maintenance/purgeChangedFiles.php

   1 <?php
   2 /**
   3  * Scan the logging table and purge affected files within a timeframe.
   4  *
   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.
   9  *
  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.
  14  *
  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
  19  *
  20  * @file
  21  * @ingroup Maintenance
  22  */
  23
  24 require_once __DIR__ . '/Maintenance.php';
  25
  26 /**
  27  * Maintenance script that scans the deletion log and purges affected files
  28  * within a timeframe.
  29  *
  30  * @ingroup Maintenance
  31  */
  32 class PurgeChangedFiles extends Maintenance {
  33         /**
  34          * Mapping from type option to log type and actions.
  35          * @var array
  36          */
  37         private static $typeMappings = [
  38                 'created' => [
  39                         'upload' => [ 'upload' ],
  40                         'import' => [ 'upload', 'interwiki' ],
  41                 ],
  42                 'deleted' => [
  43                         'delete' => [ 'delete', 'revision' ],
  44                         'suppress' => [ 'delete', 'revision' ],
  45                 ],
  46                 'modified' => [
  47                         'upload' => [ 'overwrite', 'revert' ],
  48                         'move' => [ 'move', 'move_redir' ],
  49                 ],
  50         ];
  51
  52         /**
  53          * @var string
  54          */
  55         private $startTimestamp;
  56
  57         /**
  58          * @var string
  59          */
  60         private $endTimestamp;
  61
  62         public function __construct() {
  63                 parent::__construct();
  64                 $this->addDescription( 'Scan the logging table and purge files and thumbnails.' );
  65                 $this->addOption( 'starttime', 'Starting timestamp', true, true );
  66                 $this->addOption( 'endtime', 'Ending timestamp', true, true );
  67                 $this->addOption( 'type', 'Comma-separated list of types of changes to send purges for (' .
  68                         implode( ',', array_keys( self::$typeMappings ) ) . ',all)', false, true );
  69                 $this->addOption( 'htcp-dest', 'HTCP announcement destination (IP:port)', false, true );
  70                 $this->addOption( 'dry-run', 'Do not send purge requests' );
  71                 $this->addOption( 'sleep-per-batch', 'Milliseconds to sleep between batches', false, true );
  72                 $this->addOption( 'verbose', 'Show more output', false, false, 'v' );
  73                 $this->setBatchSize( 100 );
  74         }
  75
  76         public function execute() {
  77                 global $wgHTCPRouting;
  78
  79                 if ( $this->hasOption( 'htcp-dest' ) ) {
  80                         $parts = explode( ':', $this->getOption( 'htcp-dest' ) );
  81                         if ( count( $parts ) < 2 ) {
  82                                 // Add default htcp port
  83                                 $parts[] = '4827';
  84                         }
  85
  86                         // Route all HTCP messages to provided host:port
  87                         $wgHTCPRouting = [
  88                                 '' => [ 'host' => $parts[0], 'port' => $parts[1] ],
  89                         ];
  90                         $this->verbose( "HTCP broadcasts to {$parts[0]}:{$parts[1]}\n" );
  91                 }
  92
  93                 // Find out which actions we should be concerned with
  94                 $typeOpt = $this->getOption( 'type', 'all' );
  95                 $validTypes = array_keys( self::$typeMappings );
  96                 if ( $typeOpt === 'all' ) {
  97                         // Convert 'all' to all registered types
  98                         $typeOpt = implode( ',', $validTypes );
  99                 }
 100                 $typeList = explode( ',', $typeOpt );
 101                 foreach ( $typeList as $type ) {
 102                         if ( !in_array( $type, $validTypes ) ) {
 103                                 $this->error( "\nERROR: Unknown type: {$type}\n" );
 104                                 $this->maybeHelp( true );
 105                         }
 106                 }
 107
 108                 // Validate the timestamps
 109                 $dbr = $this->getDB( DB_REPLICA );
 110                 $this->startTimestamp = $dbr->timestamp( $this->getOption( 'starttime' ) );
 111                 $this->endTimestamp = $dbr->timestamp( $this->getOption( 'endtime' ) );
 112
 113                 if ( $this->startTimestamp > $this->endTimestamp ) {
 114                         $this->error( "\nERROR: starttime after endtime\n" );
 115                         $this->maybeHelp( true );
 116                 }
 117
 118                 // Turn on verbose when dry-run is enabled
 119                 if ( $this->hasOption( 'dry-run' ) ) {
 120                         $this->mOptions['verbose'] = 1;
 121                 }
 122
 123                 $this->verbose( 'Purging files that were: ' . implode( ', ', $typeList ) . "\n" );
 124                 foreach ( $typeList as $type ) {
 125                         $this->verbose( "Checking for {$type} files...\n" );
 126                         $this->purgeFromLogType( $type );
 127                         if ( !$this->hasOption( 'dry-run' ) ) {
 128                                 $this->verbose( "...{$type} files purged.\n\n" );
 129                         }
 130                 }
 131         }
 132
 133         /**
 134          * Purge cache and thumbnails for changes of the given type.
 135          *
 136          * @param string $type Type of change to find
 137          */
 138         protected function purgeFromLogType( $type ) {
 139                 $repo = RepoGroup::singleton()->getLocalRepo();
 140                 $dbr = $this->getDB( DB_REPLICA );
 141
 142                 foreach ( self::$typeMappings[$type] as $logType => $logActions ) {
 143                         $this->verbose( "Scanning for {$logType}/" . implode( ',', $logActions ) . "\n" );
 144
 145                         $res = $dbr->select(
 146                                 'logging',
 147                                 [ 'log_title', 'log_timestamp', 'log_params' ],
 148                                 [
 149                                         'log_namespace' => NS_FILE,
 150                                         'log_type' => $logType,
 151                                         'log_action' => $logActions,
 152                                         'log_timestamp >= ' . $dbr->addQuotes( $this->startTimestamp ),
 153                                         'log_timestamp <= ' . $dbr->addQuotes( $this->endTimestamp ),
 154                                 ],
 155                                 __METHOD__
 156                         );
 157
 158                         $bSize = 0;
 159                         foreach ( $res as $row ) {
 160                                 $file = $repo->newFile( Title::makeTitle( NS_FILE, $row->log_title ) );
 161
 162                                 if ( $this->hasOption( 'dry-run' ) ) {
 163                                         $this->verbose( "{$type}[{$row->log_timestamp}]: {$row->log_title}\n" );
 164                                         continue;
 165                                 }
 166
 167                                 // Purge current version and its thumbnails
 168                                 $file->purgeCache();
 169                                 // Purge the old versions and their thumbnails
 170                                 foreach ( $file->getHistory() as $oldFile ) {
 171                                         $oldFile->purgeCache();
 172                                 }
 173
 174                                 if ( $logType === 'delete' ) {
 175                                         // If there is an orphaned storage file... delete it
 176                                         if ( !$file->exists() && $repo->fileExists( $file->getPath() ) ) {
 177                                                 $dpath = $this->getDeletedPath( $repo, $file );
 178                                                 if ( $repo->fileExists( $dpath ) ) {
 179                                                         // Sanity check to avoid data loss
 180                                                         $repo->getBackend()->delete( [ 'src' => $file->getPath() ] );
 181                                                         $this->verbose( "Deleted orphan file: {$file->getPath()}.\n" );
 182                                                 } else {
 183                                                         $this->error( "File was not deleted: {$file->getPath()}.\n" );
 184                                                 }
 185                                         }
 186
 187                                         // Purge items from fileachive table (rows are likely here)
 188                                         $this->purgeFromArchiveTable( $repo, $file );
 189                                 } elseif ( $logType === 'move' ) {
 190                                         // Purge the target file as well
 191
 192                                         $params = unserialize( $row->log_params );
 193                                         if ( isset( $params['4::target'] ) ) {
 194                                                 $target = $params['4::target'];
 195                                                 $targetFile = $repo->newFile( Title::makeTitle( NS_FILE, $target ) );
 196                                                 $targetFile->purgeCache();
 197                                                 $this->verbose( "Purged file {$target}; move target @{$row->log_timestamp}.\n" );
 198                                         }
 199                                 }
 200
 201                                 $this->verbose( "Purged file {$row->log_title}; {$type} @{$row->log_timestamp}.\n" );
 202
 203                                 if ( $this->hasOption( 'sleep-per-batch' ) && ++$bSize > $this->mBatchSize ) {
 204                                         $bSize = 0;
 205                                         // sleep-per-batch is milliseconds, usleep wants micro seconds.
 206                                         usleep( 1000 * (int)$this->getOption( 'sleep-per-batch' ) );
 207                                 }
 208                         }
 209                 }
 210         }
 211
 212         protected function purgeFromArchiveTable( LocalRepo $repo, LocalFile $file ) {
 213                 $dbr = $repo->getReplicaDB();
 214                 $res = $dbr->select(
 215                         'filearchive',
 216                         [ 'fa_archive_name' ],
 217                         [ 'fa_name' => $file->getName() ],
 218                         __METHOD__
 219                 );
 220
 221                 foreach ( $res as $row ) {
 222                         if ( $row->fa_archive_name === null ) {
 223                                 // Was not an old version (current version names checked already)
 224                                 continue;
 225                         }
 226                         $ofile = $repo->newFromArchiveName( $file->getTitle(), $row->fa_archive_name );
 227                         // If there is an orphaned storage file still there...delete it
 228                         if ( !$file->exists() && $repo->fileExists( $ofile->getPath() ) ) {
 229                                 $dpath = $this->getDeletedPath( $repo, $ofile );
 230                                 if ( $repo->fileExists( $dpath ) ) {
 231                                         // Sanity check to avoid data loss
 232                                         $repo->getBackend()->delete( [ 'src' => $ofile->getPath() ] );
 233                                         $this->output( "Deleted orphan file: {$ofile->getPath()}.\n" );
 234                                 } else {
 235                                         $this->error( "File was not deleted: {$ofile->getPath()}.\n" );
 236                                 }
 237                         }
 238                         $file->purgeOldThumbnails( $row->fa_archive_name );
 239                 }
 240         }
 241
 242         protected function getDeletedPath( LocalRepo $repo, LocalFile $file ) {
 243                 $hash = $repo->getFileSha1( $file->getPath() );
 244                 $key = "{$hash}.{$file->getExtension()}";
 245
 246                 return $repo->getDeletedHashPath( $key ) . $key;
 247         }
 248
 249         /**
 250          * Send an output message iff the 'verbose' option has been provided.
 251          *
 252          * @param string $msg Message to output
 253          */
 254         protected function verbose( $msg ) {
 255                 if ( $this->hasOption( 'verbose' ) ) {
 256                         $this->output( $msg );
 257                 }
 258         }
 259 }
 260
 261 $maintClass = "PurgeChangedFiles";
 262 require_once RUN_MAINTENANCE_IF_MAIN;