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