includes/cache/SquidUpdate.php

   1 <?php
   2 /**
   3  * Squid cache purging.
   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 Cache
  22  */
  23
  24 /**
  25  * Handles purging appropriate Squid URLs given a title (or titles)
  26  * @ingroup Cache
  27  */
  28 class SquidUpdate {
  29         var $urlArr, $mMaxTitles;
  30
  31         /**
  32          * @param $urlArr array
  33          * @param $maxTitles bool|int
  34          */
  35         function __construct( $urlArr = array(), $maxTitles = false ) {
  36                 global $wgMaxSquidPurgeTitles;
  37                 if ( $maxTitles === false ) {
  38                         $this->mMaxTitles = $wgMaxSquidPurgeTitles;
  39                 } else {
  40                         $this->mMaxTitles = $maxTitles;
  41                 }
  42                 $urlArr = array_unique( $urlArr ); // Remove duplicates
  43                 if ( count( $urlArr ) > $this->mMaxTitles ) {
  44                         $urlArr = array_slice( $urlArr, 0, $this->mMaxTitles );
  45                 }
  46                 $this->urlArr = $urlArr;
  47         }
  48
  49         /**
  50          * @param $title Title
  51          *
  52          * @return SquidUpdate
  53          */
  54         static function newFromLinksTo( &$title ) {
  55                 global $wgMaxSquidPurgeTitles;
  56                 wfProfileIn( __METHOD__ );
  57
  58                 # Get a list of URLs linking to this page
  59                 $dbr = wfGetDB( DB_SLAVE );
  60                 $res = $dbr->select( array( 'links', 'page' ),
  61                         array( 'page_namespace', 'page_title' ),
  62                         array(
  63                                 'pl_namespace' => $title->getNamespace(),
  64                                 'pl_title' => $title->getDBkey(),
  65                                 'pl_from=page_id' ),
  66                         __METHOD__ );
  67                 $blurlArr = $title->getSquidURLs();
  68                 if ( $res->numRows() <= $wgMaxSquidPurgeTitles ) {
  69                         foreach ( $res as $BL ) {
  70                                 $tobj = Title::makeTitle( $BL->page_namespace, $BL->page_title );
  71                                 $blurlArr[] = $tobj->getInternalURL();
  72                         }
  73                 }
  74
  75                 wfProfileOut( __METHOD__ );
  76                 return new SquidUpdate( $blurlArr );
  77         }
  78
  79         /**
  80          * Create a SquidUpdate from an array of Title objects, or a TitleArray object
  81          *
  82          * @param $titles array
  83          * @param $urlArr array
  84          *
  85          * @return SquidUpdate
  86          */
  87         static function newFromTitles( $titles, $urlArr = array() ) {
  88                 global $wgMaxSquidPurgeTitles;
  89                 $i = 0;
  90                 foreach ( $titles as $title ) {
  91                         $urlArr[] = $title->getInternalURL();
  92                         if ( $i++ > $wgMaxSquidPurgeTitles ) {
  93                                 break;
  94                         }
  95                 }
  96                 return new SquidUpdate( $urlArr );
  97         }
  98
  99         /**
 100          * @param $title Title
 101          *
 102          * @return SquidUpdate
 103          */
 104         static function newSimplePurge( &$title ) {
 105                 $urlArr = $title->getSquidURLs();
 106                 return new SquidUpdate( $urlArr );
 107         }
 108
 109         /**
 110          * Purges the list of URLs passed to the constructor
 111          */
 112         function doUpdate() {
 113                 SquidUpdate::purge( $this->urlArr );
 114         }
 115
 116         /**
 117          * Purges a list of Squids defined in $wgSquidServers.
 118          * $urlArr should contain the full URLs to purge as values
 119          * (example: $urlArr[] = 'http://my.host/something')
 120          * XXX report broken Squids per mail or log
 121          *
 122          * @param $urlArr array
 123          * @return void
 124          */
 125         static function purge( $urlArr ) {
 126                 global $wgSquidServers, $wgHTCPRouting;
 127
 128                 if ( !$urlArr ) {
 129                         return;
 130                 }
 131
 132                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) . "\n" );
 133
 134                 if ( $wgHTCPRouting ) {
 135                         SquidUpdate::HTCPPurge( $urlArr );
 136                 }
 137
 138                 wfProfileIn( __METHOD__ );
 139
 140                 $urlArr = array_unique( $urlArr ); // Remove duplicates
 141                 $maxSocketsPerSquid = 8; //  socket cap per Squid
 142                 $urlsPerSocket = 400; // 400 seems to be a good tradeoff, opening a socket takes a while
 143                 $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 144                 if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 145                         $socketsPerSquid = $maxSocketsPerSquid;
 146                 }
 147
 148                 $pool = new SquidPurgeClientPool;
 149                 $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 150                 foreach ( $wgSquidServers as $server ) {
 151                         foreach ( $chunks as $chunk ) {
 152                                 $client = new SquidPurgeClient( $server );
 153                                 foreach ( $chunk as $url ) {
 154                                         $client->queuePurge( $url );
 155                                 }
 156                                 $pool->addClient( $client );
 157                         }
 158                 }
 159                 $pool->run();
 160
 161                 wfProfileOut( __METHOD__ );
 162         }
 163
 164         /**
 165          * @throws MWException
 166          * @param $urlArr array
 167          */
 168         static function HTCPPurge( $urlArr ) {
 169                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 170                 wfProfileIn( __METHOD__ );
 171
 172                 $htcpOpCLR = 4; // HTCP CLR
 173
 174                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 175                 if ( !defined( "IPPROTO_IP" ) ) {
 176                         define( "IPPROTO_IP", 0 );
 177                         define( "IP_MULTICAST_LOOP", 34 );
 178                         define( "IP_MULTICAST_TTL", 33 );
 179                 }
 180
 181                 // pfsockopen doesn't work because we need set_sock_opt
 182                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 183                 if ( ! $conn ) {
 184                         $errstr = socket_strerror( socket_last_error() );
 185                         wfDebugLog( 'squid', __METHOD__ .
 186                                 ": Error opening UDP socket: $errstr\n" );
 187                         wfProfileOut( __METHOD__ );
 188                         return;
 189                 }
 190
 191                 // Set socket options
 192                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 193                 if ( $wgHTCPMulticastTTL != 1 ) {
 194                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 195                                 $wgHTCPMulticastTTL );
 196                 }
 197
 198                 $urlArr = array_unique( $urlArr ); // Remove duplicates
 199                 foreach ( $urlArr as $url ) {
 200                         if ( !is_string( $url ) ) {
 201                                 wfProfileOut( __METHOD__ );
 202                                 throw new MWException( 'Bad purge URL' );
 203                         }
 204                         $url = SquidUpdate::expand( $url );
 205                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 206                         if ( !$conf ) {
 207                                 wfDebugLog( 'squid', __METHOD__ .
 208                                         "No HTCP rule configured for URL $url , skipping\n" );
 209                                 continue;
 210                         }
 211
 212                         if( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 213                                 // Normalize single entries
 214                                 $conf = array( $conf );
 215                         }
 216                         foreach( $conf as $subconf ) {
 217                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 218                                         wfProfileOut( __METHOD__ );
 219                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 220                                 }
 221                         }
 222
 223                         // Construct a minimal HTCP request diagram
 224                         // as per RFC 2756
 225                         // Opcode 'CLR', no response desired, no auth
 226                         $htcpTransID = rand();
 227
 228                         $htcpSpecifier = pack( 'na4na*na8n',
 229                                 4, 'HEAD', strlen( $url ), $url,
 230                                 8, 'HTTP/1.0', 0 );
 231
 232                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 233                         $htcpLen = 4 + $htcpDataLen + 2;
 234
 235                         // Note! Squid gets the bit order of the first
 236                         // word wrong, wrt the RFC. Apparently no other
 237                         // implementation exists, so adapt to Squid
 238                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 239                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 240                                 $htcpTransID, $htcpSpecifier, 2 );
 241
 242                         wfDebugLog( 'squid', __METHOD__ .
 243                                 "Purging URL $url via HTCP\n" );
 244                         foreach( $conf as $subconf ) {
 245                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 246                                         $subconf['host'], $subconf['port'] );
 247                         }
 248                 }
 249                 wfProfileOut( __METHOD__ );
 250         }
 251
 252         /**
 253          * Expand local URLs to fully-qualified URLs using the internal protocol
 254          * and host defined in $wgInternalServer. Input that's already fully-
 255          * qualified will be passed through unchanged.
 256          *
 257          * This is used to generate purge URLs that may be either local to the
 258          * main wiki or include a non-native host, such as images hosted on a
 259          * second internal server.
 260          *
 261          * Client functions should not need to call this.
 262          *
 263          * @param $url string
 264          *
 265          * @return string
 266          */
 267         static function expand( $url ) {
 268                 return wfExpandUrl( $url, PROTO_INTERNAL );
 269         }
 270
 271         /**
 272          * Find the HTCP routing rule to use for a given URL.
 273          * @param string $url URL to match
 274          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 275          * @return mixed Element of $rules that matched, or false if nothing matched
 276          */
 277         static function getRuleForURL( $url, $rules ) {
 278                 foreach ( $rules as $regex => $routing ) {
 279                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 280                                 return $routing;
 281                         }
 282                 }
 283                 return false;
 284         }
 285 }