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