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