backup.inc

Go to the documentation of this file.

<?php
require_once __DIR__ . '/Maintenance.php';
require_once __DIR__ . '/../includes/export/DumpFilter.php';
 
use Wikimedia\Rdbms\LoadBalancer;
use Wikimedia\Rdbms\IDatabase;
 
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;
 
    protected $forcedDb = null;
 
    protected $lb;
 
    // @todo Unused?
    private $stubText = false; // include rev_text_id instead of text; for 2-pass dump
 
    function __construct( $args = null ) {
        parent::__construct();
        $this->stderr = fopen( "php://stderr", "wt" );
 
        // Built-in output and filter plugins
        $this->registerOutput( 'file', 'DumpFileOutput' );
        $this->registerOutput( 'gzip', 'DumpGZipOutput' );
        $this->registerOutput( 'bzip2', 'DumpBZip2Output' );
        $this->registerOutput( 'dbzip2', 'DumpDBZip2Output' );
        $this->registerOutput( '7zip', 'Dump7ZipOutput' );
 
        $this->registerFilter( 'latest', 'DumpLatestFilter' );
        $this->registerFilter( 'notalk', 'DumpNotalkFilter' );
        $this->registerFilter( 'namespace', 'DumpNamespaceFilter' );
 
        // 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();
        }
    }
 
    function registerOutput( $name, $class ) {
        $this->outputTypes[$name] = $class;
    }
 
    function registerFilter( $name, $class ) {
        $this->filterTypes[$name] = $class;
    }
 
    function loadPlugin( $class, $file ) {
        if ( $file != '' ) {
            require_once $file;
        }
        $register = [ $class, 'register' ];
        call_user_func_array( $register, [ $this ] );
    }
 
    function execute() {
        throw new MWException( 'execute() must be overridden in subclasses' );
    }
 
    function processOptions() {
        $sink = null;
        $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':
                    $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( $this->outputTypes[$type] ) ) {
                        $this->fatalError( "Unrecognized output sink type '$type'" );
                    }
                    $class = $this->outputTypes[$type];
                    if ( $type === "7zip" ) {
                        $sink = new $class( $file, intval( $this->getOption( '7ziplevel' ) ) );
                    } else {
                        $sink = new $class( $file );
                    }
 
                    break;
                case 'filter':
                    if ( is_null( $sink ) ) {
                        $sink = new DumpOutput();
                    }
 
                    $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' );
                    }
 
                    // references are lame in php...
                    unset( $sink );
                    $sink = $filter;
 
                    break;
            }
        }
 
        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 ) {
            $this->sink = new DumpMultiWriter( $sinks );
        } else {
            $this->sink = $sink;
        }
    }
 
    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' ) ) {
            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 ) {
            $exporter->openStream();
        }
        # Log item dumps: all or by range
        if ( $history & WikiExporter::LOGS ) {
            if ( $this->startId || $this->endId ) {
                $exporter->logsByRange( $this->startId, $this->endId );
            } else {
                $exporter->allLogs();
            }
        } elseif ( is_null( $this->pages ) ) {
            # Page dumps: all or by page ID range
            if ( $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();
            }
        } else {
            # Dump of specific pages
            $exporter->pagesByName( $this->pages );
        }
 
        if ( !$this->skipFooter ) {
            $exporter->closeStream();
        }
 
        $this->report( true );
    }
 
    function initProgress( $history = WikiExporter::FULL ) {
        $table = ( $history == WikiExporter::CURRENT ) ? 'page' : 'revision';
        $field = ( $history == WikiExporter::CURRENT ) ? 'page_id' : 'rev_id';
 
        $dbr = $this->forcedDb;
        if ( $this->forcedDb === null ) {
            $dbr = wfGetDB( DB_REPLICA );
        }
        $this->maxCount = $dbr->selectField( $table, "MAX($field)", '', __METHOD__ );
        $this->startTime = microtime( true );
        $this->lastTime = $this->startTime;
        $this->ID = getmypid();
    }
 
    function backupDb() {
        if ( $this->forcedDb !== null ) {
            return $this->forcedDb;
        }
 
        $this->lb = wfGetLBFactory()->newMainLB();
        $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->setSessionOptions( [ 'connTimeout' => 3600 * 24 ] );
 
        return $db;
    }
 
    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 reportPage() {
        $this->pageCount++;
    }
 
    function revCount() {
        $this->revCount++;
        $this->report();
    }
 
    function report( $final = false ) {
        if ( $final xor ( $this->revCount % $this->reportingInterval == 0 ) ) {
            $this->showReport();
        }
    }
 
    function showReport() {
        if ( $this->reporting ) {
            $now = wfTimestamp( TS_DB );
            $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 + $deltaAll / $portion;
                $etats = wfTimestamp( TS_DB, intval( $eta ) );
                $pageRate = $this->pageCount / $deltaAll;
                $revRate = $this->revCount / $deltaAll;
            } else {
                $pageRate = '-';
                $revRate = '-';
                $etats = '-';
            }
            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 ) {
        if ( $this->reporting ) {
            fwrite( $this->stderr, $string . "\n" );
        }
    }
 
    function fatalError( $msg ) {
        $this->error( "$msg\n", 1 );
    }
}
 
class ExportProgressFilter extends DumpFilter {
    function __construct( &$sink, &$progress ) {
        parent::__construct( $sink );
        $this->progress = $progress;
    }
 
    function writeClosePage( $string ) {
        parent::writeClosePage( $string );
        $this->progress->reportPage();
    }
 
    function writeRevision( $rev, $string ) {
        parent::writeRevision( $rev, $string );
        $this->progress->revCount();
    }
}