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->addDescription( '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( 'backoff', 'Stop at entries younger than this age (sec).', false, true );
  44                 $this->addOption( 'verbose', 'Verbose mode', false, false, 'v' );
  45                 $this->setBatchSize( 50 );
  46         }
  47
  48         public function execute() {
  49                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  50
  51                 $posDir = $this->getOption( 'posdir' );
  52                 $posFile = $posDir ? $posDir . '/' . wfWikiID() : false;
  53
  54                 if ( $this->hasOption( 'posdump' ) ) {
  55                         // Just dump the current position into the specified position dir
  56                         if ( !$this->hasOption( 'posdir' ) ) {
  57                                 $this->fatalError( "Param posdir required!" );
  58                         }
  59                         if ( $this->hasOption( 'postime' ) ) {
  60                                 $id = (int)$src->getJournal()->getPositionAtTime( $this->getOption( 'postime' ) );
  61                                 $this->output( "Requested journal position is $id.\n" );
  62                         } else {
  63                                 $id = (int)$src->getJournal()->getCurrentPosition();
  64                                 $this->output( "Current journal position is $id.\n" );
  65                         }
  66                         if ( file_put_contents( $posFile, $id, LOCK_EX ) !== false ) {
  67                                 $this->output( "Saved journal position file.\n" );
  68                         } else {
  69                                 $this->output( "Could not save journal position file.\n" );
  70                         }
  71                         if ( $this->isQuiet() ) {
  72                                 print $id; // give a single machine-readable number
  73                         }
  74
  75                         return;
  76                 }
  77
  78                 if ( !$this->hasOption( 'dst' ) ) {
  79                         $this->fatalError( "Param dst required!" );
  80                 }
  81                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  82
  83                 $start = $this->getOption( 'start', 0 );
  84                 if ( !$start && $posFile && is_dir( $posDir ) ) {
  85                         $start = is_file( $posFile )
  86                                 ? (int)trim( file_get_contents( $posFile ) )
  87                                 : 0;
  88                         ++$start; // we already did this ID, start with the next one
  89                         $startFromPosFile = true;
  90                 } else {
  91                         $startFromPosFile = false;
  92                 }
  93
  94                 if ( $this->hasOption( 'backoff' ) ) {
  95                         $time = time() - $this->getOption( 'backoff', 0 );
  96                         $end = (int)$src->getJournal()->getPositionAtTime( $time );
  97                 } else {
  98                         $end = $this->getOption( 'end', INF );
  99                 }
 100
 101                 $this->output( "Synchronizing backend '{$dst->getName()}' to '{$src->getName()}'...\n" );
 102                 $this->output( "Starting journal position is $start.\n" );
 103                 if ( is_finite( $end ) ) {
 104                         $this->output( "Ending journal position is $end.\n" );
 105                 }
 106
 107                 // Periodically update the position file
 108                 $callback = function ( $pos ) use ( $startFromPosFile, $posFile, $start ) {
 109                         if ( $startFromPosFile && $pos >= $start ) { // successfully advanced
 110                                 file_put_contents( $posFile, $pos, LOCK_EX );
 111                         }
 112                 };
 113
 114                 // Actually sync the dest backend with the reference backend
 115                 $lastOKPos = $this->syncBackends( $src, $dst, $start, $end, $callback );
 116
 117                 // Update the sync position file
 118                 if ( $startFromPosFile && $lastOKPos >= $start ) { // successfully advanced
 119                         if ( file_put_contents( $posFile, $lastOKPos, LOCK_EX ) !== false ) {
 120                                 $this->output( "Updated journal position file.\n" );
 121                         } else {
 122                                 $this->output( "Could not update journal position file.\n" );
 123                         }
 124                 }
 125
 126                 if ( $lastOKPos === false ) {
 127                         if ( !$start ) {
 128                                 $this->output( "No journal entries found.\n" );
 129                         } else {
 130                                 $this->output( "No new journal entries found.\n" );
 131                         }
 132                 } else {
 133                         $this->output( "Stopped synchronization at journal position $lastOKPos.\n" );
 134                 }
 135
 136                 if ( $this->isQuiet() ) {
 137                         print $lastOKPos; // give a single machine-readable number
 138                 }
 139         }
 140
 141         /**
 142          * Sync $dst backend to $src backend based on the $src logs given after $start.
 143          * Returns the journal entry ID this advanced to and handled (inclusive).
 144          *
 145          * @param FileBackend $src
 146          * @param FileBackend $dst
 147          * @param int $start Starting journal position
 148          * @param int $end Starting journal position
 149          * @param Closure $callback Callback to update any position file
 150          * @return int|bool Journal entry ID or false if there are none
 151          */
 152         protected function syncBackends(
 153                 FileBackend $src, FileBackend $dst, $start, $end, Closure $callback
 154         ) {
 155                 $lastOKPos = 0; // failed
 156                 $first = true; // first batch
 157
 158                 if ( $start > $end ) { // sanity
 159                         $this->fatalError( "Error: given starting ID greater than ending ID." );
 160                 }
 161
 162                 $next = null;
 163                 do {
 164                         $limit = min( $this->getBatchSize(), $end - $start + 1 ); // don't go pass ending ID
 165                         $this->output( "Doing id $start to " . ( $start + $limit - 1 ) . "...\n" );
 166
 167                         $entries = $src->getJournal()->getChangeEntries( $start, $limit, $next );
 168                         $start = $next; // start where we left off next time
 169                         if ( $first && !count( $entries ) ) {
 170                                 return false; // nothing to do
 171                         }
 172                         $first = false;
 173
 174                         $lastPosInBatch = 0;
 175                         $pathsInBatch = []; // changed paths
 176                         foreach ( $entries as $entry ) {
 177                                 if ( $entry['op'] !== 'null' ) { // null ops are just for reference
 178                                         $pathsInBatch[$entry['path']] = 1; // remove duplicates
 179                                 }
 180                                 $lastPosInBatch = $entry['id'];
 181                         }
 182
 183                         $status = $this->syncFileBatch( array_keys( $pathsInBatch ), $src, $dst );
 184                         if ( $status->isOK() ) {
 185                                 $lastOKPos = max( $lastOKPos, $lastPosInBatch );
 186                                 $callback( $lastOKPos ); // update position file
 187                         } else {
 188                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 189                                 break; // no gaps; everything up to $lastPos must be OK
 190                         }
 191
 192                         if ( !$start ) {
 193                                 $this->output( "End of journal entries.\n" );
 194                         }
 195                 } while ( $start && $start <= $end );
 196
 197                 return $lastOKPos;
 198         }
 199
 200         /**
 201          * Sync particular files of backend $src to the corresponding $dst backend files
 202          *
 203          * @param array $paths
 204          * @param FileBackend $src
 205          * @param FileBackend $dst
 206          * @return Status
 207          */
 208         protected function syncFileBatch( array $paths, FileBackend $src, FileBackend $dst ) {
 209                 $status = Status::newGood();
 210                 if ( !count( $paths ) ) {
 211                         return $status; // nothing to do
 212                 }
 213
 214                 // Source: convert internal backend names (FileBackendMultiWrite) to the public one
 215                 $sPaths = $this->replaceNamePaths( $paths, $src );
 216                 // Destination: get corresponding path name
 217                 $dPaths = $this->replaceNamePaths( $paths, $dst );
 218
 219                 // Lock the live backend paths from modification
 220                 $sLock = $src->getScopedFileLocks( $sPaths, LockManager::LOCK_UW, $status );
 221                 $eLock = $dst->getScopedFileLocks( $dPaths, LockManager::LOCK_EX, $status );
 222                 if ( !$status->isOK() ) {
 223                         return $status;
 224                 }
 225
 226                 $src->preloadFileStat( [ 'srcs' => $sPaths, 'latest' => 1 ] );
 227                 $dst->preloadFileStat( [ 'srcs' => $dPaths, 'latest' => 1 ] );
 228
 229                 $ops = [];
 230                 $fsFiles = [];
 231                 foreach ( $sPaths as $i => $sPath ) {
 232                         $dPath = $dPaths[$i]; // destination
 233                         $sExists = $src->fileExists( [ 'src' => $sPath, 'latest' => 1 ] );
 234                         if ( $sExists === true ) { // exists in source
 235                                 if ( $this->filesAreSame( $src, $dst, $sPath, $dPath ) ) {
 236                                         continue; // avoid local copies for non-FS backends
 237                                 }
 238                                 // Note: getLocalReference() is fast for FS backends
 239                                 $fsFile = $src->getLocalReference( [ 'src' => $sPath, 'latest' => 1 ] );
 240                                 if ( !$fsFile ) {
 241                                         $this->error( "Unable to sync '$dPath': could not get local copy." );
 242                                         $status->fatal( 'backend-fail-internal', $src->getName() );
 243
 244                                         return $status;
 245                                 }
 246                                 $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 247                                 // Note: prepare() is usually fast for key/value backends
 248                                 $status->merge( $dst->prepare( [
 249                                         'dir' => dirname( $dPath ), 'bypassReadOnly' => 1 ] ) );
 250                                 if ( !$status->isOK() ) {
 251                                         return $status;
 252                                 }
 253                                 $ops[] = [ 'op' => 'store',
 254                                         'src' => $fsFile->getPath(), 'dst' => $dPath, 'overwrite' => 1 ];
 255                         } elseif ( $sExists === false ) { // does not exist in source
 256                                 $ops[] = [ 'op' => 'delete', 'src' => $dPath, 'ignoreMissingSource' => 1 ];
 257                         } else { // error
 258                                 $this->error( "Unable to sync '$dPath': could not stat file." );
 259                                 $status->fatal( 'backend-fail-internal', $src->getName() );
 260
 261                                 return $status;
 262                         }
 263                 }
 264
 265                 $t_start = microtime( true );
 266                 $status = $dst->doQuickOperations( $ops, [ 'bypassReadOnly' => 1 ] );
 267                 if ( !$status->isOK() ) {
 268                         sleep( 10 ); // wait and retry copy again
 269                         $status = $dst->doQuickOperations( $ops, [ 'bypassReadOnly' => 1 ] );
 270                 }
 271                 $elapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 272                 if ( $status->isOK() && $this->getOption( 'verbose' ) ) {
 273                         $this->output( "Synchronized these file(s) [{$elapsed_ms}ms]:\n" .
 274                                 implode( "\n", $dPaths ) . "\n" );
 275                 }
 276
 277                 return $status;
 278         }
 279
 280         /**
 281          * Substitute the backend name of storage paths with that of a given one
 282          *
 283          * @param array|string $paths List of paths or single string path
 284          * @param FileBackend $backend
 285          * @return array|string
 286          */
 287         protected function replaceNamePaths( $paths, FileBackend $backend ) {
 288                 return preg_replace(
 289                         '!^mwstore://([^/]+)!',
 290                         StringUtils::escapeRegexReplacement( "mwstore://" . $backend->getName() ),
 291                         $paths // string or array
 292                 );
 293         }
 294
 295         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 296                 return (
 297                         ( $src->getFileSize( [ 'src' => $sPath ] )
 298                                 === $dst->getFileSize( [ 'src' => $dPath ] ) // short-circuit
 299                         ) && ( $src->getFileSha1Base36( [ 'src' => $sPath ] )
 300                                 === $dst->getFileSha1Base36( [ 'src' => $dPath ] )
 301                         )
 302                 );
 303         }
 304 }
 305
 306 $maintClass = SyncFileBackend::class;
 307 require_once RUN_MAINTENANCE_IF_MAIN;