maintenance/purgeChangedPages.php

   1 <?php
   2 /**
   3  * Send purge requests for pages edited in date range to squid/varnish.
   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 // @codeCoverageIgnoreStart
  25 require_once __DIR__ . '/Maintenance.php';
  26 // @codeCoverageIgnoreEnd
  27
  28 use MediaWiki\Maintenance\Maintenance;
  29 use MediaWiki\Title\Title;
  30 use Wikimedia\Rdbms\IResultWrapper;
  31
  32 /**
  33  * Maintenance script that sends purge requests for pages edited in a date
  34  * range to squid/varnish.
  35  *
  36  * Can be used to recover from an HTCP message partition or other major cache
  37  * layer interruption.
  38  *
  39  * @ingroup Maintenance
  40  */
  41 class PurgeChangedPages extends Maintenance {
  42
  43         public function __construct() {
  44                 parent::__construct();
  45                 $this->addDescription( 'Send purge requests for edits in date range to squid/varnish' );
  46                 $this->addOption( 'starttime', 'Starting timestamp', true, true );
  47                 $this->addOption( 'endtime', 'Ending timestamp', true, true );
  48                 $this->addOption( 'htcp-dest', 'HTCP announcement destination (IP:port)', false, true );
  49                 $this->addOption( 'sleep-per-batch', 'Milliseconds to sleep between batches', false, true );
  50                 $this->addOption( 'dry-run', 'Do not send purge requests' );
  51                 $this->addOption( 'verbose', 'Show more output', false, false, 'v' );
  52                 $this->setBatchSize( 100 );
  53         }
  54
  55         public function execute() {
  56                 global $wgHTCPRouting;
  57
  58                 if ( $this->hasOption( 'htcp-dest' ) ) {
  59                         $parts = explode( ':', $this->getOption( 'htcp-dest' ), 2 );
  60                         if ( count( $parts ) < 2 ) {
  61                                 // Add default htcp port
  62                                 $parts[] = '4827';
  63                         }
  64
  65                         // Route all HTCP messages to provided host:port
  66                         $wgHTCPRouting = [
  67                                 '' => [ 'host' => $parts[0], 'port' => $parts[1] ],
  68                         ];
  69                         if ( $this->hasOption( 'verbose' ) ) {
  70                                 $this->output( "HTCP broadcasts to {$parts[0]}:{$parts[1]}\n" );
  71                         }
  72                 }
  73
  74                 $dbr = $this->getReplicaDB();
  75                 $minTime = $dbr->timestamp( $this->getOption( 'starttime' ) );
  76                 $maxTime = $dbr->timestamp( $this->getOption( 'endtime' ) );
  77
  78                 if ( $maxTime < $minTime ) {
  79                         $this->error( "\nERROR: starttime after endtime\n" );
  80                         $this->maybeHelp( true );
  81                 }
  82
  83                 $stuckCount = 0;
  84                 while ( true ) {
  85                         // Adjust bach size if we are stuck in a second that had many changes
  86                         $bSize = ( $stuckCount + 1 ) * $this->getBatchSize();
  87
  88                         $res = $dbr->newSelectQueryBuilder()
  89                                 ->select( [ 'rev_timestamp', 'page_namespace', 'page_title', ] )
  90                                 ->from( 'revision' )
  91                                 ->join( 'page', null, 'rev_page=page_id' )
  92                                 ->where( [
  93                                         $dbr->expr( 'rev_timestamp', '>', $minTime ),
  94                                         $dbr->expr( 'rev_timestamp', '<=', $maxTime ),
  95                                 ] )
  96                                 // Only get rows where the revision is the latest for the page.
  97                                 // Other revisions would be duplicate and we don't need to purge if
  98                                 // there has been an edit after the interesting time window.
  99                                 ->andWhere( "page_latest = rev_id" )
 100                                 ->orderBy( 'rev_timestamp' )
 101                                 ->limit( $bSize )
 102                                 ->caller( __METHOD__ )->fetchResultSet();
 103
 104                         if ( !$res->numRows() ) {
 105                                 // nothing more found so we are done
 106                                 break;
 107                         }
 108
 109                         // Kludge to not get stuck in loops for batches with the same timestamp
 110                         [ $rows, $lastTime ] = $this->pageableSortedRows( $res, 'rev_timestamp', $bSize );
 111                         if ( !count( $rows ) ) {
 112                                 ++$stuckCount;
 113                                 continue;
 114                         }
 115                         // Reset suck counter
 116                         $stuckCount = 0;
 117
 118                         $this->output( "Processing changes from {$minTime} to {$lastTime}.\n" );
 119
 120                         // Advance past the last row next time
 121                         $minTime = $lastTime;
 122
 123                         // Create list of URLs from page_namespace + page_title
 124                         $urls = [];
 125                         foreach ( $rows as $row ) {
 126                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 127                                 $urls[] = $title->getInternalURL();
 128                         }
 129
 130                         if ( $this->hasOption( 'dry-run' ) || $this->hasOption( 'verbose' ) ) {
 131                                 $this->output( implode( "\n", $urls ) . "\n" );
 132                                 if ( $this->hasOption( 'dry-run' ) ) {
 133                                         continue;
 134                                 }
 135                         }
 136
 137                         // Send batch of purge requests out to CDN servers
 138                         $hcu = $this->getServiceContainer()->getHtmlCacheUpdater();
 139                         $hcu->purgeUrls( $urls, $hcu::PURGE_NAIVE );
 140
 141                         if ( $this->hasOption( 'sleep-per-batch' ) ) {
 142                                 // sleep-per-batch is milliseconds, usleep wants micro seconds.
 143                                 usleep( 1000 * (int)$this->getOption( 'sleep-per-batch' ) );
 144                         }
 145                 }
 146
 147                 $this->output( "Done!\n" );
 148         }
 149
 150         /**
 151          * Remove all the rows in a result set with the highest value for column
 152          * $column unless the number of rows is less $limit. This returns the new
 153          * array of rows and the highest value of column $column for the rows left.
 154          * The ordering of rows is maintained.
 155          *
 156          * This is useful for paging on mostly-unique values that may sometimes
 157          * have large clumps of identical values. It should be safe to do the next
 158          * query on items with a value higher than the highest of the rows returned here.
 159          * If this returns an empty array for a non-empty query result, then all the rows
 160          * had the same column value and the query should be repeated with a higher LIMIT.
 161          *
 162          * @todo move this elsewhere
 163          *
 164          * @param IResultWrapper $res Query result sorted by $column (ascending)
 165          * @param string $column
 166          * @param int $limit
 167          * @return array (array of rows, string column value)
 168          */
 169         protected function pageableSortedRows( IResultWrapper $res, $column, $limit ) {
 170                 $rows = iterator_to_array( $res, false );
 171
 172                 // Nothing to do
 173                 if ( !$rows ) {
 174                         return [ [], null ];
 175                 }
 176
 177                 $lastValue = end( $rows )->$column;
 178                 if ( count( $rows ) < $limit ) {
 179                         return [ $rows, $lastValue ];
 180                 }
 181
 182                 for ( $i = count( $rows ) - 1; $i >= 0; --$i ) {
 183                         if ( $rows[$i]->$column !== $lastValue ) {
 184                                 break;
 185                         }
 186
 187                         unset( $rows[$i] );
 188                 }
 189
 190                 // No more rows left
 191                 if ( !$rows ) {
 192                         return [ [], null ];
 193                 }
 194
 195                 return [ $rows, end( $rows )->$column ];
 196         }
 197 }
 198
 199 // @codeCoverageIgnoreStart
 200 $maintClass = PurgeChangedPages::class;
 201 require_once RUN_MAINTENANCE_IF_MAIN;
 202 // @codeCoverageIgnoreEnd