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