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', false, 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( 'posdump', 'Just dump current journal position into the position dir.' );
  42                 $this->addOption( 'postime', 'For position dumps, get the ID at this time', false, true );
  43                 $this->addOption( 'verbose', 'Verbose mode', false, false, 'v' );
  44                 $this->setBatchSize( 50 );
  45         }
  46
  47         public function execute() {
  48                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  49
  50                 $posDir = $this->getOption( 'posdir' );
  51                 $posFile = $posDir ? $posDir . '/' . wfWikiID() : false;
  52
  53                 if ( $this->hasOption( 'posdump' ) ) {
  54                         // Just dump the current position into the specified position dir
  55                         if ( !$this->hasOption( 'posdir' ) ) {
  56                                 $this->error( "Param posdir required!", 1 );
  57                         }
  58                         if ( $this->hasOption( 'postime' ) ) {
  59                                 $id = (int)$src->getJournal()->getPositionAtTime( $this->getOption( 'postime' ) );
  60                         } else {
  61                                 $id = (int)$src->getJournal()->getCurrentPosition();
  62                         }
  63                         $this->output( "Current journal position is $id.\n" );
  64                         if ( file_put_contents( $posFile, $id, LOCK_EX ) !== false ) {
  65                                 $this->output( "Saved journal position file.\n" );
  66                         } else {
  67                                 $this->output( "Could not save journal position file.\n" );
  68                         }
  69                         if ( $this->isQuiet() ) {
  70                                 print $id; // give a single machine-readable number
  71                         }
  72                         return;
  73                 }
  74
  75                 if ( !$this->hasOption( 'dst' ) ) {
  76                         $this->error( "Param dst required!", 1 );
  77                 }
  78                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  79
  80                 $start = $this->getOption( 'start', 0 );
  81                 if ( !$start && $posFile && is_dir( $posDir ) ) {
  82                         $start = is_file( $posFile )
  83                                 ? (int)trim( file_get_contents( $posFile ) )
  84                                 : 0;
  85                         ++$start; // we already did this ID, start with the next one
  86                         $startFromPosFile = true;
  87                 } else {
  88                         $startFromPosFile = false;
  89                 }
  90                 $end = $this->getOption( 'end', INF );
  91
  92                 $this->output( "Synchronizing backend '{$dst->getName()}' to '{$src->getName()}'...\n" );
  93                 $this->output( "Starting journal position is $start.\n" );
  94                 if ( is_finite( $end ) ) {
  95                         $this->output( "Ending journal position is $end.\n" );
  96                 }
  97
  98                 // Actually sync the dest backend with the reference backend
  99                 $lastOKPos = $this->syncBackends( $src, $dst, $start, $end );
 100
 101                 // Update the sync position file
 102                 if ( $startFromPosFile && $lastOKPos >= $start ) { // successfully advanced
 103                         if ( file_put_contents( $posFile, $lastOKPos, LOCK_EX ) !== false ) {
 104                                 $this->output( "Updated journal position file.\n" );
 105                         } else {
 106                                 $this->output( "Could not update journal position file.\n" );
 107                         }
 108                 }
 109
 110                 if ( $lastOKPos === false ) {
 111                         if ( !$start ) {
 112                                 $this->output( "No journal entries found.\n" );
 113                         } else {
 114                                 $this->output( "No new journal entries found.\n" );
 115                         }
 116                 } else {
 117                         $this->output( "Stopped synchronization at journal position $lastOKPos.\n" );
 118                 }
 119
 120                 if ( $this->isQuiet() ) {
 121                         print $lastOKPos; // give a single machine-readable number
 122                 }
 123         }
 124
 125         /**
 126          * Sync $dst backend to $src backend based on the $src logs given after $start.
 127          * Returns the journal entry ID this advanced to and handled (inclusive).
 128          *
 129          * @param $src FileBackend
 130          * @param $dst FileBackend
 131          * @param $start integer Starting journal position
 132          * @param $end integer Starting journal position
 133          * @return integer|false Journal entry ID or false if there are none
 134          */
 135         protected function syncBackends( FileBackend $src, FileBackend $dst, $start, $end ) {
 136                 $lastOKPos = 0; // failed
 137                 $first = true; // first batch
 138
 139                 if ( $start > $end ) { // sanity
 140                         $this->error( "Error: given starting ID greater than ending ID.", 1 );
 141                 }
 142
 143                 do {
 144                         $limit = min( $this->mBatchSize, $end - $start + 1 ); // don't go pass ending ID
 145                         $this->output( "Doing id $start to " . ( $start + $limit - 1 ) . "...\n" );
 146
 147                         $entries = $src->getJournal()->getChangeEntries( $start, $limit, $next );
 148                         $start = $next; // start where we left off next time
 149                         if ( $first && !count( $entries ) ) {
 150                                 return false; // nothing to do
 151                         }
 152                         $first = false;
 153
 154                         $lastPosInBatch = 0;
 155                         $pathsInBatch = array(); // changed paths
 156                         foreach ( $entries as $entry ) {
 157                                 if ( $entry['op'] !== 'null' ) { // null ops are just for reference
 158                                         $pathsInBatch[$entry['path']] = 1; // remove duplicates
 159                                 }
 160                                 $lastPosInBatch = $entry['id'];
 161                         }
 162
 163                         $status = $this->syncFileBatch( array_keys( $pathsInBatch ), $src, $dst );
 164                         if ( $status->isOK() ) {
 165                                 $lastOKPos = max( $lastOKPos, $lastPosInBatch );
 166                         } else {
 167                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 168                                 break; // no gaps; everything up to $lastPos must be OK
 169                         }
 170
 171                         if ( !$start ) {
 172                                 $this->output( "End of journal entries.\n" );
 173                         }
 174                 } while ( $start && $start <= $end );
 175
 176                 return $lastOKPos;
 177         }
 178
 179         /**
 180          * Sync particular files of backend $src to the corresponding $dst backend files
 181          *
 182          * @param $paths Array
 183          * @param $src FileBackend
 184          * @param $dst FileBackend
 185          * @return Status
 186          */
 187         protected function syncFileBatch( array $paths, FileBackend $src, FileBackend $dst ) {
 188                 $status = Status::newGood();
 189                 if ( !count( $paths ) ) {
 190                         return $status; // nothing to do
 191                 }
 192
 193                 // Source: convert internal backend names (FileBackendMultiWrite) to the public one
 194                 $sPaths = $this->replaceNamePaths( $paths, $src );
 195                 // Destination: get corresponding path name
 196                 $dPaths = $this->replaceNamePaths( $paths, $dst );
 197
 198                 // Lock the live backend paths from modification
 199                 $sLock = $src->getScopedFileLocks( $sPaths, LockManager::LOCK_UW, $status );
 200                 $eLock = $dst->getScopedFileLocks( $dPaths, LockManager::LOCK_EX, $status );
 201                 if ( !$status->isOK() ) {
 202                         return $status;
 203                 }
 204
 205                 $ops = array();
 206                 $fsFiles = array();
 207                 foreach ( $sPaths as $i => $sPath ) {
 208                         $dPath = $dPaths[$i]; // destination
 209                         $sExists = $src->fileExists( array( 'src' => $sPath, 'latest' => 1 ) );
 210                         if ( $sExists === true ) { // exists in source
 211                                 if ( $this->filesAreSame( $src, $dst, $sPath, $dPath ) ) {
 212                                         continue; // avoid local copies for non-FS backends
 213                                 }
 214                                 // Note: getLocalReference() is fast for FS backends
 215                                 $fsFile = $src->getLocalReference( array( 'src' => $sPath, 'latest' => 1 ) );
 216                                 if ( !$fsFile ) {
 217                                         $this->error( "Unable to sync '$dPath': could not get local copy." );
 218                                         $status->fatal( 'backend-fail-internal', $src->getName() );
 219                                         return $status;
 220                                 }
 221                                 $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 222                                 // Note: prepare() is usually fast for key/value backends
 223                                 $status->merge( $dst->prepare( array(
 224                                         'dir' => dirname( $dPath ), 'bypassReadOnly' => 1 ) ) );
 225                                 if ( !$status->isOK() ) {
 226                                         return $status;
 227                                 }
 228                                 $ops[] = array( 'op' => 'store',
 229                                         'src' => $fsFile->getPath(), 'dst' => $dPath, 'overwrite' => 1 );
 230                         } elseif ( $sExists === false ) { // does not exist in source
 231                                 $ops[] = array( 'op' => 'delete', 'src' => $dPath, 'ignoreMissingSource' => 1 );
 232                         } else { // error
 233                                 $this->error( "Unable to sync '$dPath': could not stat file." );
 234                                 $status->fatal( 'backend-fail-internal', $src->getName() );
 235                                 return $status;
 236                         }
 237                 }
 238
 239                 $t_start = microtime( true );
 240                 $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 241                 if ( !$status->isOK() ) {
 242                         sleep( 10 ); // wait and retry copy again
 243                         $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 244                 }
 245                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 246                 if ( $status->isOK() && $this->getOption( 'verbose' ) ) {
 247                         $this->output( "Synchronized these file(s) [{$ellapsed_ms}ms]:\n" .
 248                                 implode( "\n", $dPaths ) . "\n" );
 249                 }
 250
 251                 return $status;
 252         }
 253
 254         /**
 255          * Substitute the backend name of storage paths with that of a given one
 256          *
 257          * @param $paths Array|string List of paths or single string path
 258          * @return Array|string
 259          */
 260         protected function replaceNamePaths( $paths, FileBackend $backend ) {
 261                 return preg_replace(
 262                         '!^mwstore://([^/]+)!',
 263                         StringUtils::escapeRegexReplacement( "mwstore://" . $backend->getName() ),
 264                         $paths // string or array
 265                 );
 266         }
 267
 268         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 269                 return (
 270                         ( $src->getFileSize( array( 'src' => $sPath ) )
 271                                 === $dst->getFileSize( array( 'src' => $dPath ) ) // short-circuit
 272                         ) && ( $src->getFileSha1Base36( array( 'src' => $sPath ) )
 273                                 === $dst->getFileSha1Base36( array( 'src' => $dPath ) )
 274                         )
 275                 );
 276         }
 277 }
 278
 279 $maintClass = "SyncFileBackend";
 280 require_once( RUN_MAINTENANCE_IF_MAIN );