… | |
… | |
39 | Loading this module also always loads L<AnyEvent::Fork>, so you can make a |
39 | Loading this module also always loads L<AnyEvent::Fork>, so you can make a |
40 | separate C<use AnyEvent::Fork> if you wish, but you don't have to. |
40 | separate C<use AnyEvent::Fork> if you wish, but you don't have to. |
41 | |
41 | |
42 | =head1 EXAMPLES |
42 | =head1 EXAMPLES |
43 | |
43 | |
44 | =head2 Synchronous Backend |
44 | =head2 Example 1: Synchronous Backend |
45 | |
45 | |
46 | Here is a simple example that implements a backend that executes C<unlink> |
46 | Here is a simple example that implements a backend that executes C<unlink> |
47 | and C<rmdir> calls, and reports their status back. It also reports the |
47 | and C<rmdir> calls, and reports their status back. It also reports the |
48 | number of requests it has processed every three requests, which is clearly |
48 | number of requests it has processed every three requests, which is clearly |
49 | silly, but illustrates the use of events. |
49 | silly, but illustrates the use of events. |
50 | |
50 | |
51 | First the parent process: |
51 | First the parent process: |
52 | |
52 | |
53 | use AnyEvent; |
53 | use AnyEvent; |
54 | use AnyEvent::Fork; |
|
|
55 | use AnyEvent::Fork::RPC; |
54 | use AnyEvent::Fork::RPC; |
56 | |
55 | |
57 | my $done = AE::cv; |
56 | my $done = AE::cv; |
58 | |
57 | |
59 | my $rpc = AnyEvent::Fork |
58 | my $rpc = AnyEvent::Fork |
… | |
… | |
137 | |
136 | |
138 | And as a final remark, there is a fine module on CPAN that can |
137 | And as a final remark, there is a fine module on CPAN that can |
139 | asynchronously C<rmdir> and C<unlink> and a lot more, and more efficiently |
138 | asynchronously C<rmdir> and C<unlink> and a lot more, and more efficiently |
140 | than this example, namely L<IO::AIO>. |
139 | than this example, namely L<IO::AIO>. |
141 | |
140 | |
|
|
141 | =head3 Example 1a: the same with the asynchronous backend |
|
|
142 | |
|
|
143 | This example only shows what needs to be changed to use the async backend |
|
|
144 | instead. Doing this is not very useful, the purpose of this example is |
|
|
145 | to show the minimum amount of change that is required to go from the |
|
|
146 | synchronous to the asynchronous backend. |
|
|
147 | |
|
|
148 | To use the async backend in the previous example, you need to add the |
|
|
149 | C<async> parameter to the C<AnyEvent::Fork::RPC::run> call: |
|
|
150 | |
|
|
151 | ->AnyEvent::Fork::RPC::run ("MyWorker::run", |
|
|
152 | async => 1, |
|
|
153 | ... |
|
|
154 | |
|
|
155 | And since the function call protocol is now changed, you need to adopt |
|
|
156 | C<MyWorker::run> to the async API. |
|
|
157 | |
|
|
158 | First, you need to accept the extra initial C<$done> callback: |
|
|
159 | |
|
|
160 | sub run { |
|
|
161 | my ($done, $cmd, $path) = @_; |
|
|
162 | |
|
|
163 | And since a response is now generated when C<$done> is called, as opposed |
|
|
164 | to when the function returns, we need to call the C<$done> function with |
|
|
165 | the status: |
|
|
166 | |
|
|
167 | $done->($status or (0, "$!")); |
|
|
168 | |
|
|
169 | A few remarks are in order. First, it's quite pointless to use the async |
|
|
170 | backend for this example - but it I<is> possible. Second, you can call |
|
|
171 | C<$done> before or after returning from the function. Third, having both |
|
|
172 | returned from the function and having called the C<$done> callback, the |
|
|
173 | child process may exit at any time, so you should call C<$done> only when |
|
|
174 | you really I<are> done. |
|
|
175 | |
|
|
176 | =head2 Example 2: Asynchronous Backend |
|
|
177 | |
|
|
178 | This example implements multiple count-downs in the child, using |
|
|
179 | L<AnyEvent> timers. While this is a bit silly (one could use timers in te |
|
|
180 | parent just as well), it illustrates the ability to use AnyEvent in the |
|
|
181 | child and the fact that responses can arrive in a different order then the |
|
|
182 | requests. |
|
|
183 | |
|
|
184 | It also shows how to embed the actual child code into a C<__DATA__> |
|
|
185 | section, so it doesn't need any external files at all. |
|
|
186 | |
|
|
187 | And when your parent process is often busy, and you have stricter timing |
|
|
188 | requirements, then running timers in a child process suddenly doesn't look |
|
|
189 | so silly anymore. |
|
|
190 | |
|
|
191 | Without further ado, here is the code: |
|
|
192 | |
|
|
193 | use AnyEvent; |
|
|
194 | use AnyEvent::Fork::RPC; |
|
|
195 | |
|
|
196 | my $done = AE::cv; |
|
|
197 | |
|
|
198 | my $rpc = AnyEvent::Fork |
|
|
199 | ->new |
|
|
200 | ->require ("AnyEvent::Fork::RPC::Async") |
|
|
201 | ->eval (do { local $/; <DATA> }) |
|
|
202 | ->AnyEvent::Fork::RPC::run ("run", |
|
|
203 | async => 1, |
|
|
204 | on_error => sub { warn "FATAL: $_[0]"; exit 1 }, |
|
|
205 | on_event => sub { print $_[0] }, |
|
|
206 | on_destroy => $done, |
|
|
207 | ); |
|
|
208 | |
|
|
209 | for my $count (3, 2, 1) { |
|
|
210 | $rpc->($count, sub { |
|
|
211 | warn "job $count finished\n"; |
|
|
212 | }); |
|
|
213 | } |
|
|
214 | |
|
|
215 | undef $rpc; |
|
|
216 | |
|
|
217 | $done->recv; |
|
|
218 | |
|
|
219 | __DATA__ |
|
|
220 | |
|
|
221 | # this ends up in main, as we don't use a package declaration |
|
|
222 | |
|
|
223 | use AnyEvent; |
|
|
224 | |
|
|
225 | sub run { |
|
|
226 | my ($done, $count) = @_; |
|
|
227 | |
|
|
228 | my $n; |
|
|
229 | |
|
|
230 | AnyEvent::Fork::RPC::event "starting to count up to $count\n"; |
|
|
231 | |
|
|
232 | my $w; $w = AE::timer 1, 1, sub { |
|
|
233 | ++$n; |
|
|
234 | |
|
|
235 | AnyEvent::Fork::RPC::event "count $n of $count\n"; |
|
|
236 | |
|
|
237 | if ($n == $count) { |
|
|
238 | undef $w; |
|
|
239 | $done->(); |
|
|
240 | } |
|
|
241 | }; |
|
|
242 | } |
|
|
243 | |
|
|
244 | The parent part (the one before the C<__DATA__> section) isn't very |
|
|
245 | different from the earlier examples. It sets async mode, preloads |
|
|
246 | the backend module (so the C<AnyEvent::Fork::RPC::event> function is |
|
|
247 | declared), uses a slightly different C<on_event> handler (which we use |
|
|
248 | simply for logging purposes) and then, instead of loading a module with |
|
|
249 | the actual worker code, it C<eval>'s the code from the data section in the |
|
|
250 | child process. |
|
|
251 | |
|
|
252 | It then starts three countdowns, from 3 to 1 seconds downwards, destroys |
|
|
253 | the rpc object so the example finishes eventually, and then just waits for |
|
|
254 | the stuff to trickle in. |
|
|
255 | |
|
|
256 | The worker code uses the event function to log some progress messages, but |
|
|
257 | mostly just creates a recurring one-second timer. |
|
|
258 | |
|
|
259 | The timer callback increments a counter, logs a message, and eventually, |
|
|
260 | when the count has been reached, calls the finish callback. |
|
|
261 | |
|
|
262 | On my system, this results in the following output. Since all timers fire |
|
|
263 | at roughly the same time, the actual order isn't guaranteed, but the order |
|
|
264 | shown is very likely what you would get, too. |
|
|
265 | |
|
|
266 | starting to count up to 3 |
|
|
267 | starting to count up to 2 |
|
|
268 | starting to count up to 1 |
|
|
269 | count 1 of 3 |
|
|
270 | count 1 of 2 |
|
|
271 | count 1 of 1 |
|
|
272 | job 1 finished |
|
|
273 | count 2 of 2 |
|
|
274 | job 2 finished |
|
|
275 | count 2 of 3 |
|
|
276 | count 3 of 3 |
|
|
277 | job 3 finished |
|
|
278 | |
|
|
279 | While the overall ordering isn't guaranteed, the async backend still |
|
|
280 | guarantees that events and responses are delivered to the parent process |
|
|
281 | in the exact same ordering as they were generated in the child process. |
|
|
282 | |
|
|
283 | And unless your system is I<very> busy, it should clearly show that the |
|
|
284 | job started last will finish first, as it has the lowest count. |
|
|
285 | |
|
|
286 | This concludes the async example. Since L<AnyEvent::Fork> does not |
|
|
287 | actually fork, you are free to use about any module in the child, not just |
|
|
288 | L<AnyEvent>, but also L<IO::AIO>, or L<Tk> for example. |
|
|
289 | |
142 | =head1 PARENT PROCESS USAGE |
290 | =head1 PARENT PROCESS USAGE |
143 | |
291 | |
144 | This module exports nothing, and only implements a single function: |
292 | This module exports nothing, and only implements a single function: |
145 | |
293 | |
146 | =over 4 |
294 | =over 4 |
… | |
… | |
224 | |
372 | |
225 | The default server used in the child does all I/O blockingly, and only |
373 | The default server used in the child does all I/O blockingly, and only |
226 | allows a single RPC call to execute concurrently. |
374 | allows a single RPC call to execute concurrently. |
227 | |
375 | |
228 | Setting C<async> to a true value switches to another implementation that |
376 | Setting C<async> to a true value switches to another implementation that |
229 | uses L<AnyEvent> in the child and allows multiple concurrent RPC calls. |
377 | uses L<AnyEvent> in the child and allows multiple concurrent RPC calls (it |
|
|
378 | does not support recursion in the event loop however, blocking condvar |
|
|
379 | calls will fail). |
230 | |
380 | |
231 | The actual API in the child is documented in the section that describes |
381 | The actual API in the child is documented in the section that describes |
232 | the calling semantics of the returned C<$rpc> function. |
382 | the calling semantics of the returned C<$rpc> function. |
233 | |
383 | |
234 | If you want to pre-load the actual back-end modules to enable memory |
384 | If you want to pre-load the actual back-end modules to enable memory |
… | |
… | |
236 | synchronous, and C<AnyEvent::Fork::RPC::Async> for asynchronous mode. |
386 | synchronous, and C<AnyEvent::Fork::RPC::Async> for asynchronous mode. |
237 | |
387 | |
238 | If you use a template process and want to fork both sync and async |
388 | If you use a template process and want to fork both sync and async |
239 | children, then it is permissible to load both modules. |
389 | children, then it is permissible to load both modules. |
240 | |
390 | |
241 | =item serialiser => $string (default: '(sub { pack "(w/a*)*", @_ }, sub { unpack "(w/a*)*", shift })') |
391 | =item serialiser => $string (default: $AnyEvent::Fork::RPC::STRING_SERIALISER) |
242 | |
392 | |
243 | All arguments, result data and event data have to be serialised to be |
393 | All arguments, result data and event data have to be serialised to be |
244 | transferred between the processes. For this, they have to be frozen and |
394 | transferred between the processes. For this, they have to be frozen and |
245 | thawed in both parent and child processes. |
395 | thawed in both parent and child processes. |
246 | |
396 | |
247 | By default, only octet strings can be passed between the processes, which |
397 | By default, only octet strings can be passed between the processes, which |
248 | is reasonably fast and efficient. |
398 | is reasonably fast and efficient and requires no extra modules. |
249 | |
399 | |
250 | For more complicated use cases, you can provide your own freeze and thaw |
400 | For more complicated use cases, you can provide your own freeze and thaw |
251 | functions, by specifying a string with perl source code. It's supposed to |
401 | functions, by specifying a string with perl source code. It's supposed to |
252 | return two code references when evaluated: the first receives a list of |
402 | return two code references when evaluated: the first receives a list of |
253 | perl values and must return an octet string. The second receives the octet |
403 | perl values and must return an octet string. The second receives the octet |
… | |
… | |
255 | |
405 | |
256 | If you need an external module for serialisation, then you can either |
406 | If you need an external module for serialisation, then you can either |
257 | pre-load it into your L<AnyEvent::Fork> process, or you can add a C<use> |
407 | pre-load it into your L<AnyEvent::Fork> process, or you can add a C<use> |
258 | or C<require> statement into the serialiser string. Or both. |
408 | or C<require> statement into the serialiser string. Or both. |
259 | |
409 | |
|
|
410 | Here are some examples - some of them are also available as global |
|
|
411 | variables that make them easier to use. |
|
|
412 | |
|
|
413 | =over 4 |
|
|
414 | |
|
|
415 | =item octet strings - C<$AnyEvent::Fork::RPC::STRING_SERIALISER> |
|
|
416 | |
|
|
417 | This serialiser concatenates length-prefixes octet strings, and is the |
|
|
418 | default. |
|
|
419 | |
|
|
420 | Implementation: |
|
|
421 | |
|
|
422 | ( |
|
|
423 | sub { pack "(w/a*)*", @_ }, |
|
|
424 | sub { unpack "(w/a*)*", shift } |
|
|
425 | ) |
|
|
426 | |
|
|
427 | =item json - C<$AnyEvent::Fork::RPC::JSON_SERIALISER> |
|
|
428 | |
|
|
429 | This serialiser creates JSON arrays - you have to make sure the L<JSON> |
|
|
430 | module is installed for this serialiser to work. It can be beneficial for |
|
|
431 | sharing when you preload the L<JSON> module in a template process. |
|
|
432 | |
|
|
433 | L<JSON> (with L<JSON::XS> installed) is slower than the octet string |
|
|
434 | serialiser, but usually much faster than L<Storable>, unless big chunks of |
|
|
435 | binary data need to be transferred. |
|
|
436 | |
|
|
437 | Implementation: |
|
|
438 | |
|
|
439 | use JSON (); |
|
|
440 | ( |
|
|
441 | sub { JSON::encode_json \@_ }, |
|
|
442 | sub { @{ JSON::decode_json shift } } |
|
|
443 | ) |
|
|
444 | |
|
|
445 | =item storable - C<$AnyEvent::Fork::RPC::STORABLE_SERIALISER> |
|
|
446 | |
|
|
447 | This serialiser uses L<Storable>, which means it has high chance of |
|
|
448 | serialising just about anything you throw at it, at the cost of having |
|
|
449 | very high overhead per operation. It also comes with perl. |
|
|
450 | |
|
|
451 | Implementation: |
|
|
452 | |
|
|
453 | use Storable (); |
|
|
454 | ( |
|
|
455 | sub { Storable::freeze \@_ }, |
|
|
456 | sub { @{ Storable::thaw shift } } |
|
|
457 | ) |
|
|
458 | |
|
|
459 | =back |
|
|
460 | |
260 | =back |
461 | =back |
261 | |
462 | |
262 | See the examples section earlier in this document for some actual |
463 | See the examples section earlier in this document for some actual |
263 | examples. |
464 | examples. |
264 | |
465 | |
265 | =cut |
466 | =cut |
266 | |
467 | |
267 | our $STRING_SERIALISER = '(sub { pack "(w/a*)*", @_ }, sub { unpack "(w/a*)*", shift })'; |
468 | our $STRING_SERIALISER = '(sub { pack "(w/a*)*", @_ }, sub { unpack "(w/a*)*", shift })'; |
|
|
469 | our $JSON_SERIALISER = 'use JSON (); (sub { JSON::encode_json \@_ }, sub { @{ JSON::decode_json shift } })'; |
|
|
470 | our $STORABLE_SERIALISER = 'use Storable (); (sub { Storable::freeze \@_ }, sub { @{ Storable::thaw shift } })'; |
268 | |
471 | |
269 | sub run { |
472 | sub run { |
270 | my ($self, $function, %arg) = @_; |
473 | my ($self, $function, %arg) = @_; |
271 | |
474 | |
272 | my $serialiser = delete $arg{serialiser} || $STRING_SERIALISER; |
475 | my $serialiser = delete $arg{serialiser} || $STRING_SERIALISER; |
… | |
… | |
341 | } |
544 | } |
342 | } elsif (defined $len) { |
545 | } elsif (defined $len) { |
343 | undef $rw; undef $ww; # it ends here |
546 | undef $rw; undef $ww; # it ends here |
344 | |
547 | |
345 | if (@rcb || %rcb) { |
548 | if (@rcb || %rcb) { |
346 | use Data::Dump;ddx[\@rcb,\%rcb];#d# |
|
|
347 | $on_error->("unexpected eof"); |
549 | $on_error->("unexpected eof"); |
348 | } else { |
550 | } else { |
349 | $on_destroy->(); |
551 | $on_destroy->(); |
350 | } |
552 | } |
351 | } elsif ($! != Errno::EAGAIN && $! != Errno::EWOULDBLOCK) { |
553 | } elsif ($! != Errno::EAGAIN && $! != Errno::EWOULDBLOCK) { |
… | |
… | |
431 | See the examples section earlier in this document for some actual |
633 | See the examples section earlier in this document for some actual |
432 | examples. |
634 | examples. |
433 | |
635 | |
434 | =back |
636 | =back |
435 | |
637 | |
|
|
638 | =head1 ADVANCED TOPICS |
|
|
639 | |
|
|
640 | =head2 Choosing a backend |
|
|
641 | |
|
|
642 | So how do you decide which backend to use? Well, that's your problem to |
|
|
643 | solve, but here are some thoughts on the matter: |
|
|
644 | |
|
|
645 | =over 4 |
|
|
646 | |
|
|
647 | =item Synchronous |
|
|
648 | |
|
|
649 | The synchronous backend does not rely on any external modules (well, |
|
|
650 | except L<common::sense>, which works around a bug in how perl's warning |
|
|
651 | system works). This keeps the process very small, for example, on my |
|
|
652 | system, an empty perl interpreter uses 1492kB RSS, which becomes 2020kB |
|
|
653 | after C<use warnings; use strict> (for people who grew up with C64s around |
|
|
654 | them this is probably shocking every single time they see it). The worker |
|
|
655 | process in the first example in this document uses 1792kB. |
|
|
656 | |
|
|
657 | Since the calls are done synchronously, slow jobs will keep newer jobs |
|
|
658 | from executing. |
|
|
659 | |
|
|
660 | The synchronous backend also has no overhead due to running an event loop |
|
|
661 | - reading requests is therefore very efficient, while writing responses is |
|
|
662 | less so, as every response results in a write syscall. |
|
|
663 | |
|
|
664 | If the parent process is busy and a bit slow reading responses, the child |
|
|
665 | waits instead of processing further requests. This also limits the amount |
|
|
666 | of memory needed for buffering, as never more than one response has to be |
|
|
667 | buffered. |
|
|
668 | |
|
|
669 | The API in the child is simple - you just have to define a function that |
|
|
670 | does something and returns something. |
|
|
671 | |
|
|
672 | It's hard to use modules or code that relies on an event loop, as the |
|
|
673 | child cannot execute anything while it waits for more input. |
|
|
674 | |
|
|
675 | =item Asynchronous |
|
|
676 | |
|
|
677 | The asynchronous backend relies on L<AnyEvent>, which tries to be small, |
|
|
678 | but still comes at a price: On my system, the worker from example 1a uses |
|
|
679 | 3420kB RSS (for L<AnyEvent>, which loads L<EV>, which needs L<XSLoader> |
|
|
680 | which in turn loads a lot of other modules such as L<warnings>, L<strict>, |
|
|
681 | L<vars>, L<Exporter>...). |
|
|
682 | |
|
|
683 | It batches requests and responses reasonably efficiently, doing only as |
|
|
684 | few reads and writes as needed, but needs to poll for events via the event |
|
|
685 | loop. |
|
|
686 | |
|
|
687 | Responses are queued when the parent process is busy. This means the child |
|
|
688 | can continue to execute any queued requests. It also means that a child |
|
|
689 | might queue a lot of responses in memory when it generates them and the |
|
|
690 | parent process is slow accepting them. |
|
|
691 | |
|
|
692 | The API is not a straightforward RPC pattern - you have to call a |
|
|
693 | "done" callback to pass return values and signal completion. Also, more |
|
|
694 | importantly, the API starts jobs as fast as possible - when 1000 jobs |
|
|
695 | are queued and the jobs are slow, they will all run concurrently. The |
|
|
696 | child must implement some queueing/limiting mechanism if this causes |
|
|
697 | problems. Alternatively, the parent could limit the amount of rpc calls |
|
|
698 | that are outstanding. |
|
|
699 | |
|
|
700 | Blocking use of condvars is not supported. |
|
|
701 | |
|
|
702 | Using event-based modules such as L<IO::AIO>, L<Gtk2>, L<Tk> and so on is |
|
|
703 | easy. |
|
|
704 | |
|
|
705 | =back |
|
|
706 | |
|
|
707 | =head2 Passing file descriptors |
|
|
708 | |
|
|
709 | Unlike L<AnyEvent::Fork>, this module has no in-built file handle or file |
|
|
710 | descriptor passing abilities. |
|
|
711 | |
|
|
712 | The reason is that passing file descriptors is extraordinary tricky |
|
|
713 | business, and conflicts with efficient batching of messages. |
|
|
714 | |
|
|
715 | There still is a method you can use: Create a |
|
|
716 | C<AnyEvent::Util::portable_socketpair> and C<send_fh> one half of it to |
|
|
717 | the process before you pass control to C<AnyEvent::Fork::RPC::run>. |
|
|
718 | |
|
|
719 | Whenever you want to pass a file descriptor, send an rpc request to the |
|
|
720 | child process (so it expects the descriptor), then send it over the other |
|
|
721 | half of the socketpair. The child should fetch the descriptor from the |
|
|
722 | half it has passed earlier. |
|
|
723 | |
|
|
724 | Here is some (untested) pseudocode to that effect: |
|
|
725 | |
|
|
726 | use AnyEvent::Util; |
|
|
727 | use AnyEvent::Fork::RPC; |
|
|
728 | use IO::FDPass; |
|
|
729 | |
|
|
730 | my ($s1, $s2) = AnyEvent::Util::portable_socketpair; |
|
|
731 | |
|
|
732 | my $rpc = AnyEvent::Fork |
|
|
733 | ->new |
|
|
734 | ->send_fh ($s2) |
|
|
735 | ->require ("MyWorker") |
|
|
736 | ->AnyEvent::Fork::RPC::run ("MyWorker::run" |
|
|
737 | init => "MyWorker::init", |
|
|
738 | ); |
|
|
739 | |
|
|
740 | undef $s2; # no need to keep it around |
|
|
741 | |
|
|
742 | # pass an fd |
|
|
743 | $rpc->("i'll send some fd now, please expect it!", my $cv = AE::cv); |
|
|
744 | |
|
|
745 | IO::FDPass fileno $s1, fileno $handle_to_pass; |
|
|
746 | |
|
|
747 | $cv->recv; |
|
|
748 | |
|
|
749 | The MyWorker module could look like this: |
|
|
750 | |
|
|
751 | package MyWorker; |
|
|
752 | |
|
|
753 | use IO::FDPass; |
|
|
754 | |
|
|
755 | my $s2; |
|
|
756 | |
|
|
757 | sub init { |
|
|
758 | $s2 = $_[0]; |
|
|
759 | } |
|
|
760 | |
|
|
761 | sub run { |
|
|
762 | if ($_[0] eq "i'll send some fd now, please expect it!") { |
|
|
763 | my $fd = IO::FDPass::recv fileno $s2; |
|
|
764 | ... |
|
|
765 | } |
|
|
766 | } |
|
|
767 | |
|
|
768 | Of course, this might be blocking if you pass a lot of file descriptors, |
|
|
769 | so you might want to look into L<AnyEvent::FDpasser> which can handle the |
|
|
770 | gory details. |
|
|
771 | |
436 | =head1 SEE ALSO |
772 | =head1 SEE ALSO |
437 | |
773 | |
438 | L<AnyEvent::Fork> (to create the processes in the first place), |
774 | L<AnyEvent::Fork> (to create the processes in the first place), |
439 | L<AnyEvent::Fork::Pool> (to manage whole pools of processes). |
775 | L<AnyEvent::Fork::Pool> (to manage whole pools of processes). |
440 | |
776 | |