Dotaz: vlakna v PERLu

[toxLinuch]

Mam script v PERLu, ktery uklada v ruznych intervalech data do DB, ale zpozdeni je vzdy stejne. Problem je, ze script boptna v pameti a to takovym zpusobem, ze je schopen polozit i stroj. Script ceka na vstup z pipe a pote reaguje. syntaxe script1.pl | script2.pl

script2.pl
sub update {
  my ($item) = @_;
  print "item to DB: $item\n";
  sleep(2);
  if (my $dbh = DBI->connect('DBI:mysql:dataWeb', 'xxx', 'xxx')) {
    my $query="UPDATE onWeb SET name='".$item."' WHERE id=0;";
    my $sth = $dbh->prepare($query);#zkousel jsem pouzit i $dbh->do ale je to stejny
    my $rv  = $sth->execute();
    $dbh->disconnect;
  }
  else {
    print "Could not connect to database: $DBI::errstr";
  }
}


my $scThread=0;
while (my $line = <STDIN>) {
  if ($line ne "") {
    chomp ($line);
    $scThread = threads->create(\&update, $line);
  }
}

Nevite nekdo prosim vas coje spatne? Kde mam udelat zmenu, aby script bezel jak ma? Soft mysql 5.0.51a a perl 5.10.

Odpovědi

[Ivan]

Asi je to offtopic, ale opravdu je nutne kvuli kazdemu update otevirat zvlastni konexi? A jsi si jisty, ze libmysql je thread-safe?

[toxLinuch]

zkousel jsem pedat hodnotu spojeni jako parametr do vlakna bohuzel perl vyhodil error a nejelo to. postupoval sem takto :

script2.pl
sub update {
  my ($item,$dbh) = @_;
  print "item to DB: $item\n";
  sleep(2);
  my $query="UPDATE onWeb SET name='".$item."' WHERE id=0;";
  my $sth = $dbh->prepare($query);#zkousel jsem pouzit i $dbh->do ale je to stejny
  my $rv  = $sth->execute();
}


my $scThread=0;
my $dbh = DBI->connect('DBI:mysql:dataWeb', 'xxx', 'xxx');
while (my $line = <STDIN>) {
  if ($line ne "") {
    chomp ($line);
    $scThread = threads->create(\&update, $line, $dbh);
  }
}
$dbh->disconnect;

Hlaseni perlu:

Thread 2 tady dochazi ke zvysovani cisla vlakna terminated abnormally: DBD::mysql::db prepare failed: handle 2 is owned by thread 198e010 not current thread 1e2b7e0 (handles can't be shared between threads and your driver may need a CLONE method added)

Pridal jsem teda shared:

pridano
use threads::shared;
...
my $scThread=0;
my $dbh = DBI->connect('DBI:mysql:dataWeb', 'xxx', 'xxx');
share ($dbh);
while (my $line = <STDIN>) {
...

perl hlasi
Thread 1 terminated abnormally: dbih_getcom handle threads::shared::tie=SCALAR(0xe739c0) is not a DBI handle at script2.pl <STDIN> line 1
Thread 2 terminated abnormally: dbih_getcom handle threads::shared::tie=SCALAR(0xe739c0) is not a DBI handle at script2.pl <STDIN> line 2

[toxLinuch]

tak jsem ted vygooglil na cpanu:

Threads and Thread Safety
Perl 5.7 and later support a new threading model called iThreads. (The old "5.005 style" threads are not supported by the DBI.)

In the iThreads model each thread has it's own copy of the perl interpreter. When a new thread is created the original perl interpreter is 'cloned' to create a new copy for the new thread.

If the DBI and drivers are loaded and handles created before the thread is created then it will get a cloned copy of the DBI, the drivers and the handles.

However, the internal pointer data within the handles will refer to the DBI and drivers in the original interpreter. Using those handles in the new interpreter thread is not safe, so the DBI detects this and croaks on any method call using handles that don't belong to the current thread (except for DESTROY).

Because of this (possibly temporary) restriction, newly created threads must make their own connections to the database. Handles can't be shared across threads.

But BEWARE, some underlying database APIs (the code the DBD driver uses to talk to the database, often supplied by the database vendor) are not thread safe. If it's not thread safe, then allowing more than one thread to enter the code at the same time may cause subtle/serious problems. In some cases allowing more than one thread to enter the code, even if not at the same time, can cause problems. You have been warned.

Jestli to chapu dobre ma se delat v kazdem vlakne zvlast pripojeni. Pak tedy nechapu proc mi to zere pamet. Vzdyt do nekolika hodin mi padne cely system, kudla drat.

[VM]

Nikde nevidim $scThread->join() . Rekl bych ze vytvaris zombie thready.

[toxLinuch]

Nemuzu cekat na ukonceni children vlakna tim bych nezpracoval dalsi vstup. Potrebuju ,aby se operace ve vlaknu zpracovaly a pote se vlakno ukoncilo. Mezitim se vsak zpracovava dalsi vstup. Kdyz se divam na vypis vlaken k danemu procesu tak se vlakna ukoncuji. Pro zobrazeni vlaken k danemu procesu pouzivam ps -p 14814 uH

[toxLinuch]

Vyreseno-naprogramovano v ansiC. I kdyz mi to prijde jako jit s tankem na blechu, ale co uz se da delat. Vyhoda cecka je ze se da perfektne ovladat pamet apod. No co vam budu povidat. Api bezi a ma stale stejne pozadavky jako po spusteni. Takze super. Cecku zdar-Perl sux(teda na takove aplikace jinak dobry )

[NeoV]

hmmm take hinty:

1. naco su Vam vlakna pre jednoduche updaty jedneho zaznamu ? Ved samotne neustale vytvaranie spojenia kazdym novym vlaknom je x-nasobne narocnejsie ako jednovlaknova app. ktora si drzi otvorene spojenie (v pripade jeho ukoncenia ho len znovu-otvori) a po kazdom nacitani noveho riadku z pajpy hned vykona update do DB jedneho zaznamu, co je operacia porovnatelna s vyvolanim noveho vlakna...

2. pri pouzivani viacerych vlakien, ktore si kazde vytvara vlastne nove spojenie na DB sa Vam moze v jednom z miliona pripadov stat, ze prve vlakno bude vytvarat spojenie s DB dlhsie ako druhe (neskor vytvorene vlakno) a tym padom druhe vlakno vykona update ako prve...

3. ak teoreticky pride v rychlom slede velke mnozstvo 'riadkov' do pajpy, tak bud sa zdrbe DB alebo sa ti nevytvoria spojenia na DB nakolko dosiahnes max. pocet otvorenych pripojeni na DB

[toxLinuch]

naznacil jsem cast ulohy ne vsak celou . V cecku je jedno spojeni, ktere se predava zadne dementni otvirani porad dokola. Vyborna obsluha pameti- jeji uvolneni, apod.

[petr_p]

Co máš za distribuci? Perl 5.10 zapomíná dealokovat jednu proměnnou při ukončování vlákna. Pokud to distributor neopravil sám, tak by měl být postižen i tento kód:

use threads;

sub ThreadRoutine {}

while (1) {
    threads->new(\&ThreadRoutine)->join;
}

Oprava.

[toxLinuch]

distro je debian lenny. nic takoveho jsem nikde nezjistil. Diky za info. Ted uz to resit nebudu, ale je to dobre vedet,

[ph0enix]

Pokud bych mel pouzit vlakna, pouzil bych jen dve a to nasledovne - hlavni vlakno nacita data ze vstupu a pres sdilene pole je predava druhemu vlaknu, jenz nasledne aktualizuje databazi. Takto bude zajisteno zpracovani dat v poradi jak dosly. Netusim ale jak to bude rychle / jaka bude rezie.

use threads;
use threads::shared;

my $line;
my $thr_end;
my @thr_data;
my $thr_worker;

share($thr_end);
share(@thr_data);

$thr_end    = 0;
$thr_worker = threads->create(\&thr_sub);
@thr_data   = ();

sub thr_sub {
    my $dbh = DBI->connect('DBI:mysql:dataWeb;mysql_server_prepare=1', 'user', 'password');
    my $qry = "UPDATE onWeb SET name=? WHERE id=0";
    my $sth;
    my $tmp;

    die "Can't connect to database" if (!$dbh);
    $sth = $dbh->prepate($qry);

    while (!$thr_end) {
        {
            lock(@thr_data);
            $tmp = shift(@thr_data);
        }       
        $sth->execute($tmp) if(defined($tmp));
    }       
    lock(@thr_data);
    $sth->execute($_) for (@thr_data);
    $dbh->disconnect();
}

while ($line = <STDIN>) {
    chomp($line);
    next if ($line =~ /^\s*$/);
    {       
        lock(@thr_data);
        push @thr_data, $line;
    }
}

$thr_end = 1;
$thr_worker->join();