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 the given Title object.
  56          *
  57          * The resulting SquidUpdate will purge the given Title's URLs as well as
  58          * the pages that link to it. Capped at $wgMaxSquidPurgeTitles total URLs.
  59          *
  60          * @param Title $title
  61          * @return SquidUpdate
  62          */
  63         public static function newFromLinksTo( Title $title ) {
  64                 global $wgMaxSquidPurgeTitles;
  65                 wfProfileIn( __METHOD__ );
  66
  67                 # Get a list of URLs linking to this page
  68                 $dbr = wfGetDB( DB_SLAVE );
  69                 $res = $dbr->select( array( 'links', 'page' ),
  70                         array( 'page_namespace', 'page_title' ),
  71                         array(
  72                                 'pl_namespace' => $title->getNamespace(),
  73                                 'pl_title' => $title->getDBkey(),
  74                                 'pl_from=page_id' ),
  75                         __METHOD__ );
  76                 $blurlArr = $title->getSquidURLs();
  77                 if ( $res->numRows() <= $wgMaxSquidPurgeTitles ) {
  78                         foreach ( $res as $BL ) {
  79                                 $tobj = Title::makeTitle( $BL->page_namespace, $BL->page_title );
  80                                 $blurlArr[] = $tobj->getInternalURL();
  81                         }
  82                 }
  83
  84                 wfProfileOut( __METHOD__ );
  85
  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
 106                 return new SquidUpdate( $urlArr );
 107         }
 108
 109         /**
 110          * @param Title $title
 111          * @return SquidUpdate
 112          */
 113         public static function newSimplePurge( Title $title ) {
 114                 $urlArr = $title->getSquidURLs();
 115
 116                 return new SquidUpdate( $urlArr );
 117         }
 118
 119         /**
 120          * Purges the list of URLs passed to the constructor.
 121          */
 122         public function doUpdate() {
 123                 self::purge( $this->urlArr );
 124         }
 125
 126         /**
 127          * Purges a list of Squids defined in $wgSquidServers.
 128          * $urlArr should contain the full URLs to purge as values
 129          * (example: $urlArr[] = 'http://my.host/something')
 130          * XXX report broken Squids per mail or log
 131          *
 132          * @param array $urlArr List of full URLs to purge
 133          */
 134         public static function purge( $urlArr ) {
 135                 global $wgSquidServers, $wgHTCPRouting;
 136
 137                 if ( !$urlArr ) {
 138                         return;
 139                 }
 140
 141                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) . "\n" );
 142
 143                 if ( $wgHTCPRouting ) {
 144                         self::HTCPPurge( $urlArr );
 145                 }
 146
 147                 wfProfileIn( __METHOD__ );
 148
 149                 // Remove duplicate URLs
 150                 $urlArr = array_unique( $urlArr );
 151                 // Maximum number of parallel connections per squid
 152                 $maxSocketsPerSquid = 8;
 153                 // Number of requests to send per socket
 154                 // 400 seems to be a good tradeoff, opening a socket takes a while
 155                 $urlsPerSocket = 400;
 156                 $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 157                 if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 158                         $socketsPerSquid = $maxSocketsPerSquid;
 159                 }
 160
 161                 $pool = new SquidPurgeClientPool;
 162                 $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 163                 foreach ( $wgSquidServers as $server ) {
 164                         foreach ( $chunks as $chunk ) {
 165                                 $client = new SquidPurgeClient( $server );
 166                                 foreach ( $chunk as $url ) {
 167                                         $client->queuePurge( $url );
 168                                 }
 169                                 $pool->addClient( $client );
 170                         }
 171                 }
 172                 $pool->run();
 173
 174                 wfProfileOut( __METHOD__ );
 175         }
 176
 177         /**
 178          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 179          *
 180          * @throws MWException
 181          * @param array $urlArr Collection of URLs to purge
 182          */
 183         public static function HTCPPurge( $urlArr ) {
 184                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 185                 wfProfileIn( __METHOD__ );
 186
 187                 // HTCP CLR operation
 188                 $htcpOpCLR = 4;
 189
 190                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 191                 if ( !defined( "IPPROTO_IP" ) ) {
 192                         define( "IPPROTO_IP", 0 );
 193                         define( "IP_MULTICAST_LOOP", 34 );
 194                         define( "IP_MULTICAST_TTL", 33 );
 195                 }
 196
 197                 // pfsockopen doesn't work because we need set_sock_opt
 198                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 199                 if ( !$conn ) {
 200                         $errstr = socket_strerror( socket_last_error() );
 201                         wfDebugLog( 'squid', __METHOD__ .
 202                                 ": Error opening UDP socket: $errstr\n" );
 203                         wfProfileOut( __METHOD__ );
 204
 205                         return;
 206                 }
 207
 208                 // Set socket options
 209                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 210                 if ( $wgHTCPMulticastTTL != 1 ) {
 211                         // Set multicast time to live (hop count) option on socket
 212                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 213                                 $wgHTCPMulticastTTL );
 214                 }
 215
 216                 // Remove duplicate URLs from collection
 217                 $urlArr = array_unique( $urlArr );
 218                 foreach ( $urlArr as $url ) {
 219                         if ( !is_string( $url ) ) {
 220                                 wfProfileOut( __METHOD__ );
 221                                 throw new MWException( 'Bad purge URL' );
 222                         }
 223                         $url = self::expand( $url );
 224                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 225                         if ( !$conf ) {
 226                                 wfDebugLog( 'squid', __METHOD__ .
 227                                         "No HTCP rule configured for URL {$url} , skipping\n" );
 228                                 continue;
 229                         }
 230
 231                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 232                                 // Normalize single entries
 233                                 $conf = array( $conf );
 234                         }
 235                         foreach ( $conf as $subconf ) {
 236                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 237                                         wfProfileOut( __METHOD__ );
 238                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 239                                 }
 240                         }
 241
 242                         // Construct a minimal HTCP request diagram
 243                         // as per RFC 2756
 244                         // Opcode 'CLR', no response desired, no auth
 245                         $htcpTransID = rand();
 246
 247                         $htcpSpecifier = pack( 'na4na*na8n',
 248                                 4, 'HEAD', strlen( $url ), $url,
 249                                 8, 'HTTP/1.0', 0 );
 250
 251                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 252                         $htcpLen = 4 + $htcpDataLen + 2;
 253
 254                         // Note! Squid gets the bit order of the first
 255                         // word wrong, wrt the RFC. Apparently no other
 256                         // implementation exists, so adapt to Squid
 257                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 258                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 259                                 $htcpTransID, $htcpSpecifier, 2 );
 260
 261                         wfDebugLog( 'squid', __METHOD__ .
 262                                 "Purging URL $url via HTCP\n" );
 263                         foreach ( $conf as $subconf ) {
 264                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 265                                         $subconf['host'], $subconf['port'] );
 266                         }
 267                 }
 268                 wfProfileOut( __METHOD__ );
 269         }
 270
 271         /**
 272          * Expand local URLs to fully-qualified URLs using the internal protocol
 273          * and host defined in $wgInternalServer. Input that's already fully-
 274          * qualified will be passed through unchanged.
 275          *
 276          * This is used to generate purge URLs that may be either local to the
 277          * main wiki or include a non-native host, such as images hosted on a
 278          * second internal server.
 279          *
 280          * Client functions should not need to call this.
 281          *
 282          * @param string $url
 283          * @return string
 284          */
 285         public static function expand( $url ) {
 286                 return wfExpandUrl( $url, PROTO_INTERNAL );
 287         }
 288
 289         /**
 290          * Find the HTCP routing rule to use for a given URL.
 291          * @param string $url URL to match
 292          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 293          * @return mixed Element of $rules that matched, or false if nothing matched
 294          */
 295         private static function getRuleForURL( $url, $rules ) {
 296                 foreach ( $rules as $regex => $routing ) {
 297                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 298                                 return $routing;
 299                         }
 300                 }
 301
 302                 return false;
 303         }
 304 }