includes/deferred/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 implements DeferrableUpdate {
  29         /** @var string[] Collection of URLs to purge */
  30         protected $urls = array();
  31
  32         /**
  33          * @param array $urlArr Collection of URLs to purge
  34          */
  35         public function __construct( array $urlArr ) {
  36                 // Remove duplicate URLs from list
  37                 $this->urls = array_unique( $urlArr );
  38         }
  39
  40         /**
  41          * Create a SquidUpdate from an array of Title objects, or a TitleArray object
  42          *
  43          * @param Traversable|array $titles
  44          * @param array $urlArr
  45          * @return SquidUpdate
  46          */
  47         public static function newFromTitles( $titles, $urlArr = array() ) {
  48                 /** @var Title $title */
  49                 foreach ( $titles as $title ) {
  50                         $urlArr[] = $title->getInternalURL();
  51                 }
  52
  53                 return new SquidUpdate( $urlArr );
  54         }
  55
  56         /**
  57          * @param Title $title
  58          * @return SquidUpdate
  59          */
  60         public static function newSimplePurge( Title $title ) {
  61                 $urlArr = $title->getSquidURLs();
  62
  63                 return new SquidUpdate( $urlArr );
  64         }
  65
  66         /**
  67          * Purges the list of URLs passed to the constructor.
  68          */
  69         public function doUpdate() {
  70                 self::purge( $this->urls );
  71         }
  72
  73         /**
  74          * Purges a list of Squids defined in $wgSquidServers.
  75          * $urlArr should contain the full URLs to purge as values
  76          * (example: $urlArr[] = 'http://my.host/something')
  77          * XXX report broken Squids per mail or log
  78          *
  79          * @param array $urlArr List of full URLs to purge
  80          */
  81         public static function purge( array $urlArr ) {
  82                 global $wgSquidServers, $wgHTCPRouting;
  83
  84                 if ( !$urlArr ) {
  85                         return;
  86                 }
  87
  88                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) );
  89
  90                 if ( $wgHTCPRouting ) {
  91                         self::HTCPPurge( $urlArr );
  92                 }
  93
  94                 if ( $wgSquidServers ) {
  95                         // Remove duplicate URLs
  96                         $urlArr = array_unique( $urlArr );
  97                         // Maximum number of parallel connections per squid
  98                         $maxSocketsPerSquid = 8;
  99                         // Number of requests to send per socket
 100                         // 400 seems to be a good tradeoff, opening a socket takes a while
 101                         $urlsPerSocket = 400;
 102                         $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 103                         if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 104                                 $socketsPerSquid = $maxSocketsPerSquid;
 105                         }
 106
 107                         $pool = new SquidPurgeClientPool;
 108                         $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 109                         foreach ( $wgSquidServers as $server ) {
 110                                 foreach ( $chunks as $chunk ) {
 111                                         $client = new SquidPurgeClient( $server );
 112                                         foreach ( $chunk as $url ) {
 113                                                 $client->queuePurge( $url );
 114                                         }
 115                                         $pool->addClient( $client );
 116                                 }
 117                         }
 118
 119                         $pool->run();
 120                 }
 121         }
 122
 123         /**
 124          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 125          *
 126          * @throws MWException
 127          * @param array $urlArr Collection of URLs to purge
 128          */
 129         protected static function HTCPPurge( $urlArr ) {
 130                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 131
 132                 // HTCP CLR operation
 133                 $htcpOpCLR = 4;
 134
 135                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 136                 if ( !defined( "IPPROTO_IP" ) ) {
 137                         define( "IPPROTO_IP", 0 );
 138                         define( "IP_MULTICAST_LOOP", 34 );
 139                         define( "IP_MULTICAST_TTL", 33 );
 140                 }
 141
 142                 // pfsockopen doesn't work because we need set_sock_opt
 143                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 144                 if ( !$conn ) {
 145                         $errstr = socket_strerror( socket_last_error() );
 146                         wfDebugLog( 'squid', __METHOD__ .
 147                                 ": Error opening UDP socket: $errstr" );
 148
 149                         return;
 150                 }
 151
 152                 // Set socket options
 153                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 154                 if ( $wgHTCPMulticastTTL != 1 ) {
 155                         // Set multicast time to live (hop count) option on socket
 156                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 157                                 $wgHTCPMulticastTTL );
 158                 }
 159
 160                 // Remove duplicate URLs from collection
 161                 $urlArr = array_unique( $urlArr );
 162                 // Get sequential trx IDs for packet loss counting
 163                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 164                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 165                 );
 166
 167                 foreach ( $urlArr as $url ) {
 168                         if ( !is_string( $url ) ) {
 169                                 throw new MWException( 'Bad purge URL' );
 170                         }
 171                         $url = self::expand( $url );
 172                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 173                         if ( !$conf ) {
 174                                 wfDebugLog( 'squid', __METHOD__ .
 175                                         "No HTCP rule configured for URL {$url} , skipping" );
 176                                 continue;
 177                         }
 178
 179                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 180                                 // Normalize single entries
 181                                 $conf = array( $conf );
 182                         }
 183                         foreach ( $conf as $subconf ) {
 184                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 185                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 186                                 }
 187                         }
 188
 189                         // Construct a minimal HTCP request diagram
 190                         // as per RFC 2756
 191                         // Opcode 'CLR', no response desired, no auth
 192                         $htcpTransID = current( $ids );
 193                         next( $ids );
 194
 195                         $htcpSpecifier = pack( 'na4na*na8n',
 196                                 4, 'HEAD', strlen( $url ), $url,
 197                                 8, 'HTTP/1.0', 0 );
 198
 199                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 200                         $htcpLen = 4 + $htcpDataLen + 2;
 201
 202                         // Note! Squid gets the bit order of the first
 203                         // word wrong, wrt the RFC. Apparently no other
 204                         // implementation exists, so adapt to Squid
 205                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 206                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 207                                 $htcpTransID, $htcpSpecifier, 2 );
 208
 209                         wfDebugLog( 'squid', __METHOD__ .
 210                                 "Purging URL $url via HTCP" );
 211                         foreach ( $conf as $subconf ) {
 212                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 213                                         $subconf['host'], $subconf['port'] );
 214                         }
 215                 }
 216         }
 217
 218         /**
 219          * Expand local URLs to fully-qualified URLs using the internal protocol
 220          * and host defined in $wgInternalServer. Input that's already fully-
 221          * qualified will be passed through unchanged.
 222          *
 223          * This is used to generate purge URLs that may be either local to the
 224          * main wiki or include a non-native host, such as images hosted on a
 225          * second internal server.
 226          *
 227          * Client functions should not need to call this.
 228          *
 229          * @param string $url
 230          * @return string
 231          */
 232         public static function expand( $url ) {
 233                 return wfExpandUrl( $url, PROTO_INTERNAL );
 234         }
 235
 236         /**
 237          * Find the HTCP routing rule to use for a given URL.
 238          * @param string $url URL to match
 239          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 240          * @return mixed Element of $rules that matched, or false if nothing matched
 241          */
 242         private static function getRuleForURL( $url, $rules ) {
 243                 foreach ( $rules as $regex => $routing ) {
 244                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 245                                 return $routing;
 246                         }
 247                 }
 248
 249                 return false;
 250         }
 251 }