maintenance/syncFileBackend.php

   1 <?php
   2 /**
   3  * Sync one file backend to another based on the journal of later.
   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 Maintenance
  22  */
  23
  24 require_once( __DIR__ . '/Maintenance.php' );
  25
  26 /**
  27  * Maintenance script that syncs one file backend to another based on
  28  * the journal of later.
  29  *
  30  * @ingroup Maintenance
  31  */
  32 class SyncFileBackend extends Maintenance {
  33         public function __construct() {
  34                 parent::__construct();
  35                 $this->mDescription = "Sync one file backend with another using the journal";
  36                 $this->addOption( 'src', 'Name of backend to sync from', true, true );
  37                 $this->addOption( 'dst', 'Name of destination backend to sync', true, true );
  38                 $this->addOption( 'start', 'Starting journal ID', false, true );
  39                 $this->addOption( 'end', 'Ending journal ID', false, true );
  40                 $this->addOption( 'posdir', 'Directory to read/record journal positions', false, true );
  41                 $this->addOption( 'verbose', 'Verbose mode', false, false, 'v' );
  42                 $this->setBatchSize( 50 );
  43         }
  44
  45         public function execute() {
  46                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  47                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  48
  49                 $posDir = $this->getOption( 'posdir' );
  50                 $posFile = $posDir ? $posDir . '/' . wfWikiID() : false;
  51
  52                 $start = $this->getOption( 'start', 0 );
  53                 if ( !$start && $posFile && is_dir( $posDir ) ) {
  54                         $start = is_file( $posFile )
  55                                 ? (int)trim( file_get_contents( $posFile ) )
  56                                 : 0;
  57                         ++$start; // we already did this ID, start with the next one
  58                         $startFromPosFile = true;
  59                 } else {
  60                         $startFromPosFile = false;
  61                 }
  62                 $end = $this->getOption( 'end', INF );
  63
  64                 $this->output( "Synchronizing backend '{$dst->getName()}' to '{$src->getName()}'...\n" );
  65                 $this->output( "Starting journal position is $start.\n" );
  66                 if ( is_finite( $end ) ) {
  67                         $this->output( "Ending journal position is $end.\n" );
  68                 }
  69
  70                 // Actually sync the dest backend with the reference backend
  71                 $lastOKPos = $this->syncBackends( $src, $dst, $start, $end );
  72
  73                 // Update the sync position file
  74                 if ( $startFromPosFile && $lastOKPos >= $start ) { // successfully advanced
  75                         if ( file_put_contents( $posFile, $lastOKPos, LOCK_EX ) !== false ) {
  76                                 $this->output( "Updated journal position file.\n" );
  77                         } else {
  78                                 $this->output( "Could not update journal position file.\n" );
  79                         }
  80                 }
  81
  82                 if ( $lastOKPos === false ) {
  83                         if ( !$start ) {
  84                                 $this->output( "No journal entries found.\n" );
  85                         } else {
  86                                 $this->output( "No new journal entries found.\n" );
  87                         }
  88                 } else {
  89                         $this->output( "Stopped synchronization at journal position $lastOKPos.\n" );
  90                 }
  91
  92                 if ( $this->isQuiet() ) {
  93                         print $lastOKPos; // give a single machine-readable number
  94                 }
  95         }
  96
  97         /**
  98          * Sync $dst backend to $src backend based on the $src logs given after $start.
  99          * Returns the journal entry ID this advanced to and handled (inclusive).
 100          *
 101          * @param $src FileBackend
 102          * @param $dst FileBackend
 103          * @param $start integer Starting journal position
 104          * @param $end integer Starting journal position
 105          * @return integer|false Journal entry ID or false if there are none
 106          */
 107         protected function syncBackends( FileBackend $src, FileBackend $dst, $start, $end ) {
 108                 $lastOKPos = 0; // failed
 109                 $first = true; // first batch
 110
 111                 if ( $start > $end ) { // sanity
 112                         $this->error( "Error: given starting ID greater than ending ID.", 1 );
 113                 }
 114
 115                 do {
 116                         $limit = min( $this->mBatchSize, $end - $start + 1 ); // don't go pass ending ID
 117                         $this->output( "Doing id $start to " . ( $start + $limit - 1 ) . "...\n" );
 118
 119                         $entries = $src->getJournal()->getChangeEntries( $start, $limit, $next );
 120                         $start = $next; // start where we left off next time
 121                         if ( $first && !count( $entries ) ) {
 122                                 return false; // nothing to do
 123                         }
 124                         $first = false;
 125
 126                         $lastPosInBatch = 0;
 127                         $pathsInBatch = array(); // changed paths
 128                         foreach ( $entries as $entry ) {
 129                                 if ( $entry['op'] !== 'null' ) { // null ops are just for reference
 130                                         $pathsInBatch[$entry['path']] = 1; // remove duplicates
 131                                 }
 132                                 $lastPosInBatch = $entry['id'];
 133                         }
 134
 135                         $status = $this->syncFileBatch( array_keys( $pathsInBatch ), $src, $dst );
 136                         if ( $status->isOK() ) {
 137                                 $lastOKPos = max( $lastOKPos, $lastPosInBatch );
 138                         } else {
 139                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 140                                 break; // no gaps; everything up to $lastPos must be OK
 141                         }
 142
 143                         if ( !$start ) {
 144                                 $this->output( "End of journal entries.\n" );
 145                         }
 146                 } while ( $start && $start <= $end );
 147
 148                 return $lastOKPos;
 149         }
 150
 151         /**
 152          * Sync particular files of backend $src to the corresponding $dst backend files
 153          *
 154          * @param $paths Array
 155          * @param $src FileBackend
 156          * @param $dst FileBackend
 157          * @return Status
 158          */
 159         protected function syncFileBatch( array $paths, FileBackend $src, FileBackend $dst ) {
 160                 $status = Status::newGood();
 161                 if ( !count( $paths ) ) {
 162                         return $status; // nothing to do
 163                 }
 164
 165                 // Source: convert internal backend names (FileBackendMultiWrite) to the public one
 166                 $sPaths = $this->replaceNamePaths( $paths, $src );
 167                 // Destination: get corresponding path name
 168                 $dPaths = $this->replaceNamePaths( $paths, $dst );
 169
 170                 // Lock the live backend paths from modification
 171                 $sLock = $src->getScopedFileLocks( $sPaths, LockManager::LOCK_UW, $status );
 172                 $eLock = $dst->getScopedFileLocks( $dPaths, LockManager::LOCK_EX, $status );
 173                 if ( !$status->isOK() ) {
 174                         return $status;
 175                 }
 176
 177                 $ops = array();
 178                 $fsFiles = array();
 179                 foreach ( $sPaths as $i => $sPath ) {
 180                         $dPath = $dPaths[$i]; // destination
 181                         $sExists = $src->fileExists( array( 'src' => $sPath, 'latest' => 1 ) );
 182                         if ( $sExists === true ) { // exists in source
 183                                 if ( $this->filesAreSame( $src, $dst, $sPath, $dPath ) ) {
 184                                         continue; // avoid local copies for non-FS backends
 185                                 }
 186                                 // Note: getLocalReference() is fast for FS backends
 187                                 $fsFile = $src->getLocalReference( array( 'src' => $sPath, 'latest' => 1 ) );
 188                                 if ( !$fsFile ) {
 189                                         $this->error( "Unable to sync '$dPath': could not get local copy." );
 190                                         $status->fatal( 'backend-fail-internal', $src->getName() );
 191                                         return $status;
 192                                 }
 193                                 $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 194                                 // Note: prepare() is usually fast for key/value backends
 195                                 $status->merge( $dst->prepare( array(
 196                                         'dir' => dirname( $dPath ), 'bypassReadOnly' => 1 ) ) );
 197                                 if ( !$status->isOK() ) {
 198                                         return $status;
 199                                 }
 200                                 $ops[] = array( 'op' => 'store',
 201                                         'src' => $fsFile->getPath(), 'dst' => $dPath, 'overwrite' => 1 );
 202                         } elseif ( $sExists === false ) { // does not exist in source
 203                                 $ops[] = array( 'op' => 'delete', 'src' => $dPath, 'ignoreMissingSource' => 1 );
 204                         } else { // error
 205                                 $this->error( "Unable to sync '$dPath': could not stat file." );
 206                                 $status->fatal( 'backend-fail-internal', $src->getName() );
 207                                 return $status;
 208                         }
 209                 }
 210
 211                 $t_start = microtime( true );
 212                 $status->merge( $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) ) );
 213                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 214                 if ( $status->isOK() && $this->getOption( 'verbose' ) ) {
 215                         $this->output( "Synchronized these file(s) [{$ellapsed_ms}ms]:\n" .
 216                                 implode( "\n", $dPaths ) . "\n" );
 217                 }
 218
 219                 return $status;
 220         }
 221
 222         /**
 223          * Substitute the backend name of storage paths with that of a given one
 224          *
 225          * @param $paths Array|string List of paths or single string path
 226          * @return Array|string
 227          */
 228         protected function replaceNamePaths( $paths, FileBackend $backend ) {
 229                 return preg_replace(
 230                         '!^mwstore://([^/]+)!',
 231                         StringUtils::escapeRegexReplacement( "mwstore://" . $backend->getName() ),
 232                         $paths // string or array
 233                 );
 234         }
 235
 236         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 237                 return (
 238                         ( $src->getFileSize( array( 'src' => $sPath ) )
 239                                 === $dst->getFileSize( array( 'src' => $dPath ) ) // short-circuit
 240                         ) && ( $src->getFileSha1Base36( array( 'src' => $sPath ) )
 241                                 === $dst->getFileSha1Base36( array( 'src' => $dPath ) )
 242                         )
 243                 );
 244         }
 245 }
 246
 247 $maintClass = "SyncFileBackend";
 248 require_once( RUN_MAINTENANCE_IF_MAIN );