MediaWiki  master
migrateArchiveText.php
Go to the documentation of this file.
1 <?php
26 
27 require_once __DIR__ . '/Maintenance.php';
28 
36  public function __construct() {
37  parent::__construct();
38  $this->addDescription(
39  'Migrates content from pre-1.5 ar_text and ar_flags columns to text storage'
40  );
41  $this->addOption(
42  'replace-missing',
43  "For rows with missing or unloadable data, throw away whatever is there and\n"
44  . "mark them as \"error\" in the database."
45  );
46  }
47 
48  protected function getUpdateKey() {
49  return __CLASS__;
50  }
51 
52  protected function doDBUpdates() {
53  $replaceMissing = $this->hasOption( 'replace-missing' );
54  $defaultExternalStore = $this->getConfig()->get( MainConfigNames::DefaultExternalStore );
55  $blobStore = MediaWikiServices::getInstance()
56  ->getBlobStoreFactory()
57  ->newSqlBlobStore();
58  $batchSize = $this->getBatchSize();
59 
60  $dbr = $this->getDB( DB_REPLICA, [ 'vslow' ] );
61  $dbw = $this->getDB( DB_PRIMARY );
62  if ( !$dbr->fieldExists( 'archive', 'ar_text', __METHOD__ ) ||
63  !$dbw->fieldExists( 'archive', 'ar_text', __METHOD__ )
64  ) {
65  $this->output( "No ar_text field, so nothing to migrate.\n" );
66  return true;
67  }
68 
69  $this->output( "Migrating ar_text to modern storage...\n" );
70  $last = 0;
71  $count = 0;
72  $errors = 0;
73  while ( true ) {
74  $res = $dbr->select(
75  'archive',
76  [ 'ar_id', 'ar_text', 'ar_flags' ],
77  [
78  'ar_text_id' => null,
79  "ar_id > $last",
80  ],
81  __METHOD__,
82  [ 'LIMIT' => $batchSize, 'ORDER BY' => [ 'ar_id' ] ]
83  );
84  $numRows = $res->numRows();
85 
86  foreach ( $res as $row ) {
87  $last = $row->ar_id;
88 
89  // Recompress the text (and store in external storage, if
90  // applicable) if it's not already in external storage.
91  $arFlags = explode( ',', $row->ar_flags );
92  if ( !in_array( 'external', $arFlags, true ) ) {
93  $data = $blobStore->decompressData( $row->ar_text, $arFlags );
94  if ( $data !== false ) {
95  $flags = $blobStore->compressData( $data );
96 
97  if ( $defaultExternalStore ) {
98  $data = ExternalStore::insertToDefault( $data );
99  if ( $flags ) {
100  $flags .= ',';
101  }
102  $flags .= 'external';
103  }
104  } elseif ( $replaceMissing ) {
105  $this->error( "Replacing missing data for row ar_id=$row->ar_id" );
106  $data = 'Missing data in migrateArchiveText.php on ' . date( 'c' );
107  $flags = 'error';
108  } else {
109  $this->error( "No data for row ar_id=$row->ar_id" );
110  $errors++;
111  continue;
112  }
113  } else {
114  $flags = $row->ar_flags;
115  $data = $row->ar_text;
116  }
117 
118  $this->beginTransaction( $dbw, __METHOD__ );
119  $dbw->insert(
120  'text',
121  [ 'old_text' => $data, 'old_flags' => $flags ],
122  __METHOD__
123  );
124  $id = $dbw->insertId();
125  $dbw->update(
126  'archive',
127  [ 'ar_text_id' => $id, 'ar_text' => '', 'ar_flags' => '' ],
128  [ 'ar_id' => $row->ar_id, 'ar_text_id' => null ],
129  __METHOD__
130  );
131  $count += $dbw->affectedRows();
132  $this->commitTransaction( $dbw, __METHOD__ );
133  }
134 
135  if ( $numRows < $batchSize ) {
136  // We must have reached the end
137  break;
138  }
139 
140  $this->output( "... $last\n" );
141  // $this->commitTransaction() already waited for replication; no need to re-wait here
142  }
143 
144  $this->output( "Completed ar_text migration, $count rows updated, $errors missing data.\n" );
145  if ( $errors ) {
146  $this->output( "Run with --replace-missing to overwrite missing data with an error message.\n" );
147  }
148 
149  return $errors === 0;
150  }
151 }
152 
153 $maintClass = MigrateArchiveText::class;
154 require_once RUN_MAINTENANCE_IF_MAIN;
static insertToDefault( $data, array $params=[])
Like insert() above, but does more of the work for us.
Class for scripts that perform database maintenance and want to log the update in updatelog so we can...
getDB( $db, $groups=[], $dbDomain=false)
Returns a database to be used by current maintenance script.
error( $err, $die=0)
Throw an error to the user.
beginTransaction(IDatabase $dbw, $fname)
Begin a transaction on a DB.
commitTransaction(IDatabase $dbw, $fname)
Commit the transaction on a DB handle and wait for replica DBs to catch up.
output( $out, $channel=null)
Throw some output to the user.
hasOption( $name)
Checks to see if a particular option was set.
getBatchSize()
Returns batch size.
addDescription( $text)
Set the description text.
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
A class containing constants representing the names of configuration variables.
MediaWikiServices is the service locator for the application scope of MediaWiki.
Maintenance script that migrates archive.ar_text and ar_flags to text storage.
doDBUpdates()
Do the actual work.
__construct()
Default constructor.
getUpdateKey()
Get the update key name to go in the update log table.
const DB_REPLICA
Definition: defines.php:25
const DB_PRIMARY
Definition: defines.php:27