* Base classes for database dumpers
*
* Copyright © 2005 Brion Vibber <brion@pobox.com>
- * http://www.mediawiki.org/
+ * https://www.mediawiki.org/
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* @ingroup Dump Maintenance
*/
-/**
- * @ingroup Dump Maintenance
- */
-class DumpDBZip2Output extends DumpPipeOutput {
- function DumpDBZip2Output( $file ) {
- parent::__construct( "dbzip2", $file );
- }
-}
+require_once __DIR__ . '/Maintenance.php';
+require_once __DIR__ . '/../includes/export/DumpFilter.php';
+
+use Wikimedia\Rdbms\LoadBalancer;
+use Wikimedia\Rdbms\IDatabase;
/**
* @ingroup Dump Maintenance
*/
-class BackupDumper {
- var $reportingInterval = 100;
- var $reporting = true;
- var $pageCount = 0;
- var $revCount = 0;
- var $server = null; // use default
- var $pages = null; // all pages
- var $skipHeader = false; // don't output <mediawiki> and <siteinfo>
- var $skipFooter = false; // don't output </mediawiki>
- var $startId = 0;
- var $endId = 0;
- var $sink = null; // Output filters
- var $stubText = false; // include rev_text_id instead of text; for 2-pass dump
- var $dumpUploads = false;
-
- function BackupDumper( $args ) {
+class BackupDumper extends Maintenance {
+ public $reporting = true;
+ public $pages = null; // all pages
+ public $skipHeader = false; // don't output <mediawiki> and <siteinfo>
+ public $skipFooter = false; // don't output </mediawiki>
+ public $startId = 0;
+ public $endId = 0;
+ public $revStartId = 0;
+ public $revEndId = 0;
+ public $dumpUploads = false;
+ public $dumpUploadFileContents = false;
+ public $orderRevs = false;
+
+ protected $reportingInterval = 100;
+ protected $pageCount = 0;
+ protected $revCount = 0;
+ protected $server = null; // use default
+ protected $sink = null; // Output filters
+ protected $lastTime = 0;
+ protected $pageCountLast = 0;
+ protected $revCountLast = 0;
+
+ protected $outputTypes = [];
+ protected $filterTypes = [];
+
+ protected $ID = 0;
+
+ /**
+ * The dependency-injected database to use.
+ *
+ * @var IDatabase|null
+ *
+ * @see self::setDB
+ */
+ protected $forcedDb = null;
+
+ /** @var LoadBalancer */
+ protected $lb;
+
+ // @todo Unused?
+ private $stubText = false; // include rev_text_id instead of text; for 2-pass dump
+
+ /**
+ * @param array $args For backward compatibility
+ */
+ function __construct( $args = null ) {
+ parent::__construct();
$this->stderr = fopen( "php://stderr", "wt" );
// Built-in output and filter plugins
$this->registerFilter( 'notalk', 'DumpNotalkFilter' );
$this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
- $this->sink = $this->processArgs( $args );
+ // These three can be specified multiple times
+ $this->addOption( 'plugin', 'Load a dump plugin class. Specify as <class>[:<file>].',
+ false, true, false, true );
+ $this->addOption( 'output', 'Begin a filtered output stream; Specify as <type>:<file>. ' .
+ '<type>s: file, gzip, bzip2, 7zip, dbzip2', false, true, false, true );
+ $this->addOption( 'filter', 'Add a filter on an output branch. Specify as ' .
+ '<type>[:<options>]. <types>s: latest, notalk, namespace', false, true, false, true );
+ $this->addOption( 'report', 'Report position and speed after every n pages processed. ' .
+ 'Default: 100.', false, true );
+ $this->addOption( 'server', 'Force reading from MySQL server', false, true );
+ $this->addOption( '7ziplevel', '7zip compression level for all 7zip outputs. Used for ' .
+ '-mx option to 7za command.', false, true );
+
+ if ( $args ) {
+ // Args should be loaded and processed so that dump() can be called directly
+ // instead of execute()
+ $this->loadWithArgv( $args );
+ $this->processOptions();
+ }
}
/**
- * @param $name String
- * @param $class String: name of output filter plugin class
+ * @param string $name
+ * @param string $class Name of output filter plugin class
*/
function registerOutput( $name, $class ) {
$this->outputTypes[$name] = $class;
}
/**
- * @param $name String
- * @param $class String: name of filter plugin class
+ * @param string $name
+ * @param string $class Name of filter plugin class
*/
function registerFilter( $name, $class ) {
$this->filterTypes[$name] = $class;
/**
* Load a plugin and register it
*
- * @param $class String: name of plugin class; must have a static 'register'
- * method that takes a BackupDumper as a parameter.
- * @param $file String: full or relative path to the PHP file to load, or empty
+ * @param string $class Name of plugin class; must have a static 'register'
+ * method that takes a BackupDumper as a parameter.
+ * @param string $file Full or relative path to the PHP file to load, or empty
*/
function loadPlugin( $class, $file ) {
if ( $file != '' ) {
- require_once( $file );
+ require_once $file;
}
- $register = array( $class, 'register' );
- call_user_func_array( $register, array( &$this ) );
+ $register = [ $class, 'register' ];
+ call_user_func_array( $register, [ $this ] );
+ }
+
+ function execute() {
+ throw new MWException( 'execute() must be overridden in subclasses' );
}
/**
- * @param $args Array
- * @return Array
+ * Processes arguments and sets $this->$sink accordingly
*/
- function processArgs( $args ) {
+ function processOptions() {
$sink = null;
- $sinks = array();
- foreach ( $args as $arg ) {
- $matches = array();
- if ( preg_match( '/^--(.+?)(?:=(.+?)(?::(.+?))?)?$/', $arg, $matches ) ) {
- @list( /* $full */ , $opt, $val, $param ) = $matches;
- switch( $opt ) {
- case "plugin":
- $this->loadPlugin( $val, $param );
+ $sinks = [];
+
+ $options = $this->orderedOptions;
+ foreach ( $options as $arg ) {
+ $opt = $arg[0];
+ $param = $arg[1];
+
+ switch ( $opt ) {
+ case 'plugin':
+ $val = explode( ':', $param );
+
+ if ( count( $val ) === 1 ) {
+ $this->loadPlugin( $val[0], '' );
+ } elseif ( count( $val ) === 2 ) {
+ $this->loadPlugin( $val[0], $val[1] );
+ } else {
+ $this->fatalError( 'Invalid plugin parameter' );
+ return;
+ }
+
break;
- case "output":
+ case 'output':
+ $split = explode( ':', $param, 2 );
+ if ( count( $split ) !== 2 ) {
+ $this->fatalError( 'Invalid output parameter' );
+ }
+ list( $type, $file ) = $split;
if ( !is_null( $sink ) ) {
$sinks[] = $sink;
}
- if ( !isset( $this->outputTypes[$val] ) ) {
- wfDie( "Unrecognized output sink type '$val'\n" );
+ if ( !isset( $this->outputTypes[$type] ) ) {
+ $this->fatalError( "Unrecognized output sink type '$type'" );
}
- $type = $this->outputTypes[$val];
- $sink = new $type( $param );
+ $class = $this->outputTypes[$type];
+ if ( $type === "7zip" ) {
+ $sink = new $class( $file, intval( $this->getOption( '7ziplevel' ) ) );
+ } else {
+ $sink = new $class( $file );
+ }
+
break;
- case "filter":
+ case 'filter':
if ( is_null( $sink ) ) {
- $this->progress( "Warning: assuming stdout for filter output\n" );
$sink = new DumpOutput();
}
- if ( !isset( $this->filterTypes[$val] ) ) {
- wfDie( "Unrecognized filter type '$val'\n" );
+
+ $split = explode( ':', $param );
+ $key = $split[0];
+
+ if ( !isset( $this->filterTypes[$key] ) ) {
+ $this->fatalError( "Unrecognized filter type '$key'" );
+ }
+
+ $type = $this->filterTypes[$key];
+
+ if ( count( $split ) === 1 ) {
+ $filter = new $type( $sink );
+ } elseif ( count( $split ) === 2 ) {
+ $filter = new $type( $sink, $split[1] );
+ } else {
+ $this->fatalError( 'Invalid filter parameter' );
}
- $type = $this->filterTypes[$val];
- $filter = new $type( $sink, $param );
// references are lame in php...
unset( $sink );
$sink = $filter;
break;
- case "report":
- $this->reportingInterval = intval( $val );
- break;
- case "server":
- $this->server = $val;
- break;
- case "force-normal":
- if ( !function_exists( 'utf8_normalize' ) ) {
- wfDl( "php_utfnormal.so" );
- if ( !function_exists( 'utf8_normalize' ) ) {
- wfDie( "Failed to load UTF-8 normalization extension. " .
- "Install or remove --force-normal parameter to use slower code.\n" );
- }
- }
- break;
- default:
- $this->processOption( $opt, $val, $param );
- }
}
}
+ if ( $this->hasOption( 'report' ) ) {
+ $this->reportingInterval = intval( $this->getOption( 'report' ) );
+ }
+
+ if ( $this->hasOption( 'server' ) ) {
+ $this->server = $this->getOption( 'server' );
+ }
+
if ( is_null( $sink ) ) {
$sink = new DumpOutput();
}
$sinks[] = $sink;
if ( count( $sinks ) > 1 ) {
- return new DumpMultiWriter( $sinks );
+ $this->sink = new DumpMultiWriter( $sinks );
} else {
- return $sink;
+ $this->sink = $sink;
}
}
- function processOption( $opt, $val, $param ) {
- // extension point for subclasses to add options
- }
-
function dump( $history, $text = WikiExporter::TEXT ) {
# Notice messages will foul up your XML output even if they're
# relatively harmless.
- if ( ini_get( 'display_errors' ) )
+ if ( ini_get( 'display_errors' ) ) {
ini_set( 'display_errors', 'stderr' );
+ }
$this->initProgress( $history );
$db = $this->backupDb();
$exporter = new WikiExporter( $db, $history, WikiExporter::STREAM, $text );
$exporter->dumpUploads = $this->dumpUploads;
+ $exporter->dumpUploadFileContents = $this->dumpUploadFileContents;
$wrapper = new ExportProgressFilter( $this->sink, $this );
$exporter->setOutputSink( $wrapper );
- if ( !$this->skipHeader )
+ if ( !$this->skipHeader ) {
$exporter->openStream();
+ }
# Log item dumps: all or by range
if ( $history & WikiExporter::LOGS ) {
if ( $this->startId || $this->endId ) {
} else {
$exporter->allLogs();
}
- # Page dumps: all or by page ID range
- } else if ( is_null( $this->pages ) ) {
+ } elseif ( is_null( $this->pages ) ) {
+ # Page dumps: all or by page ID range
if ( $this->startId || $this->endId ) {
- $exporter->pagesByRange( $this->startId, $this->endId );
+ $exporter->pagesByRange( $this->startId, $this->endId, $this->orderRevs );
+ } elseif ( $this->revStartId || $this->revEndId ) {
+ $exporter->revsByRange( $this->revStartId, $this->revEndId );
} else {
$exporter->allPages();
}
- # Dump of specific pages
} else {
+ # Dump of specific pages
$exporter->pagesByName( $this->pages );
}
- if ( !$this->skipFooter )
+ if ( !$this->skipFooter ) {
$exporter->closeStream();
+ }
$this->report( true );
}
* Initialise starting time and maximum revision count.
* We'll make ETA calculations based an progress, assuming relatively
* constant per-revision rate.
- * @param $history Integer: WikiExporter::CURRENT or WikiExporter::FULL
+ * @param int $history WikiExporter::CURRENT or WikiExporter::FULL
*/
function initProgress( $history = WikiExporter::FULL ) {
$table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
$field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
- $dbr = wfGetDB( DB_SLAVE );
+ $dbr = $this->forcedDb;
+ if ( $this->forcedDb === null ) {
+ $dbr = wfGetDB( DB_REPLICA );
+ }
$this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
- $this->startTime = wfTime();
+ $this->startTime = microtime( true );
+ $this->lastTime = $this->startTime;
+ $this->ID = getmypid();
}
/**
* @todo Fixme: the --server parameter is currently not respected, as it
* doesn't seem terribly easy to ask the load balancer for a particular
* connection by name.
+ * @return IDatabase
*/
function backupDb() {
+ if ( $this->forcedDb !== null ) {
+ return $this->forcedDb;
+ }
+
$this->lb = wfGetLBFactory()->newMainLB();
- $db = $this->lb->getConnection( DB_SLAVE, 'backup' );
+ $db = $this->lb->getConnection( DB_REPLICA, 'dump' );
// Discourage the server from disconnecting us if it takes a long time
// to read out the big ol' batch query.
- $db->setTimeout( 3600 * 24 );
+ $db->setSessionOptions( [ 'connTimeout' => 3600 * 24 ] );
return $db;
}
+ /**
+ * Force the dump to use the provided database connection for database
+ * operations, wherever possible.
+ *
+ * @param IDatabase|null $db (Optional) the database connection to use. If null, resort to
+ * use the globally provided ways to get database connections.
+ */
+ function setDB( IDatabase $db = null ) {
+ parent::setDB( $db );
+ $this->forcedDb = $db;
+ }
+
function __destruct() {
if ( isset( $this->lb ) ) {
$this->lb->closeAll();
function backupServer() {
global $wgDBserver;
+
return $this->server
? $this->server
: $wgDBserver;
function showReport() {
if ( $this->reporting ) {
- $delta = wfTime() - $this->startTime;
$now = wfTimestamp( TS_DB );
- if ( $delta ) {
- $rate = $this->pageCount / $delta;
- $revrate = $this->revCount / $delta;
+ $nowts = microtime( true );
+ $deltaAll = $nowts - $this->startTime;
+ $deltaPart = $nowts - $this->lastTime;
+ $this->pageCountPart = $this->pageCount - $this->pageCountLast;
+ $this->revCountPart = $this->revCount - $this->revCountLast;
+
+ if ( $deltaAll ) {
$portion = $this->revCount / $this->maxCount;
- $eta = $this->startTime + $delta / $portion;
+ $eta = $this->startTime + $deltaAll / $portion;
$etats = wfTimestamp( TS_DB, intval( $eta ) );
+ $pageRate = $this->pageCount / $deltaAll;
+ $revRate = $this->revCount / $deltaAll;
} else {
- $rate = '-';
- $revrate = '-';
+ $pageRate = '-';
+ $revRate = '-';
$etats = '-';
}
- $this->progress( sprintf( "%s: %s %d pages (%0.3f/sec), %d revs (%0.3f/sec), ETA %s [max %d]",
- $now, wfWikiID(), $this->pageCount, $rate, $this->revCount, $revrate, $etats, $this->maxCount ) );
+ if ( $deltaPart ) {
+ $pageRatePart = $this->pageCountPart / $deltaPart;
+ $revRatePart = $this->revCountPart / $deltaPart;
+ } else {
+ $pageRatePart = '-';
+ $revRatePart = '-';
+ }
+ $this->progress( sprintf(
+ "%s: %s (ID %d) %d pages (%0.1f|%0.1f/sec all|curr), "
+ . "%d revs (%0.1f|%0.1f/sec all|curr), ETA %s [max %d]",
+ $now, wfWikiID(), $this->ID, $this->pageCount, $pageRate,
+ $pageRatePart, $this->revCount, $revRate, $revRatePart, $etats,
+ $this->maxCount
+ ) );
+ $this->lastTime = $nowts;
+ $this->revCountLast = $this->revCount;
}
}
function progress( $string ) {
- fwrite( $this->stderr, $string . "\n" );
+ if ( $this->reporting ) {
+ fwrite( $this->stderr, $string . "\n" );
+ }
+ }
+
+ function fatalError( $msg ) {
+ $this->error( "$msg\n", 1 );
}
}
class ExportProgressFilter extends DumpFilter {
- function ExportProgressFilter( &$sink, &$progress ) {
+ function __construct( &$sink, &$progress ) {
parent::__construct( $sink );
$this->progress = $progress;
}