includes/deferred/CdnCacheUpdate.php

   1 <?php
   2 /**
   3  * CDN 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  */
  22
  23 use Wikimedia\Assert\Assert;
  24 use MediaWiki\MediaWikiServices;
  25
  26 /**
  27  * Handles purging the appropriate CDN objects given a list of URLs or Title instances
  28  * @ingroup Cache
  29  */
  30 class CdnCacheUpdate implements DeferrableUpdate, MergeableUpdate {
  31         /** @var string[] Collection of URLs to purge */
  32         private $urls = [];
  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         public function merge( MergeableUpdate $update ) {
  42                 /** @var CdnCacheUpdate $update */
  43                 Assert::parameterType( __CLASS__, $update, '$update' );
  44
  45                 $this->urls = array_merge( $this->urls, $update->urls );
  46         }
  47
  48         /**
  49          * Create an update object from an array of Title objects, or a TitleArray object
  50          *
  51          * @param Traversable|Title[] $titles
  52          * @param string[] $urlArr
  53          * @return CdnCacheUpdate
  54          */
  55         public static function newFromTitles( $titles, $urlArr = [] ) {
  56                 ( new LinkBatch( $titles ) )->execute();
  57                 /** @var Title $title */
  58                 foreach ( $titles as $title ) {
  59                         $urlArr = array_merge( $urlArr, $title->getCdnUrls() );
  60                 }
  61
  62                 return new self( $urlArr );
  63         }
  64
  65         public function doUpdate() {
  66                 global $wgCdnReboundPurgeDelay;
  67
  68                 self::purge( $this->urls );
  69
  70                 if ( $wgCdnReboundPurgeDelay > 0 ) {
  71                         JobQueueGroup::singleton()->lazyPush( new CdnPurgeJob( [
  72                                 'urls' => $this->urls,
  73                                 'jobReleaseTimestamp' => time() + $wgCdnReboundPurgeDelay
  74                         ] ) );
  75                 }
  76         }
  77
  78         /**
  79          * Purges a list of CDN nodes defined in $wgCdnServers.
  80          * $urlArr should contain the full URLs to purge as values
  81          * (example: $urlArr[] = 'http://my.host/something')
  82          *
  83          * @param string[] $urlArr List of full URLs to purge
  84          */
  85         public static function purge( array $urlArr ) {
  86                 global $wgCdnServers, $wgHTCPRouting;
  87
  88                 if ( !$urlArr ) {
  89                         return;
  90                 }
  91
  92                 // Remove duplicate URLs from list
  93                 $urlArr = array_unique( $urlArr );
  94
  95                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) );
  96
  97                 // Reliably broadcast the purge to all edge nodes
  98                 $ts = microtime( true );
  99                 $relayerGroup = MediaWikiServices::getInstance()->getEventRelayerGroup();
 100                 $relayerGroup->getRelayer( 'cdn-url-purges' )->notifyMulti(
 101                         'cdn-url-purges',
 102                         array_map(
 103                                 function ( $url ) use ( $ts ) {
 104                                         return [
 105                                                 'url' => $url,
 106                                                 'timestamp' => $ts,
 107                                         ];
 108                                 },
 109                                 $urlArr
 110                         )
 111                 );
 112
 113                 // Send lossy UDP broadcasting if enabled
 114                 if ( $wgHTCPRouting ) {
 115                         self::HTCPPurge( $urlArr );
 116                 }
 117
 118                 // Do direct server purges if enabled (this does not scale very well)
 119                 if ( $wgCdnServers ) {
 120                         // Maximum number of parallel connections per CDN
 121                         $maxSocketsPerCdn = 8;
 122                         // Number of requests to send per socket
 123                         // 400 seems to be a good tradeoff, opening a socket takes a while
 124                         $urlsPerSocket = 400;
 125                         $socketsPerCdn = ceil( count( $urlArr ) / $urlsPerSocket );
 126                         if ( $socketsPerCdn > $maxSocketsPerCdn ) {
 127                                 $socketsPerCdn = $maxSocketsPerCdn;
 128                         }
 129
 130                         $pool = new SquidPurgeClientPool;
 131                         $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerCdn ) );
 132                         foreach ( $wgCdnServers as $server ) {
 133                                 foreach ( $chunks as $chunk ) {
 134                                         $client = new SquidPurgeClient( $server );
 135                                         foreach ( $chunk as $url ) {
 136                                                 $client->queuePurge( $url );
 137                                         }
 138                                         $pool->addClient( $client );
 139                                 }
 140                         }
 141
 142                         $pool->run();
 143                 }
 144         }
 145
 146         /**
 147          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 148          *
 149          * @throws MWException
 150          * @param string[] $urlArr Collection of URLs to purge
 151          */
 152         private static function HTCPPurge( array $urlArr ) {
 153                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 154
 155                 // HTCP CLR operation
 156                 $htcpOpCLR = 4;
 157
 158                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 159                 if ( !defined( "IPPROTO_IP" ) ) {
 160                         define( "IPPROTO_IP", 0 );
 161                         define( "IP_MULTICAST_LOOP", 34 );
 162                         define( "IP_MULTICAST_TTL", 33 );
 163                 }
 164
 165                 // pfsockopen doesn't work because we need set_sock_opt
 166                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 167                 if ( !$conn ) {
 168                         $errstr = socket_strerror( socket_last_error() );
 169                         wfDebugLog( 'squid', __METHOD__ .
 170                                 ": Error opening UDP socket: $errstr" );
 171
 172                         return;
 173                 }
 174
 175                 // Set socket options
 176                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 177                 if ( $wgHTCPMulticastTTL != 1 ) {
 178                         // Set multicast time to live (hop count) option on socket
 179                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 180                                 $wgHTCPMulticastTTL );
 181                 }
 182
 183                 // Get sequential trx IDs for packet loss counting
 184                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 185                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 186                 );
 187
 188                 foreach ( $urlArr as $url ) {
 189                         if ( !is_string( $url ) ) {
 190                                 throw new MWException( 'Bad purge URL' );
 191                         }
 192                         $url = self::expand( $url );
 193                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 194                         if ( !$conf ) {
 195                                 wfDebugLog( 'squid', __METHOD__ .
 196                                         "No HTCP rule configured for URL {$url} , skipping" );
 197                                 continue;
 198                         }
 199
 200                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 201                                 // Normalize single entries
 202                                 $conf = [ $conf ];
 203                         }
 204                         foreach ( $conf as $subconf ) {
 205                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 206                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 207                                 }
 208                         }
 209
 210                         // Construct a minimal HTCP request diagram
 211                         // as per RFC 2756
 212                         // Opcode 'CLR', no response desired, no auth
 213                         $htcpTransID = current( $ids );
 214                         next( $ids );
 215
 216                         $htcpSpecifier = pack( 'na4na*na8n',
 217                                 4, 'HEAD', strlen( $url ), $url,
 218                                 8, 'HTTP/1.0', 0 );
 219
 220                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 221                         $htcpLen = 4 + $htcpDataLen + 2;
 222
 223                         // Note! Squid gets the bit order of the first
 224                         // word wrong, wrt the RFC. Apparently no other
 225                         // implementation exists, so adapt to Squid
 226                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 227                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 228                                 $htcpTransID, $htcpSpecifier, 2 );
 229
 230                         wfDebugLog( 'squid', __METHOD__ .
 231                                 "Purging URL $url via HTCP" );
 232                         foreach ( $conf as $subconf ) {
 233                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 234                                         $subconf['host'], $subconf['port'] );
 235                         }
 236                 }
 237         }
 238
 239         /**
 240          * Expand local URLs to fully-qualified URLs using the internal protocol
 241          * and host defined in $wgInternalServer. Input that's already fully-
 242          * qualified will be passed through unchanged.
 243          *
 244          * This is used to generate purge URLs that may be either local to the
 245          * main wiki or include a non-native host, such as images hosted on a
 246          * second internal server.
 247          *
 248          * Client functions should not need to call this.
 249          *
 250          * @param string $url
 251          * @return string
 252          */
 253         public static function expand( $url ) {
 254                 return wfExpandUrl( $url, PROTO_INTERNAL );
 255         }
 256
 257         /**
 258          * Find the HTCP routing rule to use for a given URL.
 259          * @param string $url URL to match
 260          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 261          * @return mixed Element of $rules that matched, or false if nothing matched
 262          */
 263         private static function getRuleForURL( $url, $rules ) {
 264                 foreach ( $rules as $regex => $routing ) {
 265                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 266                                 return $routing;
 267                         }
 268                 }
 269
 270                 return false;
 271         }
 272 }