedContributions & $ret:the HTML line $row:the DB row for this line & $classes:the classes to add to the surrounding< li > & $attribs:associative array of other HTML attributes for the< li > element. Currently only data attributes reserved to MediaWiki are allowed(see Sanitizer::isReservedDataAttribute). 'DeleteUnknownPreferences':Called by the cleanupPreferences.php maintenance script to build a WHERE clause with which to delete preferences that are not known about. This hook is used by extensions that have dynamically-named preferences that should not be deleted in the usual cleanup process. For example, the Gadgets extension creates preferences prefixed with 'gadget-', and so anything with that prefix is excluded from the deletion. &where:An array that will be passed as the $cond parameter to IDatabase::select() to determine what will be deleted from the user_properties table. $db:The IDatabase object, useful for accessing $db->buildLike() etc. 'DifferenceEngineAfterLoadNewText':called in DifferenceEngine::loadNewText() after the new revision 's content has been loaded into the class member variable $differenceEngine->mNewContent but before returning true from this function. $differenceEngine:DifferenceEngine object 'DifferenceEngineLoadTextAfterNewContentIsLoaded':called in DifferenceEngine::loadText() after the new revision 's content has been loaded into the class member variable $differenceEngine->mNewContent but before checking if the variable 's value is null. This hook can be used to inject content into said class member variable. $differenceEngine:DifferenceEngine object 'DifferenceEngineMarkPatrolledLink':Allows extensions to change the "mark as patrolled" link which is shown both on the diff header as well as on the bottom of a page, usually wrapped in a span element which has class="patrollink". $differenceEngine:DifferenceEngine object & $markAsPatrolledLink:The "mark as patrolled" link HTML(string) $rcid:Recent change ID(rc_id) for this change(int) 'DifferenceEngineMarkPatrolledRCID':Allows extensions to possibly change the rcid parameter. For example the rcid might be set to zero due to the user being the same as the performer of the change but an extension might still want to show it under certain conditions. & $rcid:rc_id(int) of the change or 0 $differenceEngine:DifferenceEngine object $change:RecentChange object $user:User object representing the current user 'DifferenceEngineNewHeader':Allows extensions to change the $newHeader variable, which contains information about the new revision, such as the revision 's author, whether the revision was marked as a minor edit or not, etc. $differenceEngine:DifferenceEngine object & $newHeader:The string containing the various #mw-diff-otitle[1-5] divs, which include things like revision author info, revision comment, RevisionDelete link and more $formattedRevisionTools:Array containing revision tools, some of which may have been injected with the DiffRevisionTools hook $nextlink:String containing the link to the next revision(if any) $status
Definition: hooks.txt:1263
const LOCK_UW
Definition: LockManager.php:68
const LOCK_EX
Definition: LockManager.php:69
static newGood( $value=null)
Factory function for good results.
Definition: StatusValue.php:81
addDescription( $text)
Set the description text.
fileExists(array $params)
Check if a file exists at a storage path in the backend.
getFileSha1Base36(array $params)
Get a SHA-1 hash of the file at a storage path in the backend.
this hook is for auditing only or null if authentication failed before getting that far or null if we can t even determine that When $user is not null
Definition: hooks.txt:780
getFileSize(array $params)
Get the size (bytes) of a file at a storage path in the backend.
wfWikiID()
Get an ASCII string identifying this wiki This is used as a prefix in memcached keys.
output( $out, $channel=null)
Throw some output to the user.
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
doQuickOperations(array $ops, array $opts=[])
Perform a set of independent file operations on some files.
prepare(array $params)
Prepare a storage directory for usage.
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
getJournal()
Get the file journal object for this backend.
you have access to all of the normal MediaWiki so you can get a DB use the etc For full docs on the Maintenance class
Definition: maintenance.txt:52
getBatchSize()
Returns batch size.
Base class for all file backend classes (including multi-write backends).
Definition: FileBackend.php:92
syncFileBatch(array $paths, FileBackend $src, FileBackend $dst)
Sync particular files of backend $src to the corresponding $dst backend files.
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
$maintClass
fatalError( $msg, $exitCode=1)
Output a message and terminate the current script.
preloadFileStat(array $params)
Preload file stat information (concurrently if possible) into in-process cache.
getLocalReference(array $params)
Returns a file system file, identical to the file at a storage path.
filesAreSame(FileBackend $src, FileBackend $dst, $sPath, $dPath)
replaceNamePaths( $paths, FileBackend $backend)
Substitute the backend name of storage paths with that of a given one.