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 appropriate CDN URLs given a title (or titles)
  28  * @ingroup Cache
  29  */
  30 class CdnCacheUpdate implements DeferrableUpdate, MergeableUpdate {
  31         /** @var string[] Collection of URLs to purge */
  32         protected $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 self $update */
  43                 Assert::parameterType( __CLASS__, $update, '$update' );
  44                 '@phan-var self $update';
  45
  46                 $this->urls = array_merge( $this->urls, $update->urls );
  47         }
  48
  49         /**
  50          * Create an update object from an array of Title objects, or a TitleArray object
  51          *
  52          * @param Traversable|Title[] $titles
  53          * @param string[] $urlArr
  54          * @return CdnCacheUpdate
  55          */
  56         public static function newFromTitles( $titles, $urlArr = [] ) {
  57                 ( new LinkBatch( $titles ) )->execute();
  58                 /** @var Title $title */
  59                 foreach ( $titles as $title ) {
  60                         $urlArr = array_merge( $urlArr, $title->getCdnUrls() );
  61                 }
  62
  63                 return new CdnCacheUpdate( $urlArr );
  64         }
  65
  66         /**
  67          * Purges the list of URLs passed to the constructor.
  68          */
  69         public function doUpdate() {
  70                 global $wgCdnReboundPurgeDelay;
  71
  72                 self::purge( $this->urls );
  73
  74                 if ( $wgCdnReboundPurgeDelay > 0 ) {
  75                         JobQueueGroup::singleton()->lazyPush( new CdnPurgeJob( [
  76                                 'urls' => $this->urls,
  77                                 'jobReleaseTimestamp' => time() + $wgCdnReboundPurgeDelay
  78                         ] ) );
  79                 }
  80         }
  81
  82         /**
  83          * Purges a list of CDN nodes defined in $wgCdnServers.
  84          * $urlArr should contain the full URLs to purge as values
  85          * (example: $urlArr[] = 'http://my.host/something')
  86          *
  87          * @param string[] $urlArr List of full URLs to purge
  88          */
  89         public static function purge( array $urlArr ) {
  90                 global $wgCdnServers, $wgHTCPRouting;
  91
  92                 if ( !$urlArr ) {
  93                         return;
  94                 }
  95
  96                 // Remove duplicate URLs from list
  97                 $urlArr = array_unique( $urlArr );
  98
  99                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) );
 100
 101                 // Reliably broadcast the purge to all edge nodes
 102                 $relayer = MediaWikiServices::getInstance()->getEventRelayerGroup()
 103                                         ->getRelayer( 'cdn-url-purges' );
 104                 $ts = microtime( true );
 105                 $relayer->notifyMulti(
 106                         'cdn-url-purges',
 107                         array_map(
 108                                 function ( $url ) use ( $ts ) {
 109                                         return [
 110                                                 'url' => $url,
 111                                                 'timestamp' => $ts,
 112                                         ];
 113                                 },
 114                                 $urlArr
 115                         )
 116                 );
 117
 118                 // Send lossy UDP broadcasting if enabled
 119                 if ( $wgHTCPRouting ) {
 120                         self::HTCPPurge( $urlArr );
 121                 }
 122
 123                 // Do direct server purges if enabled (this does not scale very well)
 124                 if ( $wgCdnServers ) {
 125                         // Maximum number of parallel connections per CDN
 126                         $maxSocketsPerCdn = 8;
 127                         // Number of requests to send per socket
 128                         // 400 seems to be a good tradeoff, opening a socket takes a while
 129                         $urlsPerSocket = 400;
 130                         $socketsPerCdn = ceil( count( $urlArr ) / $urlsPerSocket );
 131                         if ( $socketsPerCdn > $maxSocketsPerCdn ) {
 132                                 $socketsPerCdn = $maxSocketsPerCdn;
 133                         }
 134
 135                         $pool = new SquidPurgeClientPool;
 136                         $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerCdn ) );
 137                         foreach ( $wgCdnServers as $server ) {
 138                                 foreach ( $chunks as $chunk ) {
 139                                         $client = new SquidPurgeClient( $server );
 140                                         foreach ( $chunk as $url ) {
 141                                                 $client->queuePurge( $url );
 142                                         }
 143                                         $pool->addClient( $client );
 144                                 }
 145                         }
 146
 147                         $pool->run();
 148                 }
 149         }
 150
 151         /**
 152          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 153          *
 154          * @throws MWException
 155          * @param string[] $urlArr Collection of URLs to purge
 156          */
 157         private static function HTCPPurge( array $urlArr ) {
 158                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 159
 160                 // HTCP CLR operation
 161                 $htcpOpCLR = 4;
 162
 163                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 164                 if ( !defined( "IPPROTO_IP" ) ) {
 165                         define( "IPPROTO_IP", 0 );
 166                         define( "IP_MULTICAST_LOOP", 34 );
 167                         define( "IP_MULTICAST_TTL", 33 );
 168                 }
 169
 170                 // pfsockopen doesn't work because we need set_sock_opt
 171                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 172                 if ( !$conn ) {
 173                         $errstr = socket_strerror( socket_last_error() );
 174                         wfDebugLog( 'squid', __METHOD__ .
 175                                 ": Error opening UDP socket: $errstr" );
 176
 177                         return;
 178                 }
 179
 180                 // Set socket options
 181                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 182                 if ( $wgHTCPMulticastTTL != 1 ) {
 183                         // Set multicast time to live (hop count) option on socket
 184                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 185                                 $wgHTCPMulticastTTL );
 186                 }
 187
 188                 // Get sequential trx IDs for packet loss counting
 189                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 190                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 191                 );
 192
 193                 foreach ( $urlArr as $url ) {
 194                         if ( !is_string( $url ) ) {
 195                                 throw new MWException( 'Bad purge URL' );
 196                         }
 197                         $url = self::expand( $url );
 198                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 199                         if ( !$conf ) {
 200                                 wfDebugLog( 'squid', __METHOD__ .
 201                                         "No HTCP rule configured for URL {$url} , skipping" );
 202                                 continue;
 203                         }
 204
 205                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 206                                 // Normalize single entries
 207                                 $conf = [ $conf ];
 208                         }
 209                         foreach ( $conf as $subconf ) {
 210                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 211                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 212                                 }
 213                         }
 214
 215                         // Construct a minimal HTCP request diagram
 216                         // as per RFC 2756
 217                         // Opcode 'CLR', no response desired, no auth
 218                         $htcpTransID = current( $ids );
 219                         next( $ids );
 220
 221                         $htcpSpecifier = pack( 'na4na*na8n',
 222                                 4, 'HEAD', strlen( $url ), $url,
 223                                 8, 'HTTP/1.0', 0 );
 224
 225                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 226                         $htcpLen = 4 + $htcpDataLen + 2;
 227
 228                         // Note! Squid gets the bit order of the first
 229                         // word wrong, wrt the RFC. Apparently no other
 230                         // implementation exists, so adapt to Squid
 231                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 232                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 233                                 $htcpTransID, $htcpSpecifier, 2 );
 234
 235                         wfDebugLog( 'squid', __METHOD__ .
 236                                 "Purging URL $url via HTCP" );
 237                         foreach ( $conf as $subconf ) {
 238                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 239                                         $subconf['host'], $subconf['port'] );
 240                         }
 241                 }
 242         }
 243
 244         /**
 245          * Expand local URLs to fully-qualified URLs using the internal protocol
 246          * and host defined in $wgInternalServer. Input that's already fully-
 247          * qualified will be passed through unchanged.
 248          *
 249          * This is used to generate purge URLs that may be either local to the
 250          * main wiki or include a non-native host, such as images hosted on a
 251          * second internal server.
 252          *
 253          * Client functions should not need to call this.
 254          *
 255          * @param string $url
 256          * @return string
 257          */
 258         public static function expand( $url ) {
 259                 return wfExpandUrl( $url, PROTO_INTERNAL );
 260         }
 261
 262         /**
 263          * Find the HTCP routing rule to use for a given URL.
 264          * @param string $url URL to match
 265          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 266          * @return mixed Element of $rules that matched, or false if nothing matched
 267          */
 268         private static function getRuleForURL( $url, $rules ) {
 269                 foreach ( $rules as $regex => $routing ) {
 270                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 271                                 return $routing;
 272                         }
 273                 }
 274
 275                 return false;
 276         }
 277 }