[ViewVC] Diff of: cvs/AnyEvent-Fork-Pool/Pool.pm

Comparing AnyEvent-Fork-Pool/Pool.pm (file contents):
Revision 1.1 by root, Thu Apr 18 14:24:01 2013 UTC vs.
Revision 1.8 by root, Sun Apr 21 12:28:34 2013 UTC

…		…
1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent::Fork::Pool - simple process pool manager on top of AnyEvent::Fork	3	AnyEvent::Fork::Pool - simple process pool manager on top of AnyEvent::Fork
		4
		5	THE API IS NOT FINISHED, CONSIDER THIS AN ALPHA RELEASE
4		6
5	=head1 SYNOPSIS	7	=head1 SYNOPSIS
6		8
7	use AnyEvent;	9	use AnyEvent;
8	use AnyEvent::Fork::Pool;	10	use AnyEvent::Fork::Pool;
9	# use AnyEvent::Fork is not needed	11	# use AnyEvent::Fork is not needed
10		12
11	# all parameters with default values	13	# all possible parameters shown, with default values
12	my $pool = new AnyEvent::Fork::Pool	14	my $pool = AnyEvent::Fork
13	"MyWorker::run",	15	->new
		16	->require ("MyWorker")
		17	->AnyEvent::Fork::Pool::run (
		18	"MyWorker::run", # the worker function
14		19
15	# pool management	20	# pool management
16	min => 0, # minimum # of processes
17	max => 8, # maximum # of processes	21	max => 4, # absolute maximum # of processes
		22	idle => 0, # minimum # of idle processes
		23	load => 2, # queue at most this number of jobs per process
18	busy_time => 0, # wait this before starting a new process	24	start => 0.1, # wait this many seconds before starting a new process
19	max_idle => 1, # wait this before killing an idle process	25	stop => 10, # wait this many seconds before stopping an idle process
20	idle_time => 1, # at most this many idle processes	26	on_destroy => (my $finish = AE::cv), # called when object is destroyed
21		27
22	# template process
23	template => AnyEvent::Fork->new, # the template process to use
24	require => [MyWorker::], # module(s) to load
25	eval => "# perl code to execute in template",
26	on_destroy => (my $finish = AE::cv),
27
28	# parameters passed to AnyEvent::Fork::RPC	28	# parameters passed to AnyEvent::Fork::RPC
29	async => 0,	29	async => 0,
30	on_error => sub { die "FATAL: $_[0]\n" },	30	on_error => sub { die "FATAL: $_[0]\n" },
31	on_event => sub { my @ev = @_ },	31	on_event => sub { my @ev = @_ },
32	init => "MyWorker::init",	32	init => "MyWorker::init",
33	serialiser => $AnyEvent::Fork::RPC::STRING_SERIALISER,	33	serialiser => $AnyEvent::Fork::RPC::STRING_SERIALISER,
34	;	34	);
35		35
36	for (1..10) {	36	for (1..10) {
37	$pool->call (doit => $_, sub {	37	$pool->(doit => $_, sub {
38	print "MyWorker::run returned @_\n";	38	print "MyWorker::run returned @_\n";
39	});	39	});
40	}	40	}
41		41
42	undef $pool;	42	undef $pool;
50	pool of processes that handles jobs.	50	pool of processes that handles jobs.
51		51
52	Understanding of L<AnyEvent::Fork> is helpful but not critical to be able	52	Understanding of L<AnyEvent::Fork> is helpful but not critical to be able
53	to use this module, but a thorough understanding of L<AnyEvent::Fork::RPC>	53	to use this module, but a thorough understanding of L<AnyEvent::Fork::RPC>
54	is, as it defines the actual API that needs to be implemented in the	54	is, as it defines the actual API that needs to be implemented in the
55	children.	55	worker processes.
56		56
57	=head1 EXAMPLES	57	=head1 EXAMPLES
58		58
59	=head1 API	59	=head1 PARENT USAGE
		60
		61	To create a pool, you first have to create a L<AnyEvent::Fork> object -
		62	this object becomes your template process. Whenever a new worker process
		63	is needed, it is forked from this template process. Then you need to
		64	"hand off" this template process to the C<AnyEvent::Fork::Pool> module by
		65	calling its run method on it:
		66
		67	my $template = AnyEvent::Fork
		68	->new
		69	->require ("SomeModule", "MyWorkerModule");
		70
		71	my $pool = $template->AnyEvent::Fork::Pool::run ("MyWorkerModule::myfunction");
		72
		73	The pool "object" is not a regular Perl object, but a code reference that
		74	you can call and that works roughly like calling the worker function
		75	directly, except that it returns nothing but instead you need to specify a
		76	callback to be invoked once results are in:
		77
		78	$pool->(1, 2, 3, sub { warn "myfunction(1,2,3) returned @_" });
60		79
61	=over 4	80	=over 4
62		81
63	=cut	82	=cut
64		83
65	package AnyEvent::Fork::Pool;	84	package AnyEvent::Fork::Pool;
66		85
67	use common::sense;	86	use common::sense;
68		87
		88	use Scalar::Util ();
		89
69	use Guard ();	90	use Guard ();
		91	use Array::Heap ();
70		92
71	use AnyEvent;	93	use AnyEvent;
72	use AnyEvent::Fork; # we don't actually depend on it, this is for convenience	94	use AnyEvent::Fork; # we don't actually depend on it, this is for convenience
73	use AnyEvent::Fork::RPC;	95	use AnyEvent::Fork::RPC;
74		96
		97	# these are used for the first and last argument of events
		98	# in the hope of not colliding. yes, I don't like it either,
		99	# but didn't come up with an obviously better alternative.
		100	my $magic0 = ':t6Z@HK1N%Dx@_7?=~-7NQgWDdAs6a,jFN=wLO0jD1%P';
		101	my $magic1 = '<~53rexz.U`!]X[A235^"fyEoiTF\T~oH1l/N6+Djep9b~bI9`\1x%B~vWO1q*';
		102
75	our $VERSION = 0.1;	103	our $VERSION = 0.1;
76		104
77	=item my $rpc = new AnyEvent::Fork::RPC::pool $function, [key => value...]	105	=item my $pool = AnyEvent::Fork::Pool::run $fork, $function, [key => value...]
		106
		107	The traditional way to call the pool creation function. But it is way
		108	cooler to call it in the following way:
		109
		110	=item my $pool = $fork->AnyEvent::Fork::Pool::run ($function, [key => value...])
		111
		112	Creates a new pool object with the specified C<$function> as function
		113	(name) to call for each request. The pool uses the C<$fork> object as the
		114	template when creating worker processes.
		115
		116	You can supply your own template process, or tell C<AnyEvent::Fork::Pool>
		117	to create one.
		118
		119	A relatively large number of key/value pairs can be specified to influence
		120	the behaviour. They are grouped into the categories "pool management",
		121	"template process" and "rpc parameters".
78		122
79	=over 4	123	=over 4
80		124
81	=item on_error => $cb->($msg)	125	=item Pool Management
82		126
83	Called on (fatal) errors, with a descriptive (hopefully) message. If	127	The pool consists of a certain number of worker processes. These options
84	this callback is not provided, but C<on_event> is, then the C<on_event>	128	decide how many of these processes exist and when they are started and
85	callback is called with the first argument being the string C<error>,	129	stopped.
86	followed by the error message.
87		130
88	If neither handler is provided it prints the error to STDERR and will	131	The worker pool is dynamically resized, according to (perceived :)
89	start failing badly.	132	load. The minimum size is given by the C<idle> parameter and the maximum
		133	size is given by the C<max> parameter. A new worker is started every
		134	C<start> seconds at most, and an idle worker is stopped at most every
		135	C<stop> second.
90		136
91	=item on_event => $cb->(...)	137	You can specify the amount of jobs sent to a worker concurrently using the
		138	C<load> parameter.
92		139
93	Called for every call to the C<AnyEvent::Fork::RPC::event> function in the	140	=over 4
94	child, with the arguments of that function passed to the callback.
95		141
96	Also called on errors when no C<on_error> handler is provided.	142	=item idle => $count (default: 0)
97		143
98	=item on_destroy => $cb->()	144	The minimum amount of idle processes in the pool - when there are fewer
		145	than this many idle workers, C<AnyEvent::Fork::Pool> will try to start new
		146	ones, subject to the limits set by C<max> and C<start>.
99		147
100	Called when the C<$rpc> object has been destroyed and all requests have	148	This is also the initial amount of workers in the pool. The default of
101	been successfully handled. This is useful when you queue some requests and	149	zero means that the pool starts empty and can shrink back to zero workers
102	want the child to go away after it has handled them. The problem is that	150	over time.
103	the parent must not exit either until all requests have been handled, and
104	this can be accomplished by waiting for this callback.
105		151
106	=item init => $function (default none)	152	=item max => $count (default: 4)
107		153
108	When specified (by name), this function is called in the child as the very	154	The maximum number of processes in the pool, in addition to the template
109	first thing when taking over the process, with all the arguments normally	155	process. C<AnyEvent::Fork::Pool> will never have more than this number of
110	passed to the C<AnyEvent::Fork::run> function, except the communications	156	worker processes, although there can be more temporarily when a worker is
111	socket.	157	shut down and hasn't exited yet.
112		158
113	It can be used to do one-time things in the child such as storing passed	159	=item load => $count (default: 2)
114	parameters or opening database connections.
115		160
116	It is called very early - before the serialisers are created or the	161	The maximum number of concurrent jobs sent to a single worker process.
117	C<$function> name is resolved into a function reference, so it could be	162
118	used to load any modules that provide the serialiser or function. It can	163	Jobs that cannot be sent to a worker immediately (because all workers are
119	not, however, create events.	164	busy) will be queued until a worker is available.
		165
		166	Setting this low improves latency. For example, at C<1>, every job that
		167	is sent to a worker is sent to a completely idle worker that doesn't run
		168	any other jobs. The downside is that throughput is reduced - a worker that
		169	finishes a job needs to wait for a new job from the parent.
		170
		171	The default of C<2> is usually a good compromise.
		172
		173	=item start => $seconds (default: 0.1)
		174
		175	When there are fewer than C<idle> workers (or all workers are completely
		176	busy), then a timer is started. If the timer elapses and there are still
		177	jobs that cannot be queued to a worker, a new worker is started.
		178
		179	This sets the minimum time that all workers must be busy before a new
		180	worker is started. Or, put differently, the minimum delay between starting
		181	new workers.
		182
		183	The delay is small by default, which means new workers will be started
		184	relatively quickly. A delay of C<0> is possible, and ensures that the pool
		185	will grow as quickly as possible under load.
		186
		187	Non-zero values are useful to avoid "exploding" a pool because a lot of
		188	jobs are queued in an instant.
		189
		190	Higher values are often useful to improve efficiency at the cost of
		191	latency - when fewer processes can do the job over time, starting more and
		192	more is not necessarily going to help.
		193
		194	=item stop => $seconds (default: 10)
		195
		196	When a worker has no jobs to execute it becomes idle. An idle worker that
		197	hasn't executed a job within this amount of time will be stopped, unless
		198	the other parameters say otherwise.
		199
		200	Setting this to a very high value means that workers stay around longer,
		201	even when they have nothing to do, which can be good as they don't have to
		202	be started on the netx load spike again.
		203
		204	Setting this to a lower value can be useful to avoid memory or simply
		205	process table wastage.
		206
		207	Usually, setting this to a time longer than the time between load spikes
		208	is best - if you expect a lot of requests every minute and little work
		209	in between, setting this to longer than a minute avoids having to stop
		210	and start workers. On the other hand, you have to ask yourself if letting
		211	workers run idle is a good use of your resources. Try to find a good
		212	balance between resource usage of your workers and the time to start new
		213	workers - the processes created by L<AnyEvent::Fork> itself is fats at
		214	creating workers while not using much memory for them, so most of the
		215	overhead is likely from your own code.
		216
		217	=item on_destroy => $callback->() (default: none)
		218
		219	When a pool object goes out of scope, the outstanding requests are still
		220	handled till completion. Only after handling all jobs will the workers
		221	be destroyed (and also the template process if it isn't referenced
		222	otherwise).
		223
		224	To find out when a pool I<really> has finished its work, you can set this
		225	callback, which will be called when the pool has been destroyed.
		226
		227	=back
		228
		229	=item AnyEvent::Fork::RPC Parameters
		230
		231	These parameters are all passed more or less directly to
		232	L<AnyEvent::Fork::RPC>. They are only briefly mentioned here, for
		233	their full documentation please refer to the L<AnyEvent::Fork::RPC>
		234	documentation. Also, the default values mentioned here are only documented
		235	as a best effort - the L<AnyEvent::Fork::RPC> documentation is binding.
		236
		237	=over 4
120		238
121	=item async => $boolean (default: 0)	239	=item async => $boolean (default: 0)
122		240
123	The default server used in the child does all I/O blockingly, and only	241	Whether to use the synchronous or asynchronous RPC backend.
124	allows a single RPC call to execute concurrently.
125		242
126	Setting C<async> to a true value switches to another implementation that	243	=item on_error => $callback->($message) (default: die with message)
127	uses L<AnyEvent> in the child and allows multiple concurrent RPC calls.
128		244
129	The actual API in the child is documented in the section that describes	245	The callback to call on any (fatal) errors.
130	the calling semantics of the returned C<$rpc> function.
131		246
132	If you want to pre-load the actual back-end modules to enable memory	247	=item on_event => $callback->(...) (default: C<sub { }>, unlike L<AnyEvent::Fork::RPC>)
133	sharing, then you should load C<AnyEvent::Fork::RPC::Sync> for
134	synchronous, and C<AnyEvent::Fork::RPC::Async> for asynchronous mode.
135		248
136	If you use a template process and want to fork both sync and async	249	The callback to invoke on events.
137	children, then it is permissible to load both modules.
138		250
139	=item serialiser => $string (default: '(sub { pack "(w/a)", @_ }, sub { unpack "(w/a)", shift })')	251	=item init => $initfunction (default: none)
140		252
141	All arguments, result data and event data have to be serialised to be	253	The function to call in the child, once before handling requests.
142	transferred between the processes. For this, they have to be frozen and
143	thawed in both parent and child processes.
144		254
145	By default, only octet strings can be passed between the processes, which	255	=item serialiser => $serialiser (defailt: $AnyEvent::Fork::RPC::STRING_SERIALISER)
146	is reasonably fast and efficient.
147		256
148	For more complicated use cases, you can provide your own freeze and thaw	257	The serialiser to use.
149	functions, by specifying a string with perl source code. It's supposed to
150	return two code references when evaluated: the first receives a list of
151	perl values and must return an octet string. The second receives the octet
152	string and must return the original list of values.
153
154	If you need an external module for serialisation, then you can either
155	pre-load it into your L<AnyEvent::Fork> process, or you can add a C<use>
156	or C<require> statement into the serialiser string. Or both.
157		258
158	=back	259	=back
159		260
160	See the examples section earlier in this document for some actual	261	=back
161	examples.
162		262
163	=cut	263	=cut
164		264
165	sub new {	265	sub run {
166	my ($self, $function, %arg) = @_;	266	my ($template, $function, %arg) = @_;
167		267
168	my $serialiser = delete $arg{serialiser} \|\| $STRING_SERIALISER;	268	my $max = $arg{max} \|\| 4;
		269	my $idle = $arg{idle} \|\| 0,
		270	my $load = $arg{load} \|\| 2,
		271	my $start = $arg{start} \|\| 0.1,
		272	my $stop = $arg{stop} \|\| 10,
169	my $on_event = delete $arg{on_event};	273	my $on_event = $arg{on_event} \|\| sub { },
170	my $on_error = delete $arg{on_error};
171	my $on_destroy = delete $arg{on_destroy};	274	my $on_destroy = $arg{on_destroy};
		275
		276	my @rpc = (
		277	async => $arg{async},
		278	init => $arg{init},
		279	serialiser => delete $arg{serialiser},
		280	on_error => $arg{on_error},
172		281	);
173	# default for on_error is to on_event, if specified
174	$on_error \|\|= $on_event
175	? sub { $on_event->(error => shift) }
176	: sub { die "AnyEvent::Fork::RPC: uncaught error: $_[0].\n" };
177		282
178	# default for on_event is to raise an error	283	my (@pool, @queue, $nidle, $start_w, $stop_w, $shutdown);
179	$on_event \|\|= sub { $on_error->("event received, but no on_event handler") };	284	my ($start_worker, $stop_worker, $want_start, $want_stop, $scheduler);
180		285
181	my ($f, $t) = eval $serialiser; die $@ if $@;	286	my $destroy_guard = Guard::guard {
		287	$on_destroy->()
		288	if $on_destroy;
		289	};
182		290
183	my (@rcb, %rcb, $fh, $shutdown, $wbuf, $ww);	291	$template
184	my ($rlen, $rbuf, $rw) = 512 - 16;	292	->require ("AnyEvent::Fork::RPC::" . ($arg{async} ? "Async" : "Sync"))
185		293	->eval ('
186	my $wcb = sub {	294	my ($magic0, $magic1) = @_;
187	my $len = syswrite $fh, $wbuf;	295	sub AnyEvent::Fork::Pool::retire() {
188		296	AnyEvent::Fork::RPC::event $magic0, "quit", $magic1;
189	unless (defined $len) {
190	if ($! != Errno::EAGAIN && $! != Errno::EWOULDBLOCK) {
191	undef $rw; undef $ww; # it ends here
192	$on_error->("$!");
193	}	297	}
194	}	298	', $magic0, $magic1)
195
196	substr $wbuf, 0, $len, "";
197
198	unless (length $wbuf) {
199	undef $ww;
200	$shutdown and shutdown $fh, 1;
201	}
202	};	299	;
203		300
204	my $module = "AnyEvent::Fork::RPC::" . ($arg{async} ? "Async" : "Sync");	301	$start_worker = sub {
		302	my $proc = [0, 0, undef]; # load, index, rpc
205		303
206	$self->require ($module)	304	$proc->[2] = $template
207	->send_arg ($function, $arg{init}, $serialiser)	305	->fork
208	->run ("$module\::run", sub {	306	->AnyEvent::Fork::RPC::run ($function,
209	$fh = shift;	307	@rpc,
		308	on_event => sub {
		309	if (@_ == 3 && $_[0] eq $magic0 && $_[2] eq $magic1) {
		310	$destroy_guard if 0; # keep it alive
210		311
211	my ($id, $len);	312	$_[1] eq "quit" and $stop_worker->($proc);
212	$rw = AE::io $fh, 0, sub {
213	$rlen = $rlen * 2 + 16 if $rlen - 128 < length $rbuf;
214	$len = sysread $fh, $rbuf, $rlen - length $rbuf, length $rbuf;
215
216	if ($len) {
217	while (8 <= length $rbuf) {
218	($id, $len) = unpack "LL", $rbuf;
219	8 + $len <= length $rbuf
220	or last;	313	return;
221
222	my @r = $t->(substr $rbuf, 8, $len);
223	substr $rbuf, 0, 8 + $len, "";
224
225	if ($id) {
226	if (@rcb) {
227	(shift @rcb)->(@r);
228	} elsif (my $cb = delete $rcb{$id}) {
229	$cb->(@r);
230	} else {
231	undef $rw; undef $ww;
232	$on_error->("unexpected data from child");
233	}	314	}
234	} else {	315
235	$on_event->(@r);	316	&$on_event;
236	}	317	},
237	}	318	)
238	} elsif (defined $len) {	319	;
239	undef $rw; undef $ww; # it ends here
240		320
241	if (@rcb \|\| %rcb) {	321	++$nidle;
242	$on_error->("unexpected eof");	322	Array::Heap::push_heap_idx @pool, $proc;
		323
		324	Scalar::Util::weaken $proc;
		325	};
		326
		327	$stop_worker = sub {
		328	my $proc = shift;
		329
		330	$proc->[0]
		331	or --$nidle;
		332
		333	Array::Heap::splice_heap_idx @pool, $proc->[1]
		334	if defined $proc->[1];
		335
		336	@$proc = 0; # tell others to leave it be
		337	};
		338
		339	$want_start = sub {
		340	undef $stop_w;
		341
		342	$start_w \|\|= AE::timer $start, $start, sub {
		343	if (($nidle < $idle \|\| @queue) && @pool < $max) {
		344	$start_worker->();
		345	$scheduler->();
243	} else {	346	} else {
244	$on_destroy->();	347	undef $start_w;
245	}
246	} elsif ($! != Errno::EAGAIN && $! != Errno::EWOULDBLOCK) {
247	undef $rw; undef $ww; # it ends here
248	$on_error->("read: $!");
249	}	348	}
250	};	349	};
251
252	$ww \|\|= AE::io $fh, 1, $wcb;
253	});	350	};
254		351
		352	$want_stop = sub {
		353	$stop_w \|\|= AE::timer $stop, $stop, sub {
		354	$stop_worker->($pool[0])
		355	if $nidle;
		356
		357	undef $stop_w
		358	if $nidle <= $idle;
		359	};
		360	};
		361
		362	$scheduler = sub {
		363	if (@queue) {
		364	while (@queue) {
		365	@pool or $start_worker->();
		366
		367	my $proc = $pool[0];
		368
		369	if ($proc->[0] < $load) {
		370	# found free worker, increase load
		371	unless ($proc->[0]++) {
		372	# worker became busy
		373	--$nidle
		374	or undef $stop_w;
		375
		376	$want_start->()
		377	if $nidle < $idle && @pool < $max;
		378	}
		379
		380	Array::Heap::adjust_heap_idx @pool, 0;
		381
		382	my $job = shift @queue;
		383	my $ocb = pop @$job;
		384
		385	$proc->[2]->(@$job, sub {
		386	# reduce load
		387	--$proc->[0] # worker still busy?
		388	or ++$nidle > $idle # not too many idle processes?
		389	or $want_stop->();
		390
		391	Array::Heap::adjust_heap_idx @pool, $proc->[1]
		392	if defined $proc->[1];
		393
		394	&$ocb;
		395
		396	$scheduler->();
		397	});
		398	} else {
		399	$want_start->()
		400	unless @pool >= $max;
		401
		402	last;
		403	}
		404	}
		405	} elsif ($shutdown) {
		406	@pool = ();
		407	undef $start_w;
		408	undef $start_worker; # frees $destroy_guard reference
		409
		410	$stop_worker->($pool[0])
		411	while $nidle;
		412	}
		413	};
		414
255	my $guard = Guard::guard {	415	my $shutdown_guard = Guard::guard {
256	$shutdown = 1;	416	$shutdown = 1;
257	$ww \|\|= $fh && AE::io $fh, 1, $wcb;	417	$scheduler->();
		418	};
		419
		420	$start_worker->()
		421	while @pool < $idle;
		422
		423	sub {
		424	$shutdown_guard if 0; # keep it alive
		425
		426	$start_worker->()
		427	unless @pool;
		428
		429	push @queue, [@_];
		430	$scheduler->();
258	};	431	}
259
260	my $id;
261
262	$arg{async}
263	? sub {
264	$id = ($id == 0xffffffff ? 0 : $id) + 1;
265	$id = ($id == 0xffffffff ? 0 : $id) + 1 while exists $rcb{$id}; # rarely loops
266
267	$rcb{$id} = pop;
268
269	$guard; # keep it alive
270
271	$wbuf .= pack "LL/a*", $id, &$f;
272	$ww \|\|= $fh && AE::io $fh, 1, $wcb;
273	}
274	: sub {
275	push @rcb, pop;
276
277	$guard; # keep it alive
278
279	$wbuf .= pack "L/a*", &$f;
280	$ww \|\|= $fh && AE::io $fh, 1, $wcb;
281	}
282	}	432	}
283		433
284	=item $pool->call (..., $cb->(...))	434	=item $pool->(..., $cb->(...))
		435
		436	Call the RPC function of a worker with the given arguments, and when the
		437	worker is done, call the C<$cb> with the results, just like calling the
		438	RPC object durectly - see the L<AnyEvent::Fork::RPC> documentation for
		439	details on the RPC API.
		440
		441	If there is no free worker, the call will be queued until a worker becomes
		442	available.
		443
		444	Note that there can be considerable time between calling this method and
		445	the call actually being executed. During this time, the parameters passed
		446	to this function are effectively read-only - modifying them after the call
		447	and before the callback is invoked causes undefined behaviour.
		448
		449	=cut
285		450
286	=back	451	=back
		452
		453	=head1 CHILD USAGE
		454
		455	In addition to the L<AnyEvent::Fork::RPC> API, this module implements one
		456	more child-side function:
		457
		458	=over 4
		459
		460	=item AnyEvent::Fork::Pool::retire ()
		461
		462	This function sends an event to the parent process to request retirement:
		463	the worker is removed from the pool and no new jobs will be sent to it,
		464	but it has to handle the jobs that are already queued.
		465
		466	The parentheses are part of the syntax: the function usually isn't defined
		467	when you compile your code (because that happens I<before> handing the
		468	template process over to C<AnyEvent::Fork::Pool::run>, so you need the
		469	empty parentheses to tell Perl that the function is indeed a function.
		470
		471	Retiring a worker can be useful to gracefully shut it down when the worker
		472	deems this useful. For example, after executing a job, one could check
		473	the process size or the number of jobs handled so far, and if either is
		474	too high, the worker could ask to get retired, to avoid memory leaks to
		475	accumulate.
		476
		477	=back
		478
		479	=head1 POOL PARAMETERS RECIPES
		480
		481	This section describes some recipes for pool paramaters. These are mostly
		482	meant for the synchronous RPC backend, as the asynchronous RPC backend
		483	changes the rules considerably, making workers themselves responsible for
		484	their scheduling.
		485
		486	=over 4
		487
		488	=item low latency - set load = 1
		489
		490	If you need a deterministic low latency, you should set the C<load>
		491	parameter to C<1>. This ensures that never more than one job is sent to
		492	each worker. This avoids having to wait for a previous job to finish.
		493
		494	This makes most sense with the synchronous (default) backend, as the
		495	asynchronous backend can handle multiple requests concurrently.
		496
		497	=item lowest latency - set load = 1 and idle = max
		498
		499	To achieve the lowest latency, you additionally should disable any dynamic
		500	resizing of the pool by setting C<idle> to the same value as C<max>.
		501
		502	=item high throughput, cpu bound jobs - set load >= 2, max = #cpus
		503
		504	To get high throughput with cpu-bound jobs, you should set the maximum
		505	pool size to the number of cpus in your system, and C<load> to at least
		506	C<2>, to make sure there can be another job waiting for the worker when it
		507	has finished one.
		508
		509	The value of C<2> for C<load> is the minimum value that I<can> achieve
		510	100% throughput, but if your parent process itself is sometimes busy, you
		511	might need higher values. Also there is a limit on the amount of data that
		512	can be "in flight" to the worker, so if you send big blobs of data to your
		513	worker, C<load> might have much less of an effect.
		514
		515	=item high throughput, I/O bound jobs - set load >= 2, max = 1, or very high
		516
		517	When your jobs are I/O bound, using more workers usually boils down to
		518	higher throughput, depending very much on your actual workload - sometimes
		519	having only one worker is best, for example, when you read or write big
		520	files at maixmum speed, as a second worker will increase seek times.
		521
		522	=back
		523
		524	=head1 EXCEPTIONS
		525
		526	The same "policy" as with L<AnyEvent::Fork::RPC> applies - exceptins will
		527	not be caught, and exceptions in both worker and in callbacks causes
		528	undesirable or undefined behaviour.
287		529
288	=head1 SEE ALSO	530	=head1 SEE ALSO
289		531
290	L<AnyEvent::Fork>, to create the processes in the first place.	532	L<AnyEvent::Fork>, to create the processes in the first place.
291		533

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing AnyEvent-Fork-Pool/Pool.pm (file contents): Revision 1.1 by root, Thu Apr 18 14:24:01 2013 UTC vs. Revision 1.8 by root, Sun Apr 21 12:28:34 2013 UTC

Diff Legend

Comparing AnyEvent-Fork-Pool/Pool.pm (file contents):
Revision 1.1 by root, Thu Apr 18 14:24:01 2013 UTC vs.
Revision 1.8 by root, Sun Apr 21 12:28:34 2013 UTC