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  * @ingroup Maintenance
  21  */
  22
  23 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  24
  25 class SyncFileBackend extends Maintenance {
  26         public function __construct() {
  27                 parent::__construct();
  28                 $this->mDescription = "Sync one file backend with another using the journal";
  29                 $this->addOption( 'src', 'Name of backend to sync from', true, true );
  30                 $this->addOption( 'dst', 'Name of destination backend to sync', true, true );
  31                 $this->addOption( 'start', 'Starting journal ID', false, true );
  32                 $this->addOption( 'end', 'Ending journal ID', false, true );
  33                 $this->addOption( 'posdir', 'Directory to read/record journal positions', false, true );
  34                 $this->addOption( 'verbose', 'Verbose mode', false, false, 'v' );
  35                 $this->setBatchSize( 50 );
  36         }
  37
  38         public function execute() {
  39                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  40                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  41
  42                 $posDir = $this->getOption( 'posdir' );
  43                 $posFile = $posDir ? $posDir . '/' . wfWikiID() : false;
  44
  45                 $start = $this->getOption( 'start', 0 );
  46                 if ( !$start && $posFile && is_dir( $posDir ) ) {
  47                         $start = is_file( $posFile )
  48                                 ? (int)trim( file_get_contents( $posFile ) )
  49                                 : 0;
  50                         ++$start; // we already did this ID, start with the next one
  51                         $startFromPosFile = true;
  52                 } else {
  53                         $startFromPosFile = false;
  54                 }
  55                 $end = $this->getOption( 'end', INF );
  56
  57                 $this->output( "Synchronizing backend '{$dst->getName()}' to '{$src->getName()}'...\n" );
  58                 $this->output( "Starting journal position is $start.\n" );
  59                 if ( is_finite( $end ) ) {
  60                         $this->output( "Ending journal position is $end.\n" );
  61                 }
  62
  63                 // Actually sync the dest backend with the reference backend
  64                 $lastOKPos = $this->syncBackends( $src, $dst, $start, $end );
  65
  66                 // Update the sync position file
  67                 if ( $startFromPosFile && $lastOKPos >= $start ) { // successfully advanced
  68                         if ( file_put_contents( $posFile, $lastOKPos, LOCK_EX ) !== false ) {
  69                                 $this->output( "Updated journal position file.\n" );
  70                         } else {
  71                                 $this->output( "Could not update journal position file.\n" );
  72                         }
  73                 }
  74
  75                 if ( $lastOKPos === false ) {
  76                         if ( !$start ) {
  77                                 $this->output( "No journal entries found.\n" );
  78                         } else {
  79                                 $this->output( "No new journal entries found.\n" );
  80                         }
  81                 } else {
  82                         $this->output( "Stopped synchronization at journal position $lastOKPos.\n" );
  83                 }
  84
  85                 if ( $this->isQuiet() ) {
  86                         print $lastOKPos; // give a single machine-readable number
  87                 }
  88         }
  89
  90         /**
  91          * Sync $dst backend to $src backend based on the $src logs given after $start.
  92          * Returns the journal entry ID this advanced to and handled (inclusive).
  93          *
  94          * @param $src FileBackend
  95          * @param $dst FileBackend
  96          * @param $start integer Starting journal position
  97          * @param $end integer Starting journal position
  98          * @return integer|false Journal entry ID or false if there are none
  99          */
 100         protected function syncBackends( FileBackend $src, FileBackend $dst, $start, $end ) {
 101                 $lastOKPos = 0; // failed
 102                 $first = true; // first batch
 103
 104                 if ( $start > $end ) { // sanity
 105                         $this->error( "Error: given starting ID greater than ending ID.", 1 );
 106                 }
 107
 108                 do {
 109                         $limit = min( $this->mBatchSize, $end - $start + 1 ); // don't go pass ending ID
 110                         $this->output( "Doing id $start to " . ( $start + $limit - 1 ) . "...\n" );
 111
 112                         $entries = $src->getJournal()->getChangeEntries( $start, $limit, $next );
 113                         $start = $next; // start where we left off next time
 114                         if ( $first && !count( $entries ) ) {
 115                                 return false; // nothing to do
 116                         }
 117                         $first = false;
 118
 119                         $lastPosInBatch = 0;
 120                         $pathsInBatch = array(); // changed paths
 121                         foreach ( $entries as $entry ) {
 122                                 if ( $entry['op'] !== 'null' ) { // null ops are just for reference
 123                                         $pathsInBatch[$entry['path']] = 1; // remove duplicates
 124                                 }
 125                                 $lastPosInBatch = $entry['id'];
 126                         }
 127
 128                         $status = $this->syncFileBatch( array_keys( $pathsInBatch ), $src, $dst );
 129                         if ( $status->isOK() ) {
 130                                 $lastOKPos = max( $lastOKPos, $lastPosInBatch );
 131                         } else {
 132                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 133                                 break; // no gaps; everything up to $lastPos must be OK
 134                         }
 135
 136                         if ( !$start ) {
 137                                 $this->output( "End of journal entries.\n" );
 138                         }
 139                 } while ( $start && $start <= $end );
 140
 141                 return $lastOKPos;
 142         }
 143
 144         /**
 145          * Sync particular files of backend $src to the corresponding $dst backend files
 146          *
 147          * @param $paths Array
 148          * @param $src FileBackend
 149          * @param $dst FileBackend
 150          * @return Status
 151          */
 152         protected function syncFileBatch( array $paths, FileBackend $src, FileBackend $dst ) {
 153                 $status = Status::newGood();
 154                 if ( !count( $paths ) ) {
 155                         return $status; // nothing to do
 156                 }
 157
 158                 // Source: convert internal backend names (FileBackendMultiWrite) to the public one
 159                 $sPaths = $this->replaceNamePaths( $paths, $src );
 160                 // Destination: get corresponding path name
 161                 $dPaths = $this->replaceNamePaths( $paths, $dst );
 162
 163                 // Lock the live backend paths from modification
 164                 $sLock = $src->getScopedFileLocks( $sPaths, LockManager::LOCK_UW, $status );
 165                 $eLock = $dst->getScopedFileLocks( $dPaths, LockManager::LOCK_EX, $status );
 166                 if ( !$status->isOK() ) {
 167                         return $status;
 168                 }
 169
 170                 $ops = array();
 171                 $fsFiles = array();
 172                 foreach ( $sPaths as $i => $sPath ) {
 173                         $dPath = $dPaths[$i]; // destination
 174                         $sExists = $src->fileExists( array( 'src' => $sPath, 'latest' => 1 ) );
 175                         if ( $sExists === true ) { // exists in source
 176                                 if ( $this->filesAreSame( $src, $dst, $sPath, $dPath ) ) {
 177                                         continue; // avoid local copies for non-FS backends
 178                                 }
 179                                 // Note: getLocalReference() is fast for FS backends
 180                                 $fsFile = $src->getLocalReference( array( 'src' => $sPath, 'latest' => 1 ) );
 181                                 if ( !$fsFile ) {
 182                                         $this->error( "Unable to sync '$dPath': could not get local copy." );
 183                                         $status->fatal( 'backend-fail-internal', $src->getName() );
 184                                         return $status;
 185                                 }
 186                                 $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 187                                 // Note: prepare() is usually fast for key/value backends
 188                                 $status->merge( $dst->prepare( array(
 189                                         'dir' => dirname( $dPath ), 'bypassReadOnly' => 1 ) ) );
 190                                 if ( !$status->isOK() ) {
 191                                         return $status;
 192                                 }
 193                                 $ops[] = array( 'op' => 'store',
 194                                         'src' => $fsFile->getPath(), 'dst' => $dPath, 'overwrite' => 1 );
 195                         } elseif ( $sExists === false ) { // does not exist in source
 196                                 $ops[] = array( 'op' => 'delete', 'src' => $dPath, 'ignoreMissingSource' => 1 );
 197                         } else { // error
 198                                 $this->error( "Unable to sync '$dPath': could not stat file." );
 199                                 $status->fatal( 'backend-fail-internal', $src->getName() );
 200                                 return $status;
 201                         }
 202                 }
 203
 204                 $t_start = microtime( true );
 205                 $status->merge( $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) ) );
 206                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 207                 if ( $status->isOK() && $this->getOption( 'verbose' ) ) {
 208                         $this->output( "Synchronized these file(s) [{$ellapsed_ms}ms]:\n" .
 209                                 implode( "\n", $dPaths ) . "\n" );
 210                 }
 211
 212                 return $status;
 213         }
 214
 215         /**
 216          * Substitute the backend name of storage paths with that of a given one
 217          *
 218          * @param $paths Array|string List of paths or single string path
 219          * @return Array|string
 220          */
 221         protected function replaceNamePaths( $paths, FileBackend $backend ) {
 222                 return preg_replace(
 223                         '!^mwstore://([^/]+)!',
 224                         StringUtils::escapeRegexReplacement( "mwstore://" . $backend->getName() ),
 225                         $paths // string or array
 226                 );
 227         }
 228
 229         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 230                 return (
 231                         ( $src->getFileSize( array( 'src' => $sPath ) )
 232                                 === $dst->getFileSize( array( 'src' => $dPath ) ) // short-circuit
 233                         ) && ( $src->getFileSha1Base36( array( 'src' => $sPath ) )
 234                                 === $dst->getFileSha1Base36( array( 'src' => $dPath ) )
 235                         )
 236                 );
 237         }
 238 }
 239
 240 $maintClass = "SyncFileBackend";
 241 require_once( RUN_MAINTENANCE_IF_MAIN );