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                                 $dstPathsRel = $dst->getFileList( array(
  83                                         'dir' => $dst->getRootStoragePath() . "/$backendRel" ) );
  84                                 if ( $dstPathsRel === null ) {
  85                                         $this->error( "Could not list files in $container.", 1 ); // die
  86                                 }
  87                                 // Get the list of destination files
  88                                 $relFilesDstSha1 = array();
  89                                 foreach ( $dstPathsRel as $dstPathRel ) {
  90                                         $relFilesDstSha1[sha1( $dstPathRel )] = 1;
  91                                 }
  92                                 unset( $dstPathsRel ); // free
  93                                 // Get the list of missing files
  94                                 $missingPathsRel = array();
  95                                 foreach ( $srcPathsRel as $srcPathRel ) {
  96                                         if ( !isset( $relFilesDstSha1[sha1( $srcPathRel )] ) ) {
  97                                                 $missingPathsRel[] = $srcPathRel;
  98                                         }
  99                                 }
 100                                 unset( $srcPathsRel ); // free
 101                                 // Only copy the missing files over in the next loop
 102                                 $srcPathsRel = $missingPathsRel;
 103                                 $this->output( count( $srcPathsRel ) . " file(s) need to be copied.\n" );
 104                         }
 105
 106                         $batchPaths = array();
 107                         foreach ( $srcPathsRel as $srcPathRel ) {
 108                                 // Check up on the rate file periodically to adjust the concurrency
 109                                 if ( $rateFile && ( !$count || ( $count % 500 ) == 0 ) ) {
 110                                         $this->mBatchSize = max( 1, (int)file_get_contents( $rateFile ) );
 111                                         $this->output( "Batch size is now {$this->mBatchSize}.\n" );
 112                                 }
 113                                 $batchPaths[$srcPathRel] = 1; // remove duplicates
 114                                 if ( count( $batchPaths ) >= $this->mBatchSize ) {
 115                                         $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 116                                         $batchPaths = array(); // done
 117                                 }
 118                                 ++$count;
 119                         }
 120                         if ( count( $batchPaths ) ) { // left-overs
 121                                 $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 122                                 $batchPaths = array(); // done
 123                         }
 124
 125                         if ( $subDir != '' ) {
 126                                 $this->output( "Finished container '$container', directory '$subDir'.\n" );
 127                         } else {
 128                                 $this->output( "Finished container '$container'.\n" );
 129                         }
 130                 }
 131
 132                 $this->output( "Done [$count file(s)].\n" );
 133         }
 134
 135         protected function copyFileBatch(
 136                 array $srcPathsRel, $backendRel, FileBackend $src, FileBackend $dst
 137         ) {
 138                 $ops = array();
 139                 $fsFiles = array();
 140                 $copiedRel = array(); // for output message
 141
 142                 // Download the batch of source files into backend cache...
 143                 if ( $this->hasOption( 'missingonly' ) ) {
 144                         $srcPaths = array();
 145                         foreach ( $srcPathsRel as $srcPathRel ) {
 146                                 $srcPaths[] = $src->getRootStoragePath() . "/$backendRel/$srcPathRel";
 147                         }
 148                         $t_start = microtime( true );
 149                         $fsFiles = $src->getLocalReferenceMulti( array( 'srcs' => $srcPaths, 'latest' => 1 ) );
 150                         $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 151                         $this->output( "\nDownloaded these file(s) [{$ellapsed_ms}ms]:\n" .
 152                                 implode( "\n", $srcPaths ) . "\n\n" );
 153                 }
 154
 155                 // Determine what files need to be copied over...
 156                 foreach ( $srcPathsRel as $srcPathRel ) {
 157                         $srcPath = $src->getRootStoragePath() . "/$backendRel/$srcPathRel";
 158                         $dstPath = $dst->getRootStoragePath() . "/$backendRel/$srcPathRel";
 159                         if ( $this->hasOption( 'utf8only' ) && !mb_check_encoding( $srcPath, 'UTF-8' ) ) {
 160                                 $this->error( "Detected illegal (non-UTF8) path for $srcPath." );
 161                                 continue;
 162                         } elseif ( $this->filesAreSame( $src, $dst, $srcPath, $dstPath ) ) {
 163                                 $this->output( "Already have $srcPathRel.\n" );
 164                                 continue; // assume already copied...
 165                         }
 166                         $fsFile = array_key_exists( $srcPath, $fsFiles )
 167                                 ? $fsFiles[$srcPath]
 168                                 : $src->getLocalReference( array( 'src' => $srcPath, 'latest' => 1 ) );
 169                         if ( !$fsFile ) {
 170                                 $this->error( "Could not get local copy of $srcPath.", 1 ); // die
 171                         } elseif ( !$fsFile->exists() ) {
 172                                 // FSFileBackends just return the path for getLocalReference() and paths with
 173                                 // illegal slashes may get normalized to a different path. This can cause the
 174                                 // local reference to not exist...skip these broken files.
 175                                 $this->error( "Detected possible illegal path for $srcPath." );
 176                                 continue;
 177                         }
 178                         $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 179                         // Note: prepare() is usually fast for key/value backends
 180                         $status = $dst->prepare( array( 'dir' => dirname( $dstPath ), 'bypassReadOnly' => 1 ) );
 181                         if ( !$status->isOK() ) {
 182                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 183                                 $this->error( "Could not copy $srcPath to $dstPath.", 1 ); // die
 184                         }
 185                         $ops[] = array( 'op' => 'store',
 186                                 'src' => $fsFile->getPath(), 'dst' => $dstPath, 'overwrite' => 1 );
 187                         $copiedRel[] = $srcPathRel;
 188                 }
 189
 190                 // Copy in the batch of source files...
 191                 $t_start = microtime( true );
 192                 $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 193                 if ( !$status->isOK() ) {
 194                         sleep( 10 ); // wait and retry copy again
 195                         $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 196                 }
 197                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 198                 if ( !$status->isOK() ) {
 199                         $this->error( print_r( $status->getErrorsArray(), true ) );
 200                         $this->error( "Could not copy file batch.", 1 ); // die
 201                 } elseif ( count( $copiedRel ) ) {
 202                         $this->output( "\nCopied these file(s) [{$ellapsed_ms}ms]:\n" .
 203                                 implode( "\n", $copiedRel ) . "\n\n" );
 204                 }
 205         }
 206
 207         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 208                 $skipHash = $this->hasOption( 'skiphash' );
 209                 return (
 210                         ( $src->fileExists( array( 'src' => $sPath, 'latest' => 1 ) )
 211                                 === $dst->fileExists( array( 'src' => $dPath, 'latest' => 1 ) ) // short-circuit
 212                         ) && ( $src->getFileSize( array( 'src' => $sPath, 'latest' => 1 ) )
 213                                 === $dst->getFileSize( array( 'src' => $dPath, 'latest' => 1 ) ) // short-circuit
 214                         ) && ( $skipHash || ( $src->getFileSha1Base36( array( 'src' => $sPath, 'latest' => 1 ) )
 215                                 === $dst->getFileSha1Base36( array( 'src' => $dPath, 'latest' => 1 ) )
 216                         ) )
 217                 );
 218         }
 219 }
 220
 221 $maintClass = 'CopyFileBackend';
 222 require_once( RUN_MAINTENANCE_IF_MAIN );