maintenance/storage/compressOld.inc

   1 <?php
   2 /**
   3  * Helper functions for compressOld.php script.
   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 /** @todo document */
  25 function compressOldPages( $start = 0, $extdb = '' ) {
  26         $fname = 'compressOldPages';
  27
  28         $chunksize = 50;
  29         print "Starting from old_id $start...\n";
  30         $dbw = wfGetDB( DB_MASTER );
  31         do {
  32                 $res = $dbw->select( 'text', array( 'old_id','old_flags','old_text' ),
  33                         "old_id>=$start", $fname, array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ) );
  34                 if( $dbw->numRows( $res ) == 0 ) {
  35                         break;
  36                 }
  37                 $last = $start;
  38                 foreach ( $res as $row ) {
  39                         # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
  40                         compressPage( $row, $extdb );
  41                         $last = $row->old_id;
  42                 }
  43                 $start = $last + 1; # Deletion may leave long empty stretches
  44                 print "$start...\n";
  45         } while( true );
  46 }
  47
  48 /** @todo document */
  49 function compressPage( $row, $extdb ) {
  50         $fname = 'compressPage';
  51         if ( false !== strpos( $row->old_flags, 'gzip' ) || false !== strpos( $row->old_flags, 'object' ) ) {
  52                 #print "Already compressed row {$row->old_id}\n";
  53                 return false;
  54         }
  55         $dbw = wfGetDB( DB_MASTER );
  56         $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
  57         $compress = gzdeflate( $row->old_text );
  58
  59         # Store in external storage if required
  60         if ( $extdb !== '' ) {
  61                 $storeObj = new ExternalStoreDB;
  62                 $compress = $storeObj->store( $extdb, $compress );
  63                 if ( $compress === false ) {
  64                         print "Unable to store object\n";
  65                         return false;
  66                 }
  67         }
  68
  69         # Update text row
  70         $dbw->update( 'text',
  71                 array( /* SET */
  72                         'old_flags' => $flags,
  73                         'old_text' => $compress
  74                 ), array( /* WHERE */
  75                         'old_id' => $row->old_id
  76                 ), $fname,
  77                 array( 'LIMIT' => 1 )
  78         );
  79         return true;
  80 }
  81
  82 define( 'LS_INDIVIDUAL', 0 );
  83 define( 'LS_CHUNKED', 1 );
  84
  85 /** @todo document */
  86 function compressWithConcat( $startId, $maxChunkSize, $beginDate,
  87         $endDate, $extdb="", $maxPageId = false )
  88 {
  89         $fname = 'compressWithConcat';
  90         $loadStyle = LS_CHUNKED;
  91
  92         $dbr = wfGetDB( DB_SLAVE );
  93         $dbw = wfGetDB( DB_MASTER );
  94
  95         # Set up external storage
  96         if ( $extdb != '' ) {
  97                 $storeObj = new ExternalStoreDB;
  98         }
  99
 100         # Get all articles by page_id
 101         if ( !$maxPageId ) {
 102                 $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', $fname );
 103         }
 104         print "Starting from $startId of $maxPageId\n";
 105         $pageConds = array();
 106
 107         /*
 108         if ( $exclude_ns0 ) {
 109                 print "Excluding main namespace\n";
 110                 $pageConds[] = 'page_namespace<>0';
 111         }
 112         if ( $queryExtra ) {
 113                                 $pageConds[] = $queryExtra;
 114         }
 115          */
 116
 117         # For each article, get a list of revisions which fit the criteria
 118
 119         # No recompression, use a condition on old_flags
 120         # Don't compress object type entities, because that might produce data loss when
 121         # overwriting bulk storage concat rows. Don't compress external references, because
 122         # the script doesn't yet delete rows from external storage.
 123         $conds = array(
 124                 'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() ) . ' AND old_flags NOT '
 125                         . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ) );
 126
 127         if ( $beginDate ) {
 128                 if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 129                         print "Invalid begin date \"$beginDate\"\n";
 130                         return false;
 131                 }
 132                 $conds[] = "rev_timestamp>'" . $beginDate . "'";
 133         }
 134         if ( $endDate )  {
 135                 if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 136                         print "Invalid end date \"$endDate\"\n";
 137                         return false;
 138                 }
 139                 $conds[] = "rev_timestamp<'" . $endDate . "'";
 140         }
 141         if ( $loadStyle == LS_CHUNKED ) {
 142                 $tables = array( 'revision', 'text' );
 143                 $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 144                 $conds[] = 'rev_text_id=old_id';
 145                 $revLoadOptions = 'FOR UPDATE';
 146         } else {
 147                 $tables = array( 'revision' );
 148                 $fields = array( 'rev_id', 'rev_text_id' );
 149                 $revLoadOptions = array();
 150         }
 151
 152         # Don't work with current revisions
 153         # Don't lock the page table for update either -- TS 2006-04-04
 154         #$tables[] = 'page';
 155         #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 156
 157         for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 158                 wfWaitForSlaves();
 159
 160                 # Wake up
 161                 $dbr->ping();
 162
 163                 # Get the page row
 164                 $pageRes = $dbr->select( 'page',
 165                         array('page_id', 'page_namespace', 'page_title','page_latest'),
 166                         $pageConds + array('page_id' => $pageId), $fname );
 167                 if ( $dbr->numRows( $pageRes ) == 0 ) {
 168                         continue;
 169                 }
 170                 $pageRow = $dbr->fetchObject( $pageRes );
 171
 172                 # Display progress
 173                 $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 174                 print "$pageId\t" . $titleObj->getPrefixedDBkey() . " ";
 175
 176                 # Load revisions
 177                 $revRes = $dbw->select( $tables, $fields,
 178                         array_merge( array(
 179                                 'rev_page' => $pageRow->page_id,
 180                                 # Don't operate on the current revision
 181                                 # Use < instead of <> in case the current revision has changed
 182                                 # since the page select, which wasn't locking
 183                                 'rev_id < ' . $pageRow->page_latest
 184                         ), $conds ),
 185                         $fname,
 186                         $revLoadOptions
 187                 );
 188                 $revs = array();
 189                 foreach ( $revRes as $revRow ) {
 190                         $revs[] = $revRow;
 191                 }
 192
 193                 if ( count( $revs ) < 2) {
 194                         # No revisions matching, no further processing
 195                         print "\n";
 196                         continue;
 197                 }
 198
 199                 # For each chunk
 200                 $i = 0;
 201                 while ( $i < count( $revs ) ) {
 202                         if ( $i < count( $revs ) - $maxChunkSize ) {
 203                                 $thisChunkSize = $maxChunkSize;
 204                         } else {
 205                                 $thisChunkSize = count( $revs ) - $i;
 206                         }
 207
 208                         $chunk = new ConcatenatedGzipHistoryBlob();
 209                         $stubs = array();
 210                         $dbw->begin();
 211                         $usedChunk = false;
 212                         $primaryOldid = $revs[$i]->rev_text_id;
 213
 214                         # Get the text of each revision and add it to the object
 215                         for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 216                                 $oldid = $revs[$i + $j]->rev_text_id;
 217
 218                                 # Get text
 219                                 if ( $loadStyle == LS_INDIVIDUAL ) {
 220                                         $textRow = $dbw->selectRow( 'text',
 221                                                 array( 'old_flags', 'old_text' ),
 222                                                 array( 'old_id' => $oldid ),
 223                                                 $fname,
 224                                                 'FOR UPDATE'
 225                                         );
 226                                         $text = Revision::getRevisionText( $textRow );
 227                                 } else {
 228                                         $text = Revision::getRevisionText( $revs[$i + $j] );
 229                                 }
 230
 231                                 if ( $text === false ) {
 232                                         print "\nError, unable to get text in old_id $oldid\n";
 233                                         #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 234                                 }
 235
 236                                 if ( $extdb == "" && $j == 0 ) {
 237                                         $chunk->setText( $text );
 238                                         print '.';
 239                                 } else {
 240                                         # Don't make a stub if it's going to be longer than the article
 241                                         # Stubs are typically about 100 bytes
 242                                         if ( strlen( $text ) < 120 ) {
 243                                                 $stub = false;
 244                                                 print 'x';
 245                                         } else {
 246                                                 $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 247                                                 $stub->setLocation( $primaryOldid );
 248                                                 $stub->setReferrer( $oldid );
 249                                                 print '.';
 250                                                 $usedChunk = true;
 251                                         }
 252                                         $stubs[$j] = $stub;
 253                                 }
 254                         }
 255                         $thisChunkSize = $j;
 256
 257                         # If we couldn't actually use any stubs because the pages were too small, do nothing
 258                         if ( $usedChunk ) {
 259                                 if ( $extdb != "" ) {
 260                                         # Move blob objects to External Storage
 261                                         $stored = $storeObj->store( $extdb, serialize( $chunk ));
 262                                         if ($stored === false) {
 263                                                 print "Unable to store object\n";
 264                                                 return false;
 265                                         }
 266                                         # Store External Storage URLs instead of Stub placeholders
 267                                         foreach ($stubs as $stub) {
 268                                                 if ($stub===false)
 269                                                         continue;
 270                                                 # $stored should provide base path to a BLOB
 271                                                 $url = $stored."/".$stub->getHash();
 272                                                 $dbw->update( 'text',
 273                                                         array( /* SET */
 274                                                                 'old_text' => $url,
 275                                                                 'old_flags' => 'external,utf-8',
 276                                                         ), array ( /* WHERE */
 277                                                                 'old_id' => $stub->getReferrer(),
 278                                                         )
 279                                                 );
 280                                         }
 281                                 } else {
 282                                         # Store the main object locally
 283                                         $dbw->update( 'text',
 284                                                 array( /* SET */
 285                                                         'old_text' => serialize( $chunk ),
 286                                                         'old_flags' => 'object,utf-8',
 287                                                 ), array( /* WHERE */
 288                                                         'old_id' => $primaryOldid
 289                                                 )
 290                                         );
 291
 292                                         # Store the stub objects
 293                                         for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 294                                                 # Skip if not compressing and don't overwrite the first revision
 295                                                 if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 296                                                         $dbw->update( 'text',
 297                                                                 array( /* SET */
 298                                                                         'old_text' => serialize($stubs[$j]),
 299                                                                         'old_flags' => 'object,utf-8',
 300                                                                 ), array( /* WHERE */
 301                                                                         'old_id' => $revs[$i + $j]->rev_text_id
 302                                                                 )
 303                                                         );
 304                                                 }
 305                                         }
 306                                 }
 307                         }
 308                         # Done, next
 309                         print "/";
 310                         $dbw->commit();
 311                         $i += $thisChunkSize;
 312                         wfWaitForSlaves();
 313                 }
 314                 print "\n";
 315         }
 316         return true;
 317 }