<?php
$optionsWithArgs = RecompressTracked::getOptionsWithArgs();
-require( dirname( __FILE__ ) .'/../commandLine.inc' );
+require( dirname( __FILE__ ) . '/../commandLine.inc' );
if ( count( $args ) < 1 ) {
echo "Usage: php recompressTracked.php [options] <cluster> [... <cluster>...]
Moves blobs indexed by trackBlobs.php to a specified list of destination clusters, and recompresses them in the process. Restartable.
-Options:
- --procs <procs> Set the number of child processes (default 1)
+Options:
+
--procs <procs> Set the number of child processes (default 1)
--copy-only Copy only, do not update the text table. Restart without this option to complete.
--debug-log <file> Log debugging data to the specified file
--info-log <file> Log progress messages to the specified file
var $copyOnly = false;
var $isChild = false;
var $slaveId = false;
+ var $noCount = false;
var $debugLog, $infoLog, $criticalLog;
var $store;
static $optionsWithArgs = array( 'procs', 'slave-id', 'debug-log', 'info-log', 'critical-log' );
static $cmdLineOptionMap = array(
+ 'no-count' => 'noCount',
'procs' => 'numProcs',
'copy-only' => 'copyOnly',
'child' => 'isChild',
}
function logToFile( $msg, $file ) {
- $header = '[' . date('d\TH:i:s') . '] ' . wfHostname() . ' ' . posix_getpid();
+ $header = '[' . date( 'd\TH:i:s' ) . '] ' . wfHostname() . ' ' . posix_getpid();
if ( $this->slaveId !== false ) {
$header .= "({$this->slaveId})";
}
/**
* Wait until the selected slave has caught up to the master.
- * This allows us to use the slave for things that were committed in a
- * previous part of this batch process.
+ * This allows us to use the slave for things that were committed in a
+ * previous part of this batch process.
*/
function syncDBs() {
$dbw = wfGetDB( DB_MASTER );
$cmd .= " --$cmdOption";
}
}
- $cmd .= ' --child' .
+ $cmd .= ' --child' .
' --wiki ' . wfEscapeShellArg( wfWikiID() ) .
' ' . call_user_func_array( 'wfEscapeShellArg', $this->destClusters );
$this->slavePipes = $this->slaveProcs = array();
for ( $i = 0; $i < $this->numProcs; $i++ ) {
$pipes = false;
- $spec = array(
+ $spec = array(
array( 'pipe', 'r' ),
array( 'file', 'php://stdout', 'w' ),
array( 'file', 'php://stderr', 'w' )
function dispatch( /*...*/ ) {
$args = func_get_args();
$pipes = $this->slavePipes;
- $numPipes = stream_select( $x=array(), $pipes, $y=array(), 3600 );
+ $numPipes = stream_select( $x = array(), $pipes, $y = array(), 3600 );
if ( !$numPipes ) {
$this->critical( "Error waiting to write to slaves. Aborting" );
exit( 1 );
$dbr = wfGetDB( DB_SLAVE );
$i = 0;
$startId = 0;
- $numPages = $dbr->selectField( 'blob_tracking',
- 'COUNT(DISTINCT bt_page)',
- # A condition is required so that this query uses the index
- array( 'bt_moved' => 0 ),
- __METHOD__
- );
+ if ( $this->noCount ) {
+ $numPages = '[unknown]';
+ } else {
+ $numPages = $dbr->selectField( 'blob_tracking',
+ 'COUNT(DISTINCT bt_page)',
+ # A condition is required so that this query uses the index
+ array( 'bt_moved' => 0 ),
+ __METHOD__
+ );
+ }
if ( $this->copyOnly ) {
$this->info( "Copying pages..." );
} else {
$this->info( "Moving pages..." );
}
while ( true ) {
- $res = $dbr->select( 'blob_tracking',
+ $res = $dbr->select( 'blob_tracking',
array( 'bt_page' ),
- array(
+ array(
'bt_moved' => 0,
'bt_page > ' . $dbr->addQuotes( $startId )
),
__METHOD__,
- array(
- 'DISTINCT',
+ array(
+ 'DISTINCT',
'ORDER BY' => 'bt_page',
'LIMIT' => $this->batchSize,
)
if ( $current == $end || $this->numBatches >= $this->reportingInterval ) {
$this->numBatches = 0;
$this->info( "$label: $current / $end" );
- wfWaitForSlaves( 5 );
+ $this->waitForSlaves();
}
}
$dbr = wfGetDB( DB_SLAVE );
$startId = 0;
$i = 0;
- $numOrphans = $dbr->selectField( 'blob_tracking',
- 'COUNT(DISTINCT bt_text_id)',
- array( 'bt_moved' => 0, 'bt_page' => 0 ),
- __METHOD__ );
- if ( !$numOrphans ) {
- return;
+ if ( $this->noCount ) {
+ $numOrphans = '[unknown]';
+ } else {
+ $numOrphans = $dbr->selectField( 'blob_tracking',
+ 'COUNT(DISTINCT bt_text_id)',
+ array( 'bt_moved' => 0, 'bt_page' => 0 ),
+ __METHOD__ );
+ if ( !$numOrphans ) {
+ return;
+ }
}
if ( $this->copyOnly ) {
$this->info( "Copying orphans..." );
case 'quit':
return;
}
- wfWaitForSlaves( 5 );
+ $this->waitForSlaves();
}
}
$trx = new CgzCopyTransaction( $this, $this->pageBlobClass );
while ( true ) {
- $res = $dbr->select(
- array( 'blob_tracking', 'text' ),
+ $res = $dbr->select(
+ array( 'blob_tracking', 'text' ),
'*',
array(
'bt_page' => $pageId,
'bt_text_id=old_id',
),
__METHOD__,
- array(
+ array(
'ORDER BY' => 'bt_text_id',
'LIMIT' => $this->batchSize
)
$this->debug( "$titleText: committing blob with " . $trx->getSize() . " items" );
$trx->commit();
$trx = new CgzCopyTransaction( $this, $this->pageBlobClass );
+ $this->waitForSlaves();
}
}
$startId = $row->bt_text_id;
*
* This is done in a single transaction to provide restartable behaviour
* without data loss.
- *
+ *
* The transaction is kept short to reduce locking.
*/
function moveTextRow( $textId, $url ) {
$dbr = wfGetDB( DB_SLAVE );
$startId = 0;
- $conds = array_merge( $conds, array(
+ $conds = array_merge( $conds, array(
'bt_moved' => 0,
'bt_new_url IS NOT NULL'
- ));
+ ) );
while ( true ) {
$res = $dbr->select( 'blob_tracking',
'*',
array_merge( $conds, array( 'bt_text_id > ' . $dbr->addQuotes( $startId ) ) ),
__METHOD__,
- array(
+ array(
'ORDER BY' => 'bt_text_id',
'LIMIT' => $this->batchSize,
)
$this->debug( 'Incomplete: ' . $res->numRows() . ' rows' );
foreach ( $res as $row ) {
$this->moveTextRow( $row->bt_text_id, $row->bt_new_url );
+ if ( $row->bt_text_id % 10 == 0 ) {
+ $this->waitForSlaves();
+ }
}
$startId = $row->bt_text_id;
}
$this->finishIncompleteMoves( array( 'bt_text_id' => $textIds ) );
$this->syncDBs();
}
-
+
$trx = new CgzCopyTransaction( $this, $this->orphanBlobClass );
$res = wfGetDB( DB_SLAVE )->select(
- array( 'text', 'blob_tracking' ),
- array( 'old_id', 'old_text', 'old_flags' ),
- array(
+ array( 'text', 'blob_tracking' ),
+ array( 'old_id', 'old_text', 'old_flags' ),
+ array(
'old_id' => $textIds,
'bt_text_id=old_id',
'bt_moved' => 0,
- ),
+ ),
__METHOD__,
array( 'DISTINCT' )
);
foreach ( $res as $row ) {
$text = Revision::getRevisionText( $row );
if ( $text === false ) {
- $this->critical( "Error: cannot load revision text for old_id=$textId" );
+ $this->critical( "Error: cannot load revision text for old_id={$row->old_id}" );
continue;
}
-
+
if ( !$trx->addItem( $text, $row->old_id ) ) {
$this->debug( "[orphan]: committing blob with " . $trx->getSize() . " rows" );
$trx->commit();
$trx = new CgzCopyTransaction( $this, $this->orphanBlobClass );
+ $this->waitForSlaves();
}
}
$this->debug( "[orphan]: committing blob with " . $trx->getSize() . " rows" );
$trx->commit();
}
+
+ /**
+ * Wait for slaves (quietly)
+ */
+ function waitForSlaves() {
+ $lb = wfGetLB();
+ while ( true ) {
+ list( $host, $maxLag ) = $lb->getMaxLag();
+ if ( $maxLag < 2 ) {
+ break;
+ }
+ sleep( 5 );
+ }
+ }
}
/**
// Check to see if the target text_ids have been moved already.
//
- // We originally read from the slave, so this can happen when a single
- // text_id is shared between multiple pages. It's rare, but possible
+ // We originally read from the slave, so this can happen when a single
+ // text_id is shared between multiple pages. It's rare, but possible
// if a delete/move/undelete cycle splits up a null edit.
//
// We do a locking read to prevent closer-run race conditions.
$dbw = wfGetDB( DB_MASTER );
$dbw->begin();
- $res = $dbw->select( 'blob_tracking',
+ $res = $dbw->select( 'blob_tracking',
array( 'bt_text_id', 'bt_moved' ),
array( 'bt_text_id' => array_keys( $this->referrers ) ),
__METHOD__, array( 'FOR UPDATE' ) );
scripts.mit.edu / autoinstallsdev / mediawiki.git / blobdiff