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                 /** @var Title $title */
 100                 foreach ( $titles as $title ) {
 101                         $urlArr[] = $title->getInternalURL();
 102                         if ( $i++ > $wgMaxSquidPurgeTitles ) {
 103                                 break;
 104                         }
 105                 }
 106
 107                 return new SquidUpdate( $urlArr );
 108         }
 109
 110         /**
 111          * @param Title $title
 112          * @return SquidUpdate
 113          */
 114         public static function newSimplePurge( Title $title ) {
 115                 $urlArr = $title->getSquidURLs();
 116
 117                 return new SquidUpdate( $urlArr );
 118         }
 119
 120         /**
 121          * Purges the list of URLs passed to the constructor.
 122          */
 123         public function doUpdate() {
 124                 self::purge( $this->urlArr );
 125         }
 126
 127         /**
 128          * Purges a list of Squids defined in $wgSquidServers.
 129          * $urlArr should contain the full URLs to purge as values
 130          * (example: $urlArr[] = 'http://my.host/something')
 131          * XXX report broken Squids per mail or log
 132          *
 133          * @param array $urlArr List of full URLs to purge
 134          */
 135         public static function purge( $urlArr ) {
 136                 global $wgSquidServers, $wgHTCPRouting;
 137
 138                 if ( !$urlArr ) {
 139                         return;
 140                 }
 141
 142                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) . "\n" );
 143
 144                 if ( $wgHTCPRouting ) {
 145                         self::HTCPPurge( $urlArr );
 146                 }
 147
 148                 wfProfileIn( __METHOD__ );
 149
 150                 // Remove duplicate URLs
 151                 $urlArr = array_unique( $urlArr );
 152                 // Maximum number of parallel connections per squid
 153                 $maxSocketsPerSquid = 8;
 154                 // Number of requests to send per socket
 155                 // 400 seems to be a good tradeoff, opening a socket takes a while
 156                 $urlsPerSocket = 400;
 157                 $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 158                 if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 159                         $socketsPerSquid = $maxSocketsPerSquid;
 160                 }
 161
 162                 $pool = new SquidPurgeClientPool;
 163                 $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 164                 foreach ( $wgSquidServers as $server ) {
 165                         foreach ( $chunks as $chunk ) {
 166                                 $client = new SquidPurgeClient( $server );
 167                                 foreach ( $chunk as $url ) {
 168                                         $client->queuePurge( $url );
 169                                 }
 170                                 $pool->addClient( $client );
 171                         }
 172                 }
 173                 $pool->run();
 174
 175                 wfProfileOut( __METHOD__ );
 176         }
 177
 178         /**
 179          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 180          *
 181          * @throws MWException
 182          * @param array $urlArr Collection of URLs to purge
 183          */
 184         public static function HTCPPurge( $urlArr ) {
 185                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 186                 wfProfileIn( __METHOD__ );
 187
 188                 // HTCP CLR operation
 189                 $htcpOpCLR = 4;
 190
 191                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 192                 if ( !defined( "IPPROTO_IP" ) ) {
 193                         define( "IPPROTO_IP", 0 );
 194                         define( "IP_MULTICAST_LOOP", 34 );
 195                         define( "IP_MULTICAST_TTL", 33 );
 196                 }
 197
 198                 // pfsockopen doesn't work because we need set_sock_opt
 199                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 200                 if ( !$conn ) {
 201                         $errstr = socket_strerror( socket_last_error() );
 202                         wfDebugLog( 'squid', __METHOD__ .
 203                                 ": Error opening UDP socket: $errstr\n" );
 204                         wfProfileOut( __METHOD__ );
 205
 206                         return;
 207                 }
 208
 209                 // Set socket options
 210                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 211                 if ( $wgHTCPMulticastTTL != 1 ) {
 212                         // Set multicast time to live (hop count) option on socket
 213                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 214                                 $wgHTCPMulticastTTL );
 215                 }
 216
 217                 // Remove duplicate URLs from collection
 218                 $urlArr = array_unique( $urlArr );
 219                 // Get sequential trx IDs for packet loss counting
 220                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 221                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 222                 );
 223
 224                 foreach ( $urlArr as $url ) {
 225                         if ( !is_string( $url ) ) {
 226                                 wfProfileOut( __METHOD__ );
 227                                 throw new MWException( 'Bad purge URL' );
 228                         }
 229                         $url = self::expand( $url );
 230                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 231                         if ( !$conf ) {
 232                                 wfDebugLog( 'squid', __METHOD__ .
 233                                         "No HTCP rule configured for URL {$url} , skipping\n" );
 234                                 continue;
 235                         }
 236
 237                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 238                                 // Normalize single entries
 239                                 $conf = array( $conf );
 240                         }
 241                         foreach ( $conf as $subconf ) {
 242                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 243                                         wfProfileOut( __METHOD__ );
 244                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 245                                 }
 246                         }
 247
 248                         // Construct a minimal HTCP request diagram
 249                         // as per RFC 2756
 250                         // Opcode 'CLR', no response desired, no auth
 251                         $htcpTransID = current( $ids );
 252                         next( $ids );
 253
 254                         $htcpSpecifier = pack( 'na4na*na8n',
 255                                 4, 'HEAD', strlen( $url ), $url,
 256                                 8, 'HTTP/1.0', 0 );
 257
 258                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 259                         $htcpLen = 4 + $htcpDataLen + 2;
 260
 261                         // Note! Squid gets the bit order of the first
 262                         // word wrong, wrt the RFC. Apparently no other
 263                         // implementation exists, so adapt to Squid
 264                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 265                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 266                                 $htcpTransID, $htcpSpecifier, 2 );
 267
 268                         wfDebugLog( 'squid', __METHOD__ .
 269                                 "Purging URL $url via HTCP\n" );
 270                         foreach ( $conf as $subconf ) {
 271                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 272                                         $subconf['host'], $subconf['port'] );
 273                         }
 274                 }
 275                 wfProfileOut( __METHOD__ );
 276         }
 277
 278         /**
 279          * Expand local URLs to fully-qualified URLs using the internal protocol
 280          * and host defined in $wgInternalServer. Input that's already fully-
 281          * qualified will be passed through unchanged.
 282          *
 283          * This is used to generate purge URLs that may be either local to the
 284          * main wiki or include a non-native host, such as images hosted on a
 285          * second internal server.
 286          *
 287          * Client functions should not need to call this.
 288          *
 289          * @param string $url
 290          * @return string
 291          */
 292         public static function expand( $url ) {
 293                 return wfExpandUrl( $url, PROTO_INTERNAL );
 294         }
 295
 296         /**
 297          * Find the HTCP routing rule to use for a given URL.
 298          * @param string $url URL to match
 299          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 300          * @return mixed Element of $rules that matched, or false if nothing matched
 301          */
 302         private static function getRuleForURL( $url, $rules ) {
 303                 foreach ( $rules as $regex => $routing ) {
 304                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 305                                 return $routing;
 306                         }
 307                 }
 308
 309                 return false;
 310         }
 311 }