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 require_once __DIR__ . '/Maintenance.php';
  25
  26 /**
  27  * Maintenance script that sends purge requests for pages edited in a date
  28  * range to squid/varnish.
  29  *
  30  * Can be used to recover from an HTCP message partition or other major cache
  31  * layer interruption.
  32  *
  33  * @ingroup Maintenance
  34  */
  35 class PurgeChangedPages extends Maintenance {
  36
  37         public function __construct() {
  38                 parent::__construct();
  39                 $this->addDescription( 'Send purge requests for edits in date range to squid/varnish' );
  40                 $this->addOption( 'starttime', 'Starting timestamp', true, true );
  41                 $this->addOption( 'endtime', 'Ending timestamp', true, true );
  42                 $this->addOption( 'htcp-dest', 'HTCP announcement destination (IP:port)', false, true );
  43                 $this->addOption( 'sleep-per-batch', 'Milliseconds to sleep between batches', false, true );
  44                 $this->addOption( 'dry-run', 'Do not send purge requests' );
  45                 $this->addOption( 'verbose', 'Show more output', false, false, 'v' );
  46                 $this->setBatchSize( 100 );
  47         }
  48
  49         public function execute() {
  50                 global $wgHTCPRouting;
  51
  52                 if ( $this->hasOption( 'htcp-dest' ) ) {
  53                         $parts = explode( ':', $this->getOption( 'htcp-dest' ) );
  54                         if ( count( $parts ) < 2 ) {
  55                                 // Add default htcp port
  56                                 $parts[] = '4827';
  57                         }
  58
  59                         // Route all HTCP messages to provided host:port
  60                         $wgHTCPRouting = [
  61                                 '' => [ 'host' => $parts[0], 'port' => $parts[1] ],
  62                         ];
  63                         if ( $this->hasOption( 'verbose' ) ) {
  64                                 $this->output( "HTCP broadcasts to {$parts[0]}:{$parts[1]}\n" );
  65                         }
  66                 }
  67
  68                 $dbr = $this->getDB( DB_REPLICA );
  69                 $minTime = $dbr->timestamp( $this->getOption( 'starttime' ) );
  70                 $maxTime = $dbr->timestamp( $this->getOption( 'endtime' ) );
  71
  72                 if ( $maxTime < $minTime ) {
  73                         $this->error( "\nERROR: starttime after endtime\n" );
  74                         $this->maybeHelp( true );
  75                 }
  76
  77                 $stuckCount = 0; // loop breaker
  78                 while ( true ) {
  79                         // Adjust bach size if we are stuck in a second that had many changes
  80                         $bSize = $this->mBatchSize + ( $stuckCount * $this->mBatchSize );
  81
  82                         $res = $dbr->select(
  83                                 [ 'page', 'revision' ],
  84                                 [
  85                                         'rev_timestamp',
  86                                         'page_namespace',
  87                                         'page_title',
  88                                 ],
  89                                 [
  90                                         "rev_timestamp > " . $dbr->addQuotes( $minTime ),
  91                                         "rev_timestamp <= " . $dbr->addQuotes( $maxTime ),
  92                                         // Only get rows where the revision is the latest for the page.
  93                                         // Other revisions would be duplicate and we don't need to purge if
  94                                         // there has been an edit after the interesting time window.
  95                                         "page_latest = rev_id",
  96                                 ],
  97                                 __METHOD__,
  98                                 [ 'ORDER BY' => 'rev_timestamp', 'LIMIT' => $bSize ],
  99                                 [
 100                                         'page' => [ 'INNER JOIN', 'rev_page=page_id' ],
 101                                 ]
 102                         );
 103
 104                         if ( !$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                         list( $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 squids
 138                         $squid = new CdnCacheUpdate( $urls, count( $urls ) );
 139                         $squid->doUpdate();
 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 ResultWrapper $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( ResultWrapper $res, $column, $limit ) {
 170                 $rows = iterator_to_array( $res, false );
 171                 $count = count( $rows );
 172                 if ( !$count ) {
 173                         return [ [], null ]; // nothing to do
 174                 } elseif ( $count < $limit ) {
 175                         return [ $rows, $rows[$count - 1]->$column ]; // no more rows left
 176                 }
 177                 $lastValue = $rows[$count - 1]->$column; // should be the highest
 178                 for ( $i = $count - 1; $i >= 0; --$i ) {
 179                         if ( $rows[$i]->$column === $lastValue ) {
 180                                 unset( $rows[$i] );
 181                         } else {
 182                                 break;
 183                         }
 184                 }
 185                 $lastValueLeft = count( $rows ) ? $rows[count( $rows ) - 1]->$column : null;
 186
 187                 return [ $rows, $lastValueLeft ];
 188         }
 189 }
 190
 191 $maintClass = "PurgeChangedPages";
 192 require_once RUN_MAINTENANCE_IF_MAIN;