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         /**
  31          * Collection of URLs to purge.
  32          * @var array
  33          */
  34         protected $urlArr;
  35
  36         /**
  37          * @param array $urlArr Collection of URLs to purge
  38          * @param bool|int $maxTitles Maximum number of unique URLs to purge
  39          */
  40         public function __construct( $urlArr = array(), $maxTitles = false ) {
  41                 global $wgMaxSquidPurgeTitles;
  42                 if ( $maxTitles === false ) {
  43                         $maxTitles = $wgMaxSquidPurgeTitles;
  44                 }
  45
  46                 // Remove duplicate URLs from list
  47                 $urlArr = array_unique( $urlArr );
  48                 if ( count( $urlArr ) > $maxTitles ) {
  49                         // Truncate to desired maximum URL count
  50                         $urlArr = array_slice( $urlArr, 0, $maxTitles );
  51                 }
  52                 $this->urlArr = $urlArr;
  53         }
  54
  55         /**
  56          * Create a SquidUpdate from the given Title object.
  57          *
  58          * The resulting SquidUpdate will purge the given Title's URLs as well as
  59          * the pages that link to it. Capped at $wgMaxSquidPurgeTitles total URLs.
  60          *
  61          * @param Title $title
  62          * @return SquidUpdate
  63          */
  64         public static function newFromLinksTo( Title $title ) {
  65                 global $wgMaxSquidPurgeTitles;
  66                 wfProfileIn( __METHOD__ );
  67
  68                 # Get a list of URLs linking to this page
  69                 $dbr = wfGetDB( DB_SLAVE );
  70                 $res = $dbr->select( array( 'links', 'page' ),
  71                         array( 'page_namespace', 'page_title' ),
  72                         array(
  73                                 'pl_namespace' => $title->getNamespace(),
  74                                 'pl_title' => $title->getDBkey(),
  75                                 'pl_from=page_id' ),
  76                         __METHOD__ );
  77                 $blurlArr = $title->getSquidURLs();
  78                 if ( $res->numRows() <= $wgMaxSquidPurgeTitles ) {
  79                         foreach ( $res as $BL ) {
  80                                 $tobj = Title::makeTitle( $BL->page_namespace, $BL->page_title );
  81                                 $blurlArr[] = $tobj->getInternalURL();
  82                         }
  83                 }
  84
  85                 wfProfileOut( __METHOD__ );
  86                 return new SquidUpdate( $blurlArr );
  87         }
  88
  89         /**
  90          * Create a SquidUpdate from an array of Title objects, or a TitleArray object
  91          *
  92          * @param array $titles
  93          * @param array $urlArr
  94          * @return SquidUpdate
  95          */
  96         public static function newFromTitles( $titles, $urlArr = array() ) {
  97                 global $wgMaxSquidPurgeTitles;
  98                 $i = 0;
  99                 foreach ( $titles as $title ) {
 100                         $urlArr[] = $title->getInternalURL();
 101                         if ( $i++ > $wgMaxSquidPurgeTitles ) {
 102                                 break;
 103                         }
 104                 }
 105                 return new SquidUpdate( $urlArr );
 106         }
 107
 108         /**
 109          * @param Title $title
 110          * @return SquidUpdate
 111          */
 112         public static function newSimplePurge( Title $title ) {
 113                 $urlArr = $title->getSquidURLs();
 114                 return new SquidUpdate( $urlArr );
 115         }
 116
 117         /**
 118          * Purges the list of URLs passed to the constructor.
 119          */
 120         public function doUpdate() {
 121                 self::purge( $this->urlArr );
 122         }
 123
 124         /**
 125          * Purges a list of Squids defined in $wgSquidServers.
 126          * $urlArr should contain the full URLs to purge as values
 127          * (example: $urlArr[] = 'http://my.host/something')
 128          * XXX report broken Squids per mail or log
 129          *
 130          * @param array $urlArr List of full URLs to purge
 131          */
 132         public static function purge( $urlArr ) {
 133                 global $wgSquidServers, $wgHTCPRouting;
 134
 135                 if ( !$urlArr ) {
 136                         return;
 137                 }
 138
 139                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) . "\n" );
 140
 141                 if ( $wgHTCPRouting ) {
 142                         self::HTCPPurge( $urlArr );
 143                 }
 144
 145                 wfProfileIn( __METHOD__ );
 146
 147                 // Remove duplicate URLs
 148                 $urlArr = array_unique( $urlArr );
 149                 // Maximum number of parallel connections per squid
 150                 $maxSocketsPerSquid = 8;
 151                 // Number of requests to send per socket
 152                 // 400 seems to be a good tradeoff, opening a socket takes a while
 153                 $urlsPerSocket = 400;
 154                 $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 155                 if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 156                         $socketsPerSquid = $maxSocketsPerSquid;
 157                 }
 158
 159                 $pool = new SquidPurgeClientPool;
 160                 $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 161                 foreach ( $wgSquidServers as $server ) {
 162                         foreach ( $chunks as $chunk ) {
 163                                 $client = new SquidPurgeClient( $server );
 164                                 foreach ( $chunk as $url ) {
 165                                         $client->queuePurge( $url );
 166                                 }
 167                                 $pool->addClient( $client );
 168                         }
 169                 }
 170                 $pool->run();
 171
 172                 wfProfileOut( __METHOD__ );
 173         }
 174
 175         /**
 176          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 177          *
 178          * @throws MWException
 179          * @param array $urlArr Collection of URLs to purge
 180          */
 181         public static function HTCPPurge( $urlArr ) {
 182                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 183                 wfProfileIn( __METHOD__ );
 184
 185                 // HTCP CLR operation
 186                 $htcpOpCLR = 4;
 187
 188                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 189                 if ( !defined( "IPPROTO_IP" ) ) {
 190                         define( "IPPROTO_IP", 0 );
 191                         define( "IP_MULTICAST_LOOP", 34 );
 192                         define( "IP_MULTICAST_TTL", 33 );
 193                 }
 194
 195                 // pfsockopen doesn't work because we need set_sock_opt
 196                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 197                 if ( ! $conn ) {
 198                         $errstr = socket_strerror( socket_last_error() );
 199                         wfDebugLog( 'squid', __METHOD__ .
 200                                 ": Error opening UDP socket: $errstr\n" );
 201                         wfProfileOut( __METHOD__ );
 202                         return;
 203                 }
 204
 205                 // Set socket options
 206                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 207                 if ( $wgHTCPMulticastTTL != 1 ) {
 208                         // Set multicast time to live (hop count) option on socket
 209                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 210                                 $wgHTCPMulticastTTL );
 211                 }
 212
 213                 // Remove duplicate URLs from collection
 214                 $urlArr = array_unique( $urlArr );
 215                 foreach ( $urlArr as $url ) {
 216                         if ( !is_string( $url ) ) {
 217                                 wfProfileOut( __METHOD__ );
 218                                 throw new MWException( 'Bad purge URL' );
 219                         }
 220                         $url = self::expand( $url );
 221                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 222                         if ( !$conf ) {
 223                                 wfDebugLog( 'squid', __METHOD__ .
 224                                         "No HTCP rule configured for URL {$url} , skipping\n" );
 225                                 continue;
 226                         }
 227
 228                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 229                                 // Normalize single entries
 230                                 $conf = array( $conf );
 231                         }
 232                         foreach ( $conf as $subconf ) {
 233                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 234                                         wfProfileOut( __METHOD__ );
 235                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 236                                 }
 237                         }
 238
 239                         // Construct a minimal HTCP request diagram
 240                         // as per RFC 2756
 241                         // Opcode 'CLR', no response desired, no auth
 242                         $htcpTransID = rand();
 243
 244                         $htcpSpecifier = pack( 'na4na*na8n',
 245                                 4, 'HEAD', strlen( $url ), $url,
 246                                 8, 'HTTP/1.0', 0 );
 247
 248                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 249                         $htcpLen = 4 + $htcpDataLen + 2;
 250
 251                         // Note! Squid gets the bit order of the first
 252                         // word wrong, wrt the RFC. Apparently no other
 253                         // implementation exists, so adapt to Squid
 254                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 255                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 256                                 $htcpTransID, $htcpSpecifier, 2 );
 257
 258                         wfDebugLog( 'squid', __METHOD__ .
 259                                 "Purging URL $url via HTCP\n" );
 260                         foreach ( $conf as $subconf ) {
 261                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 262                                         $subconf['host'], $subconf['port'] );
 263                         }
 264                 }
 265                 wfProfileOut( __METHOD__ );
 266         }
 267
 268         /**
 269          * Expand local URLs to fully-qualified URLs using the internal protocol
 270          * and host defined in $wgInternalServer. Input that's already fully-
 271          * qualified will be passed through unchanged.
 272          *
 273          * This is used to generate purge URLs that may be either local to the
 274          * main wiki or include a non-native host, such as images hosted on a
 275          * second internal server.
 276          *
 277          * Client functions should not need to call this.
 278          *
 279          * @param string $url
 280          * @return string
 281          */
 282         public static function expand( $url ) {
 283                 return wfExpandUrl( $url, PROTO_INTERNAL );
 284         }
 285
 286         /**
 287          * Find the HTCP routing rule to use for a given URL.
 288          * @param string $url URL to match
 289          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 290          * @return mixed Element of $rules that matched, or false if nothing matched
 291          */
 292         private static function getRuleForURL( $url, $rules ) {
 293                 foreach ( $rules as $regex => $routing ) {
 294                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 295                                 return $routing;
 296                         }
 297                 }
 298                 return false;
 299         }
 300 }