I've been doing more and more threading recently - simple, but implicitly parallel tasks, like bulk statistics collection and graphing on servers.

Whilst I initially did a 'one thread per server' type of approach, I'm finding that doesn't scale as well as I'd like - a few hundred is not a problem, but a few thousand and it just doesn't work.

Thankfully, Thread::Queue and a set of 'worker' threads works quite nicely. Queue your 'server list', spawn N threads, and have them process the queue. Thus I tend to use something like:

sub worker_thread
{
  while ( my $item = $processing_q -> dequeue() )
  {
     # do stuff to $item
  }
}

for ( my $count = 0; $count < $nthreads; $count++ )
{
  my $thr = threads -> create ( \&worker_thread );
}

$processing_q -> enqueue ( @server_list ); 
$processing_q -> end;

foreach my $thread ( threads -> list() )
{
  $thread -> join();
}
[download]

This works quite nicely, but recently I've found I've needed to do two different 'sets' of worker threads - one to 'go and collect', and another set to process (and maybe some 'helper' threads, such as one to serialise data/logging output). I've been using queues again, and having the 'workers' feed the queue. However this is where I'm starting to have a bit more difficulty.

Does anyone have a neat idiom for 'handling' sub threads?

Currently, I've tried a variety of approaches:

• A semaphore, to count when the 'worker' threads have finished their queue, and send a 'close' to the 'sub thread' queue. (setting the 'semaphore' to a negative number, and have each of the workers perform an $semaphore -> up() when it finishes.)

  my $thread_exclusion = Thread::Semaphore -> new ( -1 * $nthreads + 1
+ );
[download]

In the 'worker' threads:

  $thread_exclusion -> up();
[download]

And a 'helper' thread that all it does is close queues.

sub close_queues
{
  $exclusion_lock -> down();
  $process_q -> end(); 
}
[download]

This will block until the exclusion lock goes positive, e.g. all the 'worker' threads have finished.

• A 'timeout' mechanism' which checks activity, and if it's been idle for 'too long', close queues (and thus terminate threads).

$SIG{'ALRM'} = \&close_queues();
alarm ( 60 );
[download]

This has the advantage of being able to 'reset' the alarm - only one 'alarm' can be pending, so you can perform an 'alarm(60)' in each loop iteration to reset the clock.

• And something a bit more complicate that involves monitoring queues via 'pending()' and 'peek()' that sits in a loop and polls. (and for bonus points reaps threads as the exit).

I was wondering though - does anyone have a more elegant solution? I like how neat the 'while' loop is for queue processing.

---

---

Posts are HTML formatted. Put <p> </p> tags around your paragraphs. Put <code> </code> tags around your code and data!

Titles consisting of a single word are discouraged, and in most cases are disallowed outright.

Read Where should I post X? if you're not absolutely sure you're posting in the right place.

Please read these before you post! —

• How do I compose an effective node title?
• How do I post a question effectively?
• Markup in the Monastery

Posts may use any of the Perl Monks Approved HTML tags:

a, abbr, b, big, blockquote, br, caption, center, col, colgroup, dd, del, details, div, dl, dt, em, font, h1, h2, h3, h4, h5, h6, hr, i, ins, li, ol, p, pre, readmore, small, span, spoiler, strike, strong, sub, summary, sup, table, tbody, td, tfoot, th, thead, tr, tt, u, ul, wbr

You may need to use entities for some characters, as follows. (Exception: Within code tags, you can put the characters literally.)

	For:		Use:
	&		&
	<		&lt;
	>		&gt;
	[		&#91;
	]		&#93;

Link using PerlMonks shortcuts! What shortcuts can I use for linking?

See Writeup Formatting Tips and other pages linked from there for more info.