maintenance/copyFileBackend.php

   1 <?php
   2 /**
   3  * Copy all files in some containers of one backend to another.
   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  * Copy all files in one container of one backend to another.
  28  *
  29  * This can also be used to re-shard the files for one backend using the
  30  * config of second backend. The second backend should have the same config
  31  * as the first, except for it having a different name and different sharding
  32  * configuration. The backend should be made read-only while this runs.
  33  * After this script finishes, the old files in the containers can be deleted.
  34  *
  35  * @ingroup Maintenance
  36  */
  37 class CopyFileBackend extends Maintenance {
  38         public function __construct() {
  39                 parent::__construct();
  40                 $this->mDescription = "Copy files in one backend to another.";
  41                 $this->addOption( 'src', 'Backend containing the source files', true, true );
  42                 $this->addOption( 'dst', 'Backend where files should be copied to', true, true );
  43                 $this->addOption( 'containers', 'Pipe separated list of containers', true, true );
  44                 $this->addOption( 'subdir', 'Only do items in this child directory', false, true );
  45                 $this->addOption( 'ratefile', 'File to check periodically for batch size', false, true );
  46                 $this->addOption( 'skiphash', 'Skip SHA-1 sync checks for files' );
  47                 $this->addOption( 'missingonly', 'Only copy files missing from destination listing' );
  48                 $this->addOption( 'utf8only', 'Skip source files that do not have valid UTF-8 names' );
  49                 $this->setBatchSize( 50 );
  50         }
  51
  52         public function execute() {
  53                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  54                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  55                 $containers = explode( '|', $this->getOption( 'containers' ) );
  56                 $subDir = $this->getOption( rtrim( 'subdir', '/' ), '' );
  57
  58                 $rateFile = $this->getOption( 'ratefile' );
  59
  60                 if ( $this->hasOption( 'utf8only' ) && !extension_loaded( 'mbstring' ) ) {
  61                         $this->error( "Cannot check for UTF-8, mbstring extension missing.", 1 ); // die
  62                 }
  63
  64                 $count = 0;
  65                 foreach ( $containers as $container ) {
  66                         if ( $subDir != '' ) {
  67                                 $backendRel = "$container/$subDir";
  68                                 $this->output( "Doing container '$container', directory '$subDir'...\n" );
  69                         } else {
  70                                 $backendRel = $container;
  71                                 $this->output( "Doing container '$container'...\n" );
  72                         }
  73
  74                         $srcPathsRel = $src->getFileList( array(
  75                                 'dir' => $src->getRootStoragePath() . "/$backendRel" ) );
  76                         if ( $srcPathsRel === null ) {
  77                                 $this->error( "Could not list files in $container.", 1 ); // die
  78                         }
  79
  80                         // Do a listing comparison if specified
  81                         if ( $this->hasOption( 'missingonly' ) ) {
  82                                 $relFilesSrc = array();
  83                                 $relFilesDst = array();
  84                                 foreach ( $srcPathsRel as $srcPathRel ) {
  85                                         $relFilesSrc[] = $srcPathRel;
  86                                 }
  87                                 $dstPathsRel = $dst->getFileList( array(
  88                                         'dir' => $dst->getRootStoragePath() . "/$backendRel" ) );
  89                                 if ( $dstPathsRel === null ) {
  90                                         $this->error( "Could not list files in $container.", 1 ); // die
  91                                 }
  92                                 foreach ( $dstPathsRel as $dstPathRel ) {
  93                                         $relFilesDst[] = $dstPathRel;
  94                                 }
  95                                 // Only copy the missing files over in the next loop
  96                                 $srcPathsRel = array_diff( $relFilesSrc, $relFilesDst );
  97                                 $this->output( count( $srcPathsRel ) . " file(s) need to be copied.\n" );
  98                                 unset( $relFilesSrc );
  99                                 unset( $relFilesDst );
 100                         }
 101
 102                         $batchPaths = array();
 103                         foreach ( $srcPathsRel as $srcPathRel ) {
 104                                 // Check up on the rate file periodically to adjust the concurrency
 105                                 if ( $rateFile && ( !$count || ( $count % 500 ) == 0 ) ) {
 106                                         $this->mBatchSize = max( 1, (int)file_get_contents( $rateFile ) );
 107                                         $this->output( "Batch size is now {$this->mBatchSize}.\n" );
 108                                 }
 109                                 $batchPaths[$srcPathRel] = 1; // remove duplicates
 110                                 if ( count( $batchPaths ) >= $this->mBatchSize ) {
 111                                         $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 112                                         $batchPaths = array(); // done
 113                                 }
 114                                 ++$count;
 115                         }
 116                         if ( count( $batchPaths ) ) { // left-overs
 117                                 $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 118                                 $batchPaths = array(); // done
 119                         }
 120
 121                         if ( $subDir != '' ) {
 122                                 $this->output( "Finished container '$container', directory '$subDir'.\n" );
 123                         } else {
 124                                 $this->output( "Finished container '$container'.\n" );
 125                         }
 126                 }
 127
 128                 $this->output( "Done [$count file(s)].\n" );
 129         }
 130
 131         protected function copyFileBatch(
 132                 array $srcPathsRel, $backendRel, FileBackend $src, FileBackend $dst
 133         ) {
 134                 $ops = array();
 135                 $fsFiles = array();
 136                 $copiedRel = array(); // for output message
 137                 foreach ( $srcPathsRel as $srcPathRel ) {
 138                         $srcPath = $src->getRootStoragePath() . "/$backendRel/$srcPathRel";
 139                         $dstPath = $dst->getRootStoragePath() . "/$backendRel/$srcPathRel";
 140                         if ( $this->hasOption( 'utf8only' ) && !mb_check_encoding( $srcPath, 'UTF-8' ) ) {
 141                                 $this->error( "Detected illegal (non-UTF8) path for $srcPath." );
 142                                 continue;
 143                         } elseif ( $this->filesAreSame( $src, $dst, $srcPath, $dstPath ) ) {
 144                                 $this->output( "Already have $srcPathRel.\n" );
 145                                 continue; // assume already copied...
 146                         }
 147                         // Note: getLocalReference() is fast for FS backends
 148                         $fsFile = $src->getLocalReference( array( 'src' => $srcPath, 'latest' => 1 ) );
 149                         if ( !$fsFile ) {
 150                                 $this->error( "Could not get local copy of $srcPath.", 1 ); // die
 151                         } elseif ( !$fsFile->exists() ) {
 152                                 // FSFileBackends just return the path for getLocalReference() and paths with
 153                                 // illegal slashes may get normalized to a different path. This can cause the
 154                                 // local reference to not exist...skip these broken files.
 155                                 $this->error( "Detected possible illegal path for $srcPath." );
 156                                 continue;
 157                         }
 158                         $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 159                         // Note: prepare() is usually fast for key/value backends
 160                         $status = $dst->prepare( array( 'dir' => dirname( $dstPath ), 'bypassReadOnly' => 1 ) );
 161                         if ( !$status->isOK() ) {
 162                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 163                                 $this->error( "Could not copy $srcPath to $dstPath.", 1 ); // die
 164                         }
 165                         $ops[] = array( 'op' => 'store',
 166                                 'src' => $fsFile->getPath(), 'dst' => $dstPath, 'overwrite' => 1 );
 167                         $copiedRel[] = $srcPathRel;
 168                 }
 169
 170                 $t_start = microtime( true );
 171                 $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 172                 if ( !$status->isOK() ) {
 173                         sleep( 10 ); // wait and retry copy again
 174                         $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 175                 }
 176                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 177                 if ( !$status->isOK() ) {
 178                         $this->error( print_r( $status->getErrorsArray(), true ) );
 179                         $this->error( "Could not copy file batch.", 1 ); // die
 180                 } elseif ( count( $copiedRel ) ) {
 181                         $this->output( "\nCopied these file(s) [{$ellapsed_ms}ms]:\n" .
 182                                 implode( "\n", $copiedRel ) . "\n\n" );
 183                 }
 184         }
 185
 186         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 187                 $skipHash = $this->hasOption( 'skiphash' );
 188                 return (
 189                         ( $src->fileExists( array( 'src' => $sPath, 'latest' => 1 ) )
 190                                 === $dst->fileExists( array( 'src' => $dPath, 'latest' => 1 ) ) // short-circuit
 191                         ) && ( $src->getFileSize( array( 'src' => $sPath, 'latest' => 1 ) )
 192                                 === $dst->getFileSize( array( 'src' => $dPath, 'latest' => 1 ) ) // short-circuit
 193                         ) && ( $skipHash || ( $src->getFileSha1Base36( array( 'src' => $sPath, 'latest' => 1 ) )
 194                                 === $dst->getFileSha1Base36( array( 'src' => $dPath, 'latest' => 1 ) )
 195                         ) )
 196                 );
 197         }
 198 }
 199
 200 $maintClass = 'CopyFileBackend';
 201 require_once( RUN_MAINTENANCE_IF_MAIN );