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 use Wikimedia\Assert\Assert;
  25
  26 /**
  27  * Handles purging appropriate Squid URLs given a title (or titles)
  28  * @ingroup Cache
  29  */
  30 class SquidUpdate implements DeferrableUpdate, MergeableUpdate {
  31         /** @var string[] Collection of URLs to purge */
  32         protected $urls = array();
  33
  34         /**
  35          * @param string[] $urlArr Collection of URLs to purge
  36          */
  37         public function __construct( array $urlArr ) {
  38                 $this->urls = $urlArr;
  39         }
  40
  41         /**
  42          * Create a SquidUpdate from an array of Title objects, or a TitleArray object
  43          *
  44          * @param Traversable|array $titles
  45          * @param string[] $urlArr
  46          * @return SquidUpdate
  47          */
  48         public static function newFromTitles( $titles, $urlArr = array() ) {
  49                 /** @var Title $title */
  50                 foreach ( $titles as $title ) {
  51                         $urlArr = array_merge( $urlArr, $title->getSquidURLs() );
  52                 }
  53
  54                 return new SquidUpdate( $urlArr );
  55         }
  56
  57         /**
  58          * @param Title $title
  59          * @return SquidUpdate
  60          * @deprecated 1.27
  61          */
  62         public static function newSimplePurge( Title $title ) {
  63                 return new SquidUpdate( $title->getSquidURLs() );
  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         public function merge( MergeableUpdate $update ) {
  74                 /** @var SquidUpdate $update */
  75                 Assert::parameterType( __CLASS__, $update, '$update' );
  76
  77                 $this->urls = array_merge( $this->urls, $update->urls );
  78         }
  79
  80         /**
  81          * Purges a list of Squids defined in $wgSquidServers.
  82          * $urlArr should contain the full URLs to purge as values
  83          * (example: $urlArr[] = 'http://my.host/something')
  84          * XXX report broken Squids per mail or log
  85          *
  86          * @param string[] $urlArr List of full URLs to purge
  87          */
  88         public static function purge( array $urlArr ) {
  89                 global $wgSquidServers, $wgHTCPRouting;
  90
  91                 if ( !$urlArr ) {
  92                         return;
  93                 }
  94
  95                 // Remove duplicate URLs from list
  96                 $urlArr = array_unique( $urlArr );
  97
  98                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) );
  99
 100                 if ( $wgHTCPRouting ) {
 101                         self::HTCPPurge( $urlArr );
 102                 }
 103
 104                 if ( $wgSquidServers ) {
 105                         // Maximum number of parallel connections per squid
 106                         $maxSocketsPerSquid = 8;
 107                         // Number of requests to send per socket
 108                         // 400 seems to be a good tradeoff, opening a socket takes a while
 109                         $urlsPerSocket = 400;
 110                         $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 111                         if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 112                                 $socketsPerSquid = $maxSocketsPerSquid;
 113                         }
 114
 115                         $pool = new SquidPurgeClientPool;
 116                         $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 117                         foreach ( $wgSquidServers as $server ) {
 118                                 foreach ( $chunks as $chunk ) {
 119                                         $client = new SquidPurgeClient( $server );
 120                                         foreach ( $chunk as $url ) {
 121                                                 $client->queuePurge( $url );
 122                                         }
 123                                         $pool->addClient( $client );
 124                                 }
 125                         }
 126
 127                         $pool->run();
 128                 }
 129         }
 130
 131         /**
 132          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 133          *
 134          * @throws MWException
 135          * @param string[] $urlArr Collection of URLs to purge
 136          */
 137         private static function HTCPPurge( array $urlArr ) {
 138                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 139
 140                 // HTCP CLR operation
 141                 $htcpOpCLR = 4;
 142
 143                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 144                 if ( !defined( "IPPROTO_IP" ) ) {
 145                         define( "IPPROTO_IP", 0 );
 146                         define( "IP_MULTICAST_LOOP", 34 );
 147                         define( "IP_MULTICAST_TTL", 33 );
 148                 }
 149
 150                 // pfsockopen doesn't work because we need set_sock_opt
 151                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 152                 if ( !$conn ) {
 153                         $errstr = socket_strerror( socket_last_error() );
 154                         wfDebugLog( 'squid', __METHOD__ .
 155                                 ": Error opening UDP socket: $errstr" );
 156
 157                         return;
 158                 }
 159
 160                 // Set socket options
 161                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 162                 if ( $wgHTCPMulticastTTL != 1 ) {
 163                         // Set multicast time to live (hop count) option on socket
 164                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 165                                 $wgHTCPMulticastTTL );
 166                 }
 167
 168                 // Get sequential trx IDs for packet loss counting
 169                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 170                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 171                 );
 172
 173                 foreach ( $urlArr as $url ) {
 174                         if ( !is_string( $url ) ) {
 175                                 throw new MWException( 'Bad purge URL' );
 176                         }
 177                         $url = self::expand( $url );
 178                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 179                         if ( !$conf ) {
 180                                 wfDebugLog( 'squid', __METHOD__ .
 181                                         "No HTCP rule configured for URL {$url} , skipping" );
 182                                 continue;
 183                         }
 184
 185                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 186                                 // Normalize single entries
 187                                 $conf = array( $conf );
 188                         }
 189                         foreach ( $conf as $subconf ) {
 190                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 191                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 192                                 }
 193                         }
 194
 195                         // Construct a minimal HTCP request diagram
 196                         // as per RFC 2756
 197                         // Opcode 'CLR', no response desired, no auth
 198                         $htcpTransID = current( $ids );
 199                         next( $ids );
 200
 201                         $htcpSpecifier = pack( 'na4na*na8n',
 202                                 4, 'HEAD', strlen( $url ), $url,
 203                                 8, 'HTTP/1.0', 0 );
 204
 205                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 206                         $htcpLen = 4 + $htcpDataLen + 2;
 207
 208                         // Note! Squid gets the bit order of the first
 209                         // word wrong, wrt the RFC. Apparently no other
 210                         // implementation exists, so adapt to Squid
 211                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 212                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 213                                 $htcpTransID, $htcpSpecifier, 2 );
 214
 215                         wfDebugLog( 'squid', __METHOD__ .
 216                                 "Purging URL $url via HTCP" );
 217                         foreach ( $conf as $subconf ) {
 218                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 219                                         $subconf['host'], $subconf['port'] );
 220                         }
 221                 }
 222         }
 223
 224         /**
 225          * Expand local URLs to fully-qualified URLs using the internal protocol
 226          * and host defined in $wgInternalServer. Input that's already fully-
 227          * qualified will be passed through unchanged.
 228          *
 229          * This is used to generate purge URLs that may be either local to the
 230          * main wiki or include a non-native host, such as images hosted on a
 231          * second internal server.
 232          *
 233          * Client functions should not need to call this.
 234          *
 235          * @param string $url
 236          * @return string
 237          */
 238         public static function expand( $url ) {
 239                 return wfExpandUrl( $url, PROTO_INTERNAL );
 240         }
 241
 242         /**
 243          * Find the HTCP routing rule to use for a given URL.
 244          * @param string $url URL to match
 245          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 246          * @return mixed Element of $rules that matched, or false if nothing matched
 247          */
 248         private static function getRuleForURL( $url, $rules ) {
 249                 foreach ( $rules as $regex => $routing ) {
 250                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 251                                 return $routing;
 252                         }
 253                 }
 254
 255                 return false;
 256         }
 257 }