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  * @ingroup Cache
  22  */
  23
  24 use Wikimedia\Assert\Assert;
  25 use MediaWiki\MediaWikiServices;
  26
  27 /**
  28  * Handles purging appropriate CDN URLs given a title (or titles)
  29  * @ingroup Cache
  30  */
  31 class CdnCacheUpdate implements DeferrableUpdate, MergeableUpdate {
  32         /** @var string[] Collection of URLs to purge */
  33         protected $urls = [];
  34
  35         /**
  36          * @param string[] $urlArr Collection of URLs to purge
  37          */
  38         public function __construct( array $urlArr ) {
  39                 $this->urls = $urlArr;
  40         }
  41
  42         public function merge( MergeableUpdate $update ) {
  43                 /** @var CdnCacheUpdate $update */
  44                 Assert::parameterType( __CLASS__, $update, '$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          * @param Title $title
  68          * @return CdnCacheUpdate
  69          * @deprecated since 1.27
  70          */
  71         public static function newSimplePurge( Title $title ) {
  72                 return new CdnCacheUpdate( $title->getCdnUrls() );
  73         }
  74
  75         /**
  76          * Purges the list of URLs passed to the constructor.
  77          */
  78         public function doUpdate() {
  79                 global $wgCdnReboundPurgeDelay;
  80
  81                 self::purge( $this->urls );
  82
  83                 if ( $wgCdnReboundPurgeDelay > 0 ) {
  84                         JobQueueGroup::singleton()->lazyPush( new CdnPurgeJob(
  85                                 Title::makeTitle( NS_SPECIAL, 'Badtitle/' . __CLASS__ ),
  86                                 [
  87                                         'urls' => $this->urls,
  88                                         'jobReleaseTimestamp' => time() + $wgCdnReboundPurgeDelay
  89                                 ]
  90                         ) );
  91                 }
  92         }
  93
  94         /**
  95          * Purges a list of CDN nodes defined in $wgSquidServers.
  96          * $urlArr should contain the full URLs to purge as values
  97          * (example: $urlArr[] = 'http://my.host/something')
  98          *
  99          * @param string[] $urlArr List of full URLs to purge
 100          */
 101         public static function purge( array $urlArr ) {
 102                 global $wgSquidServers, $wgHTCPRouting;
 103
 104                 if ( !$urlArr ) {
 105                         return;
 106                 }
 107
 108                 // Remove duplicate URLs from list
 109                 $urlArr = array_unique( $urlArr );
 110
 111                 wfDebugLog( 'squid', __METHOD__ . ': ' . implode( ' ', $urlArr ) );
 112
 113                 // Reliably broadcast the purge to all edge nodes
 114                 $relayer = MediaWikiServices::getInstance()->getEventRelayerGroup()
 115                                         ->getRelayer( 'cdn-url-purges' );
 116                 $ts = microtime( true );
 117                 $relayer->notifyMulti(
 118                         'cdn-url-purges',
 119                         array_map(
 120                                 function ( $url ) use ( $ts ) {
 121                                         return [
 122                                                 'url' => $url,
 123                                                 'timestamp' => $ts,
 124                                         ];
 125                                 },
 126                                 $urlArr
 127                         )
 128                 );
 129
 130                 // Send lossy UDP broadcasting if enabled
 131                 if ( $wgHTCPRouting ) {
 132                         self::HTCPPurge( $urlArr );
 133                 }
 134
 135                 // Do direct server purges if enabled (this does not scale very well)
 136                 if ( $wgSquidServers ) {
 137                         // Maximum number of parallel connections per squid
 138                         $maxSocketsPerSquid = 8;
 139                         // Number of requests to send per socket
 140                         // 400 seems to be a good tradeoff, opening a socket takes a while
 141                         $urlsPerSocket = 400;
 142                         $socketsPerSquid = ceil( count( $urlArr ) / $urlsPerSocket );
 143                         if ( $socketsPerSquid > $maxSocketsPerSquid ) {
 144                                 $socketsPerSquid = $maxSocketsPerSquid;
 145                         }
 146
 147                         $pool = new SquidPurgeClientPool;
 148                         $chunks = array_chunk( $urlArr, ceil( count( $urlArr ) / $socketsPerSquid ) );
 149                         foreach ( $wgSquidServers as $server ) {
 150                                 foreach ( $chunks as $chunk ) {
 151                                         $client = new SquidPurgeClient( $server );
 152                                         foreach ( $chunk as $url ) {
 153                                                 $client->queuePurge( $url );
 154                                         }
 155                                         $pool->addClient( $client );
 156                                 }
 157                         }
 158
 159                         $pool->run();
 160                 }
 161         }
 162
 163         /**
 164          * Send Hyper Text Caching Protocol (HTCP) CLR requests.
 165          *
 166          * @throws MWException
 167          * @param string[] $urlArr Collection of URLs to purge
 168          */
 169         private static function HTCPPurge( array $urlArr ) {
 170                 global $wgHTCPRouting, $wgHTCPMulticastTTL;
 171
 172                 // HTCP CLR operation
 173                 $htcpOpCLR = 4;
 174
 175                 // @todo FIXME: PHP doesn't support these socket constants (include/linux/in.h)
 176                 if ( !defined( "IPPROTO_IP" ) ) {
 177                         define( "IPPROTO_IP", 0 );
 178                         define( "IP_MULTICAST_LOOP", 34 );
 179                         define( "IP_MULTICAST_TTL", 33 );
 180                 }
 181
 182                 // pfsockopen doesn't work because we need set_sock_opt
 183                 $conn = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
 184                 if ( !$conn ) {
 185                         $errstr = socket_strerror( socket_last_error() );
 186                         wfDebugLog( 'squid', __METHOD__ .
 187                                 ": Error opening UDP socket: $errstr" );
 188
 189                         return;
 190                 }
 191
 192                 // Set socket options
 193                 socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_LOOP, 0 );
 194                 if ( $wgHTCPMulticastTTL != 1 ) {
 195                         // Set multicast time to live (hop count) option on socket
 196                         socket_set_option( $conn, IPPROTO_IP, IP_MULTICAST_TTL,
 197                                 $wgHTCPMulticastTTL );
 198                 }
 199
 200                 // Get sequential trx IDs for packet loss counting
 201                 $ids = UIDGenerator::newSequentialPerNodeIDs(
 202                         'squidhtcppurge', 32, count( $urlArr ), UIDGenerator::QUICK_VOLATILE
 203                 );
 204
 205                 foreach ( $urlArr as $url ) {
 206                         if ( !is_string( $url ) ) {
 207                                 throw new MWException( 'Bad purge URL' );
 208                         }
 209                         $url = self::expand( $url );
 210                         $conf = self::getRuleForURL( $url, $wgHTCPRouting );
 211                         if ( !$conf ) {
 212                                 wfDebugLog( 'squid', __METHOD__ .
 213                                         "No HTCP rule configured for URL {$url} , skipping" );
 214                                 continue;
 215                         }
 216
 217                         if ( isset( $conf['host'] ) && isset( $conf['port'] ) ) {
 218                                 // Normalize single entries
 219                                 $conf = [ $conf ];
 220                         }
 221                         foreach ( $conf as $subconf ) {
 222                                 if ( !isset( $subconf['host'] ) || !isset( $subconf['port'] ) ) {
 223                                         throw new MWException( "Invalid HTCP rule for URL $url\n" );
 224                                 }
 225                         }
 226
 227                         // Construct a minimal HTCP request diagram
 228                         // as per RFC 2756
 229                         // Opcode 'CLR', no response desired, no auth
 230                         $htcpTransID = current( $ids );
 231                         next( $ids );
 232
 233                         $htcpSpecifier = pack( 'na4na*na8n',
 234                                 4, 'HEAD', strlen( $url ), $url,
 235                                 8, 'HTTP/1.0', 0 );
 236
 237                         $htcpDataLen = 8 + 2 + strlen( $htcpSpecifier );
 238                         $htcpLen = 4 + $htcpDataLen + 2;
 239
 240                         // Note! Squid gets the bit order of the first
 241                         // word wrong, wrt the RFC. Apparently no other
 242                         // implementation exists, so adapt to Squid
 243                         $htcpPacket = pack( 'nxxnCxNxxa*n',
 244                                 $htcpLen, $htcpDataLen, $htcpOpCLR,
 245                                 $htcpTransID, $htcpSpecifier, 2 );
 246
 247                         wfDebugLog( 'squid', __METHOD__ .
 248                                 "Purging URL $url via HTCP" );
 249                         foreach ( $conf as $subconf ) {
 250                                 socket_sendto( $conn, $htcpPacket, $htcpLen, 0,
 251                                         $subconf['host'], $subconf['port'] );
 252                         }
 253                 }
 254         }
 255
 256         /**
 257          * Expand local URLs to fully-qualified URLs using the internal protocol
 258          * and host defined in $wgInternalServer. Input that's already fully-
 259          * qualified will be passed through unchanged.
 260          *
 261          * This is used to generate purge URLs that may be either local to the
 262          * main wiki or include a non-native host, such as images hosted on a
 263          * second internal server.
 264          *
 265          * Client functions should not need to call this.
 266          *
 267          * @param string $url
 268          * @return string
 269          */
 270         public static function expand( $url ) {
 271                 return wfExpandUrl( $url, PROTO_INTERNAL );
 272         }
 273
 274         /**
 275          * Find the HTCP routing rule to use for a given URL.
 276          * @param string $url URL to match
 277          * @param array $rules Array of rules, see $wgHTCPRouting for format and behavior
 278          * @return mixed Element of $rules that matched, or false if nothing matched
 279          */
 280         private static function getRuleForURL( $url, $rules ) {
 281                 foreach ( $rules as $regex => $routing ) {
 282                         if ( $regex === '' || preg_match( $regex, $url ) ) {
 283                                 return $routing;
 284                         }
 285                 }
 286
 287                 return false;
 288         }
 289 }
 290
 291 /**
 292  * @deprecated since 1.27
 293  */
 294 class SquidUpdate extends CdnCacheUpdate {
 295         // Keep class name for b/c
 296 }