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         protected $statCache = array();
  39
  40         public function __construct() {
  41                 parent::__construct();
  42                 $this->mDescription = "Copy files in one backend to another.";
  43                 $this->addOption( 'src', 'Backend containing the source files', true, true );
  44                 $this->addOption( 'dst', 'Backend where files should be copied to', true, true );
  45                 $this->addOption( 'containers', 'Pipe separated list of containers', true, true );
  46                 $this->addOption( 'subdir', 'Only do items in this child directory', false, true );
  47                 $this->addOption( 'ratefile', 'File to check periodically for batch size', false, true );
  48                 $this->addOption( 'prestat', 'Stat the destination files first (try to use listings)' );
  49                 $this->addOption( 'skiphash', 'Skip SHA-1 sync checks for files' );
  50                 $this->addOption( 'missingonly', 'Only copy files missing from destination listing' );
  51                 $this->addOption( 'utf8only', 'Skip source files that do not have valid UTF-8 names' );
  52                 $this->setBatchSize( 50 );
  53         }
  54
  55         public function execute() {
  56                 $src = FileBackendGroup::singleton()->get( $this->getOption( 'src' ) );
  57                 $dst = FileBackendGroup::singleton()->get( $this->getOption( 'dst' ) );
  58                 $containers = explode( '|', $this->getOption( 'containers' ) );
  59                 $subDir = $this->getOption( rtrim( 'subdir', '/' ), '' );
  60
  61                 $rateFile = $this->getOption( 'ratefile' );
  62
  63                 if ( $this->hasOption( 'utf8only' ) && !extension_loaded( 'mbstring' ) ) {
  64                         $this->error( "Cannot check for UTF-8, mbstring extension missing.", 1 ); // die
  65                 }
  66
  67                 $count = 0;
  68                 foreach ( $containers as $container ) {
  69                         if ( $subDir != '' ) {
  70                                 $backendRel = "$container/$subDir";
  71                                 $this->output( "Doing container '$container', directory '$subDir'...\n" );
  72                         } else {
  73                                 $backendRel = $container;
  74                                 $this->output( "Doing container '$container'...\n" );
  75                         }
  76
  77                         $srcPathsRel = $src->getFileList( array(
  78                                 'dir' => $src->getRootStoragePath() . "/$backendRel",
  79                                 'adviseStat' => !$this->hasOption( 'missingonly' ) // avoid HEADs
  80                         ) );
  81                         if ( $srcPathsRel === null ) {
  82                                 $this->error( "Could not list files in $container.", 1 ); // die
  83                         }
  84
  85                         if ( $this->hasOption( 'missingonly' ) ) {
  86                                 $dstPathsRel = $dst->getFileList( array(
  87                                         'dir' => $dst->getRootStoragePath() . "/$backendRel" ) );
  88                                 if ( $dstPathsRel === null ) {
  89                                         $this->error( "Could not list files in $container.", 1 ); // die
  90                                 }
  91                                 // Get the list of destination files
  92                                 $relFilesDstSha1 = array();
  93                                 foreach ( $dstPathsRel as $dstPathRel ) {
  94                                         $relFilesDstSha1[sha1( $dstPathRel )] = 1;
  95                                 }
  96                                 unset( $dstPathsRel ); // free
  97                                 // Get the list of missing files
  98                                 $missingPathsRel = array();
  99                                 foreach ( $srcPathsRel as $srcPathRel ) {
 100                                         if ( !isset( $relFilesDstSha1[sha1( $srcPathRel )] ) ) {
 101                                                 $missingPathsRel[] = $srcPathRel;
 102                                         }
 103                                 }
 104                                 unset( $srcPathsRel ); // free
 105                                 // Only copy the missing files over in the next loop
 106                                 $srcPathsRel = $missingPathsRel;
 107                                 $this->output( count( $srcPathsRel ) . " file(s) need to be copied.\n" );
 108                         } elseif ( $this->getOption( 'prestat' ) ) {
 109                                 // Build the stat cache for the destination files
 110                                 $this->output( "Building destination stat cache..." );
 111                                 $dstPathsRel = $dst->getFileList( array(
 112                                         'dir' => $dst->getRootStoragePath() . "/$backendRel",
 113                                         'adviseStat' => true // avoid HEADs
 114                                 ) );
 115                                 if ( $dstPathsRel === null ) {
 116                                         $this->error( "Could not list files in $container.", 1 ); // die
 117                                 }
 118                                 $this->statCache = array(); // clear
 119                                 foreach ( $dstPathsRel as $dstPathRel ) {
 120                                         $path = $dst->getRootStoragePath() . "/$backendRel/$dstPathRel";
 121                                         $this->statCache[sha1( $path )] = $dst->getFileStat( array( 'src' => $path ) );
 122                                 }
 123                                 $this->output( "done [" . count( $this->statCache ) . " file(s)]\n" );
 124                         }
 125
 126                         $batchPaths = array();
 127                         foreach ( $srcPathsRel as $srcPathRel ) {
 128                                 // Check up on the rate file periodically to adjust the concurrency
 129                                 if ( $rateFile && ( !$count || ( $count % 500 ) == 0 ) ) {
 130                                         $this->mBatchSize = max( 1, (int)file_get_contents( $rateFile ) );
 131                                         $this->output( "Batch size is now {$this->mBatchSize}.\n" );
 132                                 }
 133                                 $batchPaths[$srcPathRel] = 1; // remove duplicates
 134                                 if ( count( $batchPaths ) >= $this->mBatchSize ) {
 135                                         $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 136                                         $batchPaths = array(); // done
 137                                 }
 138                                 ++$count;
 139                         }
 140                         if ( count( $batchPaths ) ) { // left-overs
 141                                 $this->copyFileBatch( array_keys( $batchPaths ), $backendRel, $src, $dst );
 142                                 $batchPaths = array(); // done
 143                         }
 144
 145                         if ( $subDir != '' ) {
 146                                 $this->output( "Finished container '$container', directory '$subDir'.\n" );
 147                         } else {
 148                                 $this->output( "Finished container '$container'.\n" );
 149                         }
 150                 }
 151
 152                 $this->output( "Done [$count file(s)].\n" );
 153         }
 154
 155         protected function copyFileBatch(
 156                 array $srcPathsRel, $backendRel, FileBackend $src, FileBackend $dst
 157         ) {
 158                 $ops = array();
 159                 $fsFiles = array();
 160                 $copiedRel = array(); // for output message
 161                 $wikiId = $src->getWikiId();
 162
 163                 // Download the batch of source files into backend cache...
 164                 if ( $this->hasOption( 'missingonly' ) ) {
 165                         $srcPaths = array();
 166                         foreach ( $srcPathsRel as $srcPathRel ) {
 167                                 $srcPaths[] = $src->getRootStoragePath() . "/$backendRel/$srcPathRel";
 168                         }
 169                         $t_start = microtime( true );
 170                         $fsFiles = $src->getLocalReferenceMulti( array( 'srcs' => $srcPaths, 'latest' => 1 ) );
 171                         $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 172                         $this->output( "\nDownloaded these file(s) [{$ellapsed_ms}ms]:\n" .
 173                                 implode( "\n", $srcPaths ) . "\n\n" );
 174                 }
 175
 176                 // Determine what files need to be copied over...
 177                 foreach ( $srcPathsRel as $srcPathRel ) {
 178                         $srcPath = $src->getRootStoragePath() . "/$backendRel/$srcPathRel";
 179                         $dstPath = $dst->getRootStoragePath() . "/$backendRel/$srcPathRel";
 180                         if ( $this->hasOption( 'utf8only' ) && !mb_check_encoding( $srcPath, 'UTF-8' ) ) {
 181                                 $this->error( "$wikiId: Detected illegal (non-UTF8) path for $srcPath." );
 182                                 continue;
 183                         } elseif ( !$this->hasOption( 'missingonly' )
 184                                 && $this->filesAreSame( $src, $dst, $srcPath, $dstPath ) )
 185                         {
 186                                 $this->output( "Already have $srcPathRel.\n" );
 187                                 continue; // assume already copied...
 188                         }
 189                         $fsFile = array_key_exists( $srcPath, $fsFiles )
 190                                 ? $fsFiles[$srcPath]
 191                                 : $src->getLocalReference( array( 'src' => $srcPath, 'latest' => 1 ) );
 192                         if ( !$fsFile ) {
 193                                 $src->clearCache( array( $srcPath ) );
 194                                 if ( $src->fileExists( array( 'src' => $srcPath, 'latest' => 1 ) ) === false ) {
 195                                         $this->error( "$wikiId: File '$srcPath' was listed but does not exist." );
 196                                 } else {
 197                                         $this->error( "$wikiId: Could not get local copy of $srcPath." );
 198                                 }
 199                                 continue;
 200                         } elseif ( !$fsFile->exists() ) {
 201                                 // FSFileBackends just return the path for getLocalReference() and paths with
 202                                 // illegal slashes may get normalized to a different path. This can cause the
 203                                 // local reference to not exist...skip these broken files.
 204                                 $this->error( "$wikiId: Detected possible illegal path for $srcPath." );
 205                                 continue;
 206                         }
 207                         $fsFiles[] = $fsFile; // keep TempFSFile objects alive as needed
 208                         // Note: prepare() is usually fast for key/value backends
 209                         $status = $dst->prepare( array( 'dir' => dirname( $dstPath ), 'bypassReadOnly' => 1 ) );
 210                         if ( !$status->isOK() ) {
 211                                 $this->error( print_r( $status->getErrorsArray(), true ) );
 212                                 $this->error( "$wikiId: Could not copy $srcPath to $dstPath.", 1 ); // die
 213                         }
 214                         $ops[] = array( 'op' => 'store',
 215                                 'src' => $fsFile->getPath(), 'dst' => $dstPath, 'overwrite' => 1 );
 216                         $copiedRel[] = $srcPathRel;
 217                 }
 218
 219                 // Copy in the batch of source files...
 220                 $t_start = microtime( true );
 221                 $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 222                 if ( !$status->isOK() ) {
 223                         sleep( 10 ); // wait and retry copy again
 224                         $status = $dst->doQuickOperations( $ops, array( 'bypassReadOnly' => 1 ) );
 225                 }
 226                 $ellapsed_ms = floor( ( microtime( true ) - $t_start ) * 1000 );
 227                 if ( !$status->isOK() ) {
 228                         $this->error( print_r( $status->getErrorsArray(), true ) );
 229                         $this->error( "$wikiId: Could not copy file batch.", 1 ); // die
 230                 } elseif ( count( $copiedRel ) ) {
 231                         $this->output( "\nCopied these file(s) [{$ellapsed_ms}ms]:\n" .
 232                                 implode( "\n", $copiedRel ) . "\n\n" );
 233                 }
 234         }
 235
 236         protected function filesAreSame( FileBackend $src, FileBackend $dst, $sPath, $dPath ) {
 237                 $skipHash = $this->hasOption( 'skiphash' );
 238                 $srcStat = $src->getFileStat( array( 'src' => $sPath ) );
 239                 $dPathSha1 = sha1( $dPath );
 240                 $dstStat = isset( $this->statCache[$dPathSha1] )
 241                         ? $this->statCache[$dPathSha1]
 242                         : $dst->getFileStat( array( 'src' => $dPath ) );
 243                 return (
 244                         is_array( $srcStat ) // sanity check that source exists
 245                         && is_array( $dstStat ) // dest exists
 246                         && $srcStat['size'] === $dstStat['size']
 247                         && ( !$skipHash || $srcStat['mtime'] <= $dstStat['mtime'] )
 248                         && ( $skipHash || $src->getFileSha1Base36( array( 'src' => $sPath, 'latest' => 1 ) )
 249                                 === $dst->getFileSha1Base36( array( 'src' => $dPath, 'latest' => 1 ) )
 250                         )
 251                 );
 252         }
 253 }
 254
 255 $maintClass = 'CopyFileBackend';
 256 require_once RUN_MAINTENANCE_IF_MAIN;