maintenance/storage/fixT22757.php

   1 <?php
   2 /**
   3  * Script to fix T22757.
   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 ExternalStorage
  22  */
  23
  24 require_once __DIR__ . '/../Maintenance.php';
  25
  26 /**
  27  * Maintenance script to fix T22757.
  28  *
  29  * @ingroup Maintenance ExternalStorage
  30  */
  31 class FixT22757 extends Maintenance {
  32         public $batchSize = 10000;
  33         public $mapCache = [];
  34         public $mapCacheSize = 0;
  35         public $maxMapCacheSize = 1000000;
  36
  37         function __construct() {
  38                 parent::__construct();
  39                 $this->addDescription( 'Script to fix T22757 assuming that blob_tracking is intact' );
  40                 $this->addOption( 'dry-run', 'Report only' );
  41                 $this->addOption( 'start', 'old_id to start at', false, true );
  42         }
  43
  44         function execute() {
  45                 $dbr = $this->getDB( DB_REPLICA );
  46                 $dbw = $this->getDB( DB_MASTER );
  47
  48                 $dryRun = $this->getOption( 'dry-run' );
  49                 if ( $dryRun ) {
  50                         print "Dry run only.\n";
  51                 }
  52
  53                 $startId = $this->getOption( 'start', 0 );
  54                 $numGood = 0;
  55                 $numFixed = 0;
  56                 $numBad = 0;
  57
  58                 $totalRevs = $dbr->selectField( 'text', 'MAX(old_id)', '', __METHOD__ );
  59
  60                 // In MySQL 4.1+, the binary field old_text has a non-working LOWER() function
  61                 $lowerLeft = 'LOWER(CONVERT(LEFT(old_text,22) USING latin1))';
  62
  63                 while ( true ) {
  64                         print "ID: $startId / $totalRevs\r";
  65
  66                         $res = $dbr->select(
  67                                 'text',
  68                                 [ 'old_id', 'old_flags', 'old_text' ],
  69                                 [
  70                                         'old_id > ' . intval( $startId ),
  71                                         'old_flags LIKE \'%object%\' AND old_flags NOT LIKE \'%external%\'',
  72                                         "$lowerLeft = 'o:15:\"historyblobstub\"'",
  73                                 ],
  74                                 __METHOD__,
  75                                 [
  76                                         'ORDER BY' => 'old_id',
  77                                         'LIMIT' => $this->batchSize,
  78                                 ]
  79                         );
  80
  81                         if ( !$res->numRows() ) {
  82                                 break;
  83                         }
  84
  85                         $secondaryIds = [];
  86                         $stubs = [];
  87
  88                         foreach ( $res as $row ) {
  89                                 $startId = $row->old_id;
  90
  91                                 // Basic sanity checks
  92                                 $obj = unserialize( $row->old_text );
  93                                 if ( $obj === false ) {
  94                                         print "{$row->old_id}: unrecoverable: cannot unserialize\n";
  95                                         ++$numBad;
  96                                         continue;
  97                                 }
  98
  99                                 if ( !is_object( $obj ) ) {
 100                                         print "{$row->old_id}: unrecoverable: unserialized to type " .
 101                                                 gettype( $obj ) . ", possible double-serialization\n";
 102                                         ++$numBad;
 103                                         continue;
 104                                 }
 105
 106                                 if ( strtolower( get_class( $obj ) ) !== 'historyblobstub' ) {
 107                                         print "{$row->old_id}: unrecoverable: unexpected object class " .
 108                                                 get_class( $obj ) . "\n";
 109                                         ++$numBad;
 110                                         continue;
 111                                 }
 112
 113                                 // Process flags
 114                                 $flags = explode( ',', $row->old_flags );
 115                                 if ( in_array( 'utf-8', $flags ) || in_array( 'utf8', $flags ) ) {
 116                                         $legacyEncoding = false;
 117                                 } else {
 118                                         $legacyEncoding = true;
 119                                 }
 120
 121                                 // Queue the stub for future batch processing
 122                                 $id = intval( $obj->mOldId );
 123                                 $secondaryIds[] = $id;
 124                                 $stubs[$row->old_id] = [
 125                                         'legacyEncoding' => $legacyEncoding,
 126                                         'secondaryId' => $id,
 127                                         'hash' => $obj->mHash,
 128                                 ];
 129                         }
 130
 131                         $secondaryIds = array_unique( $secondaryIds );
 132
 133                         if ( !count( $secondaryIds ) ) {
 134                                 continue;
 135                         }
 136
 137                         // Run the batch query on blob_tracking
 138                         $res = $dbr->select(
 139                                 'blob_tracking',
 140                                 '*',
 141                                 [
 142                                         'bt_text_id' => $secondaryIds,
 143                                 ],
 144                                 __METHOD__
 145                         );
 146                         $trackedBlobs = [];
 147                         foreach ( $res as $row ) {
 148                                 $trackedBlobs[$row->bt_text_id] = $row;
 149                         }
 150
 151                         // Process the stubs
 152                         foreach ( $stubs as $primaryId => $stub ) {
 153                                 $secondaryId = $stub['secondaryId'];
 154                                 if ( !isset( $trackedBlobs[$secondaryId] ) ) {
 155                                         // No tracked blob. Work out what went wrong
 156                                         $secondaryRow = $dbr->selectRow(
 157                                                 'text',
 158                                                 [ 'old_flags', 'old_text' ],
 159                                                 [ 'old_id' => $secondaryId ],
 160                                                 __METHOD__
 161                                         );
 162                                         if ( !$secondaryRow ) {
 163                                                 print "$primaryId: unrecoverable: secondary row is missing\n";
 164                                                 ++$numBad;
 165                                         } elseif ( $this->isUnbrokenStub( $stub, $secondaryRow ) ) {
 166                                                 // Not broken yet, and not in the tracked clusters so it won't get
 167                                                 // broken by the current RCT run.
 168                                                 ++$numGood;
 169                                         } elseif ( strpos( $secondaryRow->old_flags, 'external' ) !== false ) {
 170                                                 print "$primaryId: unrecoverable: secondary gone to {$secondaryRow->old_text}\n";
 171                                                 ++$numBad;
 172                                         } else {
 173                                                 print "$primaryId: unrecoverable: miscellaneous corruption of secondary row\n";
 174                                                 ++$numBad;
 175                                         }
 176                                         unset( $stubs[$primaryId] );
 177                                         continue;
 178                                 }
 179                                 $trackRow = $trackedBlobs[$secondaryId];
 180
 181                                 // Check that the specified text really is available in the tracked source row
 182                                 $url = "DB://{$trackRow->bt_cluster}/{$trackRow->bt_blob_id}/{$stub['hash']}";
 183                                 $text = ExternalStore::fetchFromURL( $url );
 184                                 if ( $text === false ) {
 185                                         print "$primaryId: unrecoverable: source text missing\n";
 186                                         ++$numBad;
 187                                         unset( $stubs[$primaryId] );
 188                                         continue;
 189                                 }
 190                                 if ( md5( $text ) !== $stub['hash'] ) {
 191                                         print "$primaryId: unrecoverable: content hashes do not match\n";
 192                                         ++$numBad;
 193                                         unset( $stubs[$primaryId] );
 194                                         continue;
 195                                 }
 196
 197                                 // Find the page_id and rev_id
 198                                 // The page is probably the same as the page of the secondary row
 199                                 $pageId = intval( $trackRow->bt_page );
 200                                 if ( !$pageId ) {
 201                                         $revId = $pageId = 0;
 202                                 } else {
 203                                         $revId = $this->findTextIdInPage( $pageId, $primaryId );
 204                                         if ( !$revId ) {
 205                                                 // Actually an orphan
 206                                                 $pageId = $revId = 0;
 207                                         }
 208                                 }
 209
 210                                 $newFlags = $stub['legacyEncoding'] ? 'external' : 'external,utf-8';
 211
 212                                 if ( !$dryRun ) {
 213                                         // Reset the text row to point to the original copy
 214                                         $this->beginTransaction( $dbw, __METHOD__ );
 215                                         $dbw->update(
 216                                                 'text',
 217                                                 // SET
 218                                                 [
 219                                                         'old_flags' => $newFlags,
 220                                                         'old_text' => $url
 221                                                 ],
 222                                                 // WHERE
 223                                                 [ 'old_id' => $primaryId ],
 224                                                 __METHOD__
 225                                         );
 226
 227                                         // Add a blob_tracking row so that the new reference can be recompressed
 228                                         // without needing to run trackBlobs.php again
 229                                         $dbw->insert( 'blob_tracking',
 230                                                 [
 231                                                         'bt_page' => $pageId,
 232                                                         'bt_rev_id' => $revId,
 233                                                         'bt_text_id' => $primaryId,
 234                                                         'bt_cluster' => $trackRow->bt_cluster,
 235                                                         'bt_blob_id' => $trackRow->bt_blob_id,
 236                                                         'bt_cgz_hash' => $stub['hash'],
 237                                                         'bt_new_url' => null,
 238                                                         'bt_moved' => 0,
 239                                                 ],
 240                                                 __METHOD__
 241                                         );
 242                                         $this->commitTransaction( $dbw, __METHOD__ );
 243                                 }
 244
 245                                 print "$primaryId: resolved to $url\n";
 246                                 ++$numFixed;
 247                         }
 248                 }
 249
 250                 print "\n";
 251                 print "Fixed: $numFixed\n";
 252                 print "Unrecoverable: $numBad\n";
 253                 print "Good stubs: $numGood\n";
 254         }
 255
 256         function findTextIdInPage( $pageId, $textId ) {
 257                 $ids = $this->getRevTextMap( $pageId );
 258                 if ( !isset( $ids[$textId] ) ) {
 259                         return null;
 260                 } else {
 261                         return $ids[$textId];
 262                 }
 263         }
 264
 265         function getRevTextMap( $pageId ) {
 266                 if ( !isset( $this->mapCache[$pageId] ) ) {
 267                         // Limit cache size
 268                         while ( $this->mapCacheSize > $this->maxMapCacheSize ) {
 269                                 $key = key( $this->mapCache );
 270                                 $this->mapCacheSize -= count( $this->mapCache[$key] );
 271                                 unset( $this->mapCache[$key] );
 272                         }
 273
 274                         $dbr = $this->getDB( DB_REPLICA );
 275                         $map = [];
 276                         $res = $dbr->select( 'revision',
 277                                 [ 'rev_id', 'rev_text_id' ],
 278                                 [ 'rev_page' => $pageId ],
 279                                 __METHOD__
 280                         );
 281                         foreach ( $res as $row ) {
 282                                 $map[$row->rev_text_id] = $row->rev_id;
 283                         }
 284                         $this->mapCache[$pageId] = $map;
 285                         $this->mapCacheSize += count( $map );
 286                 }
 287
 288                 return $this->mapCache[$pageId];
 289         }
 290
 291         /**
 292          * This is based on part of HistoryBlobStub::getText().
 293          * Determine if the text can be retrieved from the row in the normal way.
 294          * @param array $stub
 295          * @param stdClass $secondaryRow
 296          * @return bool
 297          */
 298         function isUnbrokenStub( $stub, $secondaryRow ) {
 299                 $flags = explode( ',', $secondaryRow->old_flags );
 300                 $text = $secondaryRow->old_text;
 301                 if ( in_array( 'external', $flags ) ) {
 302                         $url = $text;
 303                         Wikimedia\suppressWarnings();
 304                         list( /* $proto */, $path ) = explode( '://', $url, 2 );
 305                         Wikimedia\restoreWarnings();
 306
 307                         if ( $path == "" ) {
 308                                 return false;
 309                         }
 310                         $text = ExternalStore::fetchFromURL( $url );
 311                 }
 312                 if ( !in_array( 'object', $flags ) ) {
 313                         return false;
 314                 }
 315
 316                 if ( in_array( 'gzip', $flags ) ) {
 317                         $obj = unserialize( gzinflate( $text ) );
 318                 } else {
 319                         $obj = unserialize( $text );
 320                 }
 321
 322                 if ( !is_object( $obj ) ) {
 323                         // Correct for old double-serialization bug.
 324                         $obj = unserialize( $obj );
 325                 }
 326
 327                 if ( !is_object( $obj ) ) {
 328                         return false;
 329                 }
 330
 331                 $obj->uncompress();
 332                 $text = $obj->getItem( $stub['hash'] );
 333
 334                 return $text !== false;
 335         }
 336 }
 337
 338 $maintClass = FixT22757::class;
 339 require_once RUN_MAINTENANCE_IF_MAIN;