Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 188 |
|
0.00% |
0 / 8 |
CRAP | |
0.00% |
0 / 1 |
| MoveToExternal | |
0.00% |
0 / 185 |
|
0.00% |
0 / 8 |
2352 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
2 | |||
| execute | |
0.00% |
0 / 33 |
|
0.00% |
0 / 1 |
56 | |||
| doMoveToExternal | |
0.00% |
0 / 86 |
|
0.00% |
0 / 1 |
420 | |||
| compress | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
12 | |||
| resolveLegacyEncoding | |
0.00% |
0 / 19 |
|
0.00% |
0 / 1 |
72 | |||
| resolveStubs | |
0.00% |
0 / 21 |
|
0.00% |
0 / 1 |
56 | |||
| getConditions | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
| resolveText | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Move text from the text table to external storage |
| 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 | use MediaWiki\MainConfigNames; |
| 25 | use MediaWiki\Maintenance\UndoLog; |
| 26 | use MediaWiki\Storage\SqlBlobStore; |
| 27 | use Wikimedia\AtEase\AtEase; |
| 28 | use Wikimedia\Rdbms\IExpression; |
| 29 | use Wikimedia\Rdbms\LikeValue; |
| 30 | |
| 31 | require_once __DIR__ . '/../Maintenance.php'; |
| 32 | |
| 33 | class MoveToExternal extends Maintenance { |
| 34 | /** @var ResolveStubs */ |
| 35 | private $resolveStubs; |
| 36 | /** @var int */ |
| 37 | private $reportingInterval; |
| 38 | /** @var int */ |
| 39 | private $minID; |
| 40 | /** @var int */ |
| 41 | private $maxID; |
| 42 | /** @var string */ |
| 43 | private $esType; |
| 44 | /** @var string */ |
| 45 | private $esLocation; |
| 46 | /** @var int */ |
| 47 | private $threshold; |
| 48 | /** @var bool */ |
| 49 | private $gzip; |
| 50 | /** @var bool */ |
| 51 | private $skipResolve; |
| 52 | /** @var string|null */ |
| 53 | private $legacyEncoding; |
| 54 | /** @var bool */ |
| 55 | private $dryRun; |
| 56 | /** @var UndoLog */ |
| 57 | private $undoLog; |
| 58 | |
| 59 | public function __construct() { |
| 60 | parent::__construct(); |
| 61 | |
| 62 | $this->setBatchSize( 1000 ); |
| 63 | |
| 64 | $this->addOption( 'start', 'start old_id', false, true, 's' ); |
| 65 | $this->addOption( 'end', 'end old_id', false, true, 'e' ); |
| 66 | $this->addOption( 'threshold', 'minimum size in bytes', false, true ); |
| 67 | $this->addOption( 'reporting-interval', |
| 68 | 'show a message after this many revisions', false, true ); |
| 69 | $this->addOption( 'undo', 'filename for undo SQL', false, true ); |
| 70 | |
| 71 | $this->addOption( 'skip-gzip', 'Don\'t compress individual revisions' ); |
| 72 | $this->addOption( 'skip-resolve', |
| 73 | 'Don\'t replace HistoryBlobStub objects with direct external store pointers' ); |
| 74 | $this->addOption( 'iconv', 'Resolve legacy character encoding' ); |
| 75 | $this->addOption( 'dry-run', 'Don\'t modify any rows' ); |
| 76 | |
| 77 | $this->addArg( 'type', 'The external store type, e.g. "DB" or "mwstore"' ); |
| 78 | $this->addArg( 'location', 'e.g. "cluster12" or "global-swift"' ); |
| 79 | } |
| 80 | |
| 81 | public function execute() { |
| 82 | $this->resolveStubs = new ResolveStubs; |
| 83 | $this->esType = $this->getArg( 0 ); // e.g. "DB" or "mwstore" |
| 84 | $this->esLocation = $this->getArg( 1 ); // e.g. "cluster12" or "global-swift" |
| 85 | $dbw = $this->getPrimaryDB(); |
| 86 | |
| 87 | $maxID = $this->getOption( 'end' ); |
| 88 | if ( $maxID === null ) { |
| 89 | $maxID = $dbw->newSelectQueryBuilder() |
| 90 | ->select( 'MAX(old_id)' ) |
| 91 | ->from( 'text' ) |
| 92 | ->caller( __METHOD__ )->fetchField(); |
| 93 | } |
| 94 | $this->maxID = (int)$maxID; |
| 95 | $this->minID = (int)$this->getOption( 'start', 1 ); |
| 96 | |
| 97 | $this->reportingInterval = $this->getOption( 'reporting-interval', 100 ); |
| 98 | $this->threshold = (int)$this->getOption( 'threshold', 0 ); |
| 99 | |
| 100 | if ( $this->getOption( 'skip-gzip' ) ) { |
| 101 | $this->gzip = false; |
| 102 | } elseif ( !function_exists( 'gzdeflate' ) ) { |
| 103 | $this->fatalError( "gzdeflate() not found. " . |
| 104 | "Please run with --skip-gzip if you don't want to compress revisions." ); |
| 105 | } else { |
| 106 | $this->gzip = true; |
| 107 | } |
| 108 | |
| 109 | $this->skipResolve = $this->getOption( 'skip-resolve' ); |
| 110 | |
| 111 | if ( $this->getOption( 'iconv' ) ) { |
| 112 | $legacyEncoding = $this->getConfig()->get( MainConfigNames::LegacyEncoding ); |
| 113 | if ( $legacyEncoding ) { |
| 114 | $this->legacyEncoding = $legacyEncoding; |
| 115 | } else { |
| 116 | $this->output( "iconv requested but the wiki has no legacy encoding\n" ); |
| 117 | } |
| 118 | } |
| 119 | $this->dryRun = $this->getOption( 'dry-run', false ); |
| 120 | |
| 121 | $undo = $this->getOption( 'undo' ); |
| 122 | try { |
| 123 | $this->undoLog = new UndoLog( $undo, $dbw ); |
| 124 | } catch ( RuntimeException $e ) { |
| 125 | $this->fatalError( "Unable to open undo log" ); |
| 126 | } |
| 127 | $this->resolveStubs->setUndoLog( $this->undoLog ); |
| 128 | |
| 129 | return $this->doMoveToExternal(); |
| 130 | } |
| 131 | |
| 132 | private function doMoveToExternal() { |
| 133 | $success = true; |
| 134 | $dbr = $this->getReplicaDB(); |
| 135 | |
| 136 | $count = $this->maxID - $this->minID + 1; |
| 137 | $blockSize = $this->getBatchSize(); |
| 138 | $numBlocks = ceil( $count / $blockSize ); |
| 139 | print "Moving text rows from {$this->minID} to {$this->maxID} to external storage\n"; |
| 140 | |
| 141 | $esFactory = $this->getServiceContainer()->getExternalStoreFactory(); |
| 142 | $extStore = $esFactory->getStore( $this->esType ); |
| 143 | $numMoved = 0; |
| 144 | $stubIDs = []; |
| 145 | |
| 146 | for ( $block = 0; $block < $numBlocks; $block++ ) { |
| 147 | $blockStart = $block * $blockSize + $this->minID; |
| 148 | $blockEnd = $blockStart + $blockSize - 1; |
| 149 | |
| 150 | if ( $this->reportingInterval && !( $block % $this->reportingInterval ) ) { |
| 151 | $this->output( "oldid=$blockStart, moved=$numMoved\n" ); |
| 152 | $this->waitForReplication(); |
| 153 | } |
| 154 | |
| 155 | $res = $dbr->newSelectQueryBuilder() |
| 156 | ->select( [ 'old_id', 'old_flags', 'old_text' ] ) |
| 157 | ->from( 'text' ) |
| 158 | ->where( $this->getConditions( $blockStart, $blockEnd, $dbr ) ) |
| 159 | ->caller( __METHOD__ )->fetchResultSet(); |
| 160 | foreach ( $res as $row ) { |
| 161 | $text = $row->old_text; |
| 162 | $id = $row->old_id; |
| 163 | $flags = SqlBlobStore::explodeFlags( $row->old_flags ); |
| 164 | [ $text, $flags ] = $this->resolveText( $text, $flags ); |
| 165 | |
| 166 | if ( $text === false ) { |
| 167 | $success = false; |
| 168 | } |
| 169 | |
| 170 | if ( in_array( 'error', $flags ) ) { |
| 171 | continue; |
| 172 | } elseif ( in_array( 'object', $flags ) ) { |
| 173 | $obj = unserialize( $text ); |
| 174 | if ( $obj instanceof HistoryBlobStub ) { |
| 175 | // Handle later, after CGZ resolution |
| 176 | if ( !$this->skipResolve ) { |
| 177 | $stubIDs[] = $id; |
| 178 | } |
| 179 | continue; |
| 180 | } elseif ( $obj instanceof HistoryBlobCurStub ) { |
| 181 | // Copy cur text to ES |
| 182 | $newText = $obj->getText(); |
| 183 | if ( $newText === false ) { |
| 184 | print "Warning: Could not fetch revision blob {$id}: {$text}\n"; |
| 185 | $success = false; |
| 186 | continue; |
| 187 | } |
| 188 | |
| 189 | [ $text, $flags ] = $this->resolveLegacyEncoding( $newText, [] ); |
| 190 | |
| 191 | if ( $text === false ) { |
| 192 | print "Warning: Could not decode legacy-encoded gzip\'ed revision blob {$id}: {$newText}\n"; |
| 193 | $success = false; |
| 194 | continue; |
| 195 | } |
| 196 | |
| 197 | [ $text, $flags ] = $this->compress( $text, $flags ); |
| 198 | } elseif ( $obj instanceof ConcatenatedGzipHistoryBlob ) { |
| 199 | // Store as is |
| 200 | } else { |
| 201 | $className = get_class( $obj ); |
| 202 | print "Warning: old_id=$id unrecognised object class \"$className\"\n"; |
| 203 | $success = false; |
| 204 | continue; |
| 205 | } |
| 206 | } elseif ( strlen( $text ) < $this->threshold ) { |
| 207 | // Don't move small revisions |
| 208 | continue; |
| 209 | } else { |
| 210 | [ $text, $flags ] = $this->resolveLegacyEncoding( $text, $flags ); |
| 211 | [ $newText, $flags ] = $this->compress( $text, $flags ); |
| 212 | if ( $newText === false ) { |
| 213 | print "Warning: Could not compress revision blob {$id}: {$text}\n"; |
| 214 | $success = false; |
| 215 | continue; |
| 216 | } |
| 217 | $text = $newText; |
| 218 | } |
| 219 | $flags[] = 'external'; |
| 220 | $flagsString = implode( ',', $flags ); |
| 221 | |
| 222 | if ( $this->dryRun ) { |
| 223 | $this->output( "Move $id => $flagsString " . |
| 224 | addcslashes( substr( $text, 0, 30 ), "\0..\x1f\x7f..\xff" ) . |
| 225 | "\n" |
| 226 | ); |
| 227 | continue; |
| 228 | } |
| 229 | |
| 230 | $url = $extStore->store( $this->esLocation, $text ); |
| 231 | if ( !$url ) { |
| 232 | $this->fatalError( "Error writing to external storage" ); |
| 233 | } |
| 234 | $moved = $this->undoLog->update( |
| 235 | 'text', |
| 236 | [ 'old_flags' => $flagsString, 'old_text' => $url ], |
| 237 | (array)$row, |
| 238 | __METHOD__ |
| 239 | ); |
| 240 | if ( $moved ) { |
| 241 | $numMoved++; |
| 242 | } else { |
| 243 | print "Update of old_id $id failed, affected zero rows\n"; |
| 244 | $success = false; |
| 245 | } |
| 246 | } |
| 247 | } |
| 248 | |
| 249 | if ( count( $stubIDs ) ) { |
| 250 | $this->resolveStubs( $stubIDs ); |
| 251 | } |
| 252 | |
| 253 | return $success; |
| 254 | } |
| 255 | |
| 256 | private function compress( $text, $flags ) { |
| 257 | if ( $this->gzip && !in_array( 'gzip', $flags ) ) { |
| 258 | $flags[] = 'gzip'; |
| 259 | $text = gzdeflate( $text ); |
| 260 | } |
| 261 | return [ $text, $flags ]; |
| 262 | } |
| 263 | |
| 264 | private function resolveLegacyEncoding( $text, $flags ) { |
| 265 | if ( $this->legacyEncoding !== null |
| 266 | && !in_array( 'utf-8', $flags ) |
| 267 | && !in_array( 'utf8', $flags ) |
| 268 | ) { |
| 269 | // First decompress the entry so we don't try to convert a binary gzip to utf-8 |
| 270 | if ( in_array( 'gzip', $flags ) ) { |
| 271 | if ( !$this->gzip ) { |
| 272 | return [ $text, $flags ]; |
| 273 | } |
| 274 | $flags = array_diff( $flags, [ 'gzip' ] ); |
| 275 | $newText = gzinflate( $text ); |
| 276 | if ( $newText === false ) { |
| 277 | return [ false, $flags ]; |
| 278 | } |
| 279 | $text = $newText; |
| 280 | } |
| 281 | AtEase::suppressWarnings(); |
| 282 | $newText = iconv( $this->legacyEncoding, 'UTF-8//IGNORE', $text ); |
| 283 | AtEase::restoreWarnings(); |
| 284 | if ( $newText === false ) { |
| 285 | return [ false, $flags ]; |
| 286 | } |
| 287 | $text = $newText; |
| 288 | $flags[] = 'utf-8'; |
| 289 | } |
| 290 | return [ $text, $flags ]; |
| 291 | } |
| 292 | |
| 293 | private function resolveStubs( $stubIDs ) { |
| 294 | if ( $this->dryRun ) { |
| 295 | print "Note: resolving stubs in dry run mode is expected to fail, " . |
| 296 | "because the main blobs have not been moved to external storage.\n"; |
| 297 | } |
| 298 | |
| 299 | $dbr = $this->getReplicaDB(); |
| 300 | $this->output( "Resolving " . count( $stubIDs ) . " stubs\n" ); |
| 301 | $numResolved = 0; |
| 302 | $numTotal = 0; |
| 303 | foreach ( array_chunk( $stubIDs, $this->getBatchSize() ) as $stubBatch ) { |
| 304 | $res = $dbr->newSelectQueryBuilder() |
| 305 | ->select( [ 'old_id', 'old_flags', 'old_text' ] ) |
| 306 | ->from( 'text' ) |
| 307 | ->where( [ 'old_id' => $stubBatch ] ) |
| 308 | ->caller( __METHOD__ )->fetchResultSet(); |
| 309 | foreach ( $res as $row ) { |
| 310 | $numResolved += $this->resolveStubs->resolveStub( $row, $this->dryRun ) ? 1 : 0; |
| 311 | $numTotal++; |
| 312 | if ( $this->reportingInterval |
| 313 | && $numTotal % $this->reportingInterval == 0 |
| 314 | ) { |
| 315 | $this->output( "$numTotal stubs processed\n" ); |
| 316 | $this->waitForReplication(); |
| 317 | } |
| 318 | } |
| 319 | } |
| 320 | $this->output( "$numResolved of $numTotal stubs resolved\n" ); |
| 321 | } |
| 322 | |
| 323 | protected function getConditions( $blockStart, $blockEnd, $dbr ) { |
| 324 | return [ |
| 325 | $dbr->expr( 'old_id', '>=', $blockStart ), |
| 326 | $dbr->expr( 'old_id', '>=', $blockEnd ), |
| 327 | $dbr->expr( 'old_flags', IExpression::NOT_LIKE, |
| 328 | new LikeValue( $dbr->anyString(), 'external', $dbr->anyString() ) ), |
| 329 | ]; |
| 330 | } |
| 331 | |
| 332 | protected function resolveText( $text, $flags ) { |
| 333 | return [ $text, $flags ]; |
| 334 | } |
| 335 | } |
| 336 | |
| 337 | $maintClass = MoveToExternal::class; |
| 338 | require_once RUN_MAINTENANCE_IF_MAIN; |