1 |
=head1 Message Passing for the Non-Blocked Mind |
2 |
|
3 |
=head1 Introduction and Terminology |
4 |
|
5 |
This is a tutorial about how to get the swing of the new L<AnyEvent::MP> |
6 |
module, which allows programs to transparently pass messages within the |
7 |
process and to other processes on the same or a different host. |
8 |
|
9 |
What kind of messages? Basically a message here means a list of Perl |
10 |
strings, numbers, hashes and arrays, anything that can be expressed as a |
11 |
L<JSON> text (as JSON is the default serialiser in the protocol). Here are |
12 |
two examples: |
13 |
|
14 |
write_log => 1251555874, "action was successful.\n" |
15 |
123, ["a", "b", "c"], { foo => "bar" } |
16 |
|
17 |
When using L<AnyEvent::MP> it is customary to use a descriptive string as |
18 |
first element of a message that indicates the type of the message. This |
19 |
element is called a I<tag> in L<AnyEvent::MP>, as some API functions |
20 |
(C<rcv>) support matching it directly. |
21 |
|
22 |
Supposedly you want to send a ping message with your current time to |
23 |
somewhere, this is how such a message might look like (in Perl syntax): |
24 |
|
25 |
ping => 1251381636 |
26 |
|
27 |
Now that we know what a message is, to which entities are those |
28 |
messages being I<passed>? They are I<passed> to I<ports>. A I<port> is |
29 |
a destination for messages but also a context to execute code: when |
30 |
a runtime error occurs while executing code belonging to a port, the |
31 |
exception will be raised on the port and can even travel to interested |
32 |
parties on other nodes, which makes supervision of distributed processes |
33 |
easy. |
34 |
|
35 |
How do these ports relate to things you know? Each I<port> belongs |
36 |
to a I<node>, and a I<node> is just the UNIX process that runs your |
37 |
L<AnyEvent::MP> application. |
38 |
|
39 |
Each I<node> is distinguished from other I<nodes> running on the same or |
40 |
another host in a network by its I<node ID>. A I<node ID> is simply a |
41 |
unique string chosen manually or assigned by L<AnyEvent::MP> in some way |
42 |
(UNIX nodename, random string...). |
43 |
|
44 |
Here is a diagram about how I<nodes>, I<ports> and UNIX processes relate |
45 |
to each other. The setup consists of two nodes (more are of course |
46 |
possible): Node C<A> (in UNIX process 7066) with the ports C<ABC> and |
47 |
C<DEF>. And the node C<B> (in UNIX process 8321) with the ports C<FOO> and |
48 |
C<BAR>. |
49 |
|
50 |
|
51 |
|- PID: 7066 -| |- PID: 8321 -| |
52 |
| | | | |
53 |
| Node ID: A | | Node ID: B | |
54 |
| | | | |
55 |
| Port ABC =|= <----\ /-----> =|= Port FOO | |
56 |
| | X | | |
57 |
| Port DEF =|= <----/ \-----> =|= Port BAR | |
58 |
| | | | |
59 |
|-------------| |-------------| |
60 |
|
61 |
The strings for the I<port IDs> here are just for illustrative |
62 |
purposes: Even though I<ports> in L<AnyEvent::MP> are also identified by |
63 |
strings, they can't be chosen manually and are assigned by the system |
64 |
dynamically. These I<port IDs> are unique within a network and can also be |
65 |
used to identify senders, or even as message tags for instance. |
66 |
|
67 |
The next sections will explain the API of L<AnyEvent::MP> by going through |
68 |
a few simple examples. Later some more complex idioms are introduced, |
69 |
which are hopefully useful to solve some real world problems. |
70 |
|
71 |
=head2 Passing Your First Message |
72 |
|
73 |
For starters, let's have a look at the messaging API. The following |
74 |
example is just a demo to show the basic elements of message passing with |
75 |
L<AnyEvent::MP>. |
76 |
|
77 |
The example should print: C<Ending with: 123>, in a rather complicated |
78 |
way, by passing some message to a port. |
79 |
|
80 |
use AnyEvent; |
81 |
use AnyEvent::MP; |
82 |
|
83 |
my $end_cv = AnyEvent->condvar; |
84 |
|
85 |
my $port = port; |
86 |
|
87 |
rcv $port, test => sub { |
88 |
my ($data) = @_; |
89 |
$end_cv->send ($data); |
90 |
}; |
91 |
|
92 |
snd $port, test => 123; |
93 |
|
94 |
print "Ending with: " . $end_cv->recv . "\n"; |
95 |
|
96 |
It already uses most of the essential functions inside |
97 |
L<AnyEvent::MP>: First there is the C<port> function which creates a |
98 |
I<port> and will return it's I<port ID>, a simple string. |
99 |
|
100 |
This I<port ID> can be used to send messages to the port and install |
101 |
handlers to receive messages on the port. Since it is a simple string |
102 |
it can be safely passed to other I<nodes> in the network when you want |
103 |
to refer to that specific port (usually used for RPC, where you need |
104 |
to tell the other end which I<port> to send the reply to - messages in |
105 |
L<AnyEvent::MP> have a destination, but no source). |
106 |
|
107 |
The next function is C<rcv>: |
108 |
|
109 |
rcv $port, test => sub { ... }; |
110 |
|
111 |
It installs a receiver callback on the I<port> that specified as the first |
112 |
argument (it only works for "local" ports, i.e. ports created on the same |
113 |
node). The next argument, in this example C<test>, specifies a I<tag> to |
114 |
match. This means that whenever a message with the first element being |
115 |
the string C<test> is received, the callback is called with the remaining |
116 |
parts of that message. |
117 |
|
118 |
Messages can be sent with the C<snd> function, which is used like this in |
119 |
the example above: |
120 |
|
121 |
snd $port, test => 123; |
122 |
|
123 |
This will send the message C<'test', 123> to the I<port> with the I<port |
124 |
ID> stored in C<$port>. Since in this case the receiver has a I<tag> match |
125 |
on C<test> it will call the callback with the first argument being the |
126 |
number C<123>. |
127 |
|
128 |
The callback is a typical AnyEvent idiom: the callback just passes |
129 |
that number on to the I<condition variable> C<$end_cv> which will then |
130 |
pass the value to the print. Condition variables are out of the scope |
131 |
of this tutorial and not often used with ports, so please consult the |
132 |
L<AnyEvent::Intro> about them. |
133 |
|
134 |
Passing messages inside just one process is boring. Before we can move on |
135 |
and do interprocess message passing we first have to make sure some things |
136 |
have been set up correctly for our nodes to talk to each other. |
137 |
|
138 |
=head2 System Requirements and System Setup |
139 |
|
140 |
Before we can start with real IPC we have to make sure some things work on |
141 |
your system. |
142 |
|
143 |
First we have to setup a I<shared secret>: for two L<AnyEvent::MP> |
144 |
I<nodes> to be able to communicate with each other over the network it is |
145 |
necessary to setup the same I<shared secret> for both of them, so they can |
146 |
prove their trustworthyness to each other. |
147 |
|
148 |
The easiest way is to set this up is to use the F<aemp> utility: |
149 |
|
150 |
aemp gensecret |
151 |
|
152 |
This creates a F<$HOME/.perl-anyevent-mp> config file and generates a |
153 |
random shared secret. You can copy this file to any other system and |
154 |
then communicate over the network (via TCP) with it. You can also select |
155 |
your own shared secret (F<aemp setsecret>) and for increased security |
156 |
requirements you can even create (or configure) a TLS certificate (F<aemp |
157 |
gencert>), causing connections to not just be securely authenticated, but |
158 |
also to be encrypted and protected against tinkering. |
159 |
|
160 |
Connections will only be successfully established when the I<nodes> |
161 |
that want to connect to each other have the same I<shared secret> (or |
162 |
successfully verify the TLS certificate of the other side, in which case |
163 |
no shared secret is required). |
164 |
|
165 |
B<If something does not work as expected, and for example tcpdump shows |
166 |
that the connections are closed almost immediately, you should make sure |
167 |
that F<~/.perl-anyevent-mp> is the same on all hosts/user accounts that |
168 |
you try to connect with each other!> |
169 |
|
170 |
Thats is all for now, you will find some more advanced fiddling with the |
171 |
C<aemp> utility later. |
172 |
|
173 |
=head2 Shooting the Trouble |
174 |
|
175 |
Sometimes things go wrong, and AnyEvent::MP, being a professional module, |
176 |
does not gratuitously spill out messages to your screen. |
177 |
|
178 |
To help troubleshooting any issues, there are two environment variables |
179 |
that you can set. The first, C<PERL_ANYEVENT_MP_WARNLEVEL> sets the |
180 |
logging level. The default is C<5>, which means nothing much is |
181 |
printed. You can increase it to C<8> or C<9> to get more verbose |
182 |
output. This is example output when starting a node: |
183 |
|
184 |
2012-03-04 19:41:10 <8> node cerebro starting up. |
185 |
2012-03-04 19:41:10 <8> node listens on [10.0.0.1:4040]. |
186 |
2012-03-04 19:41:10 <9> trying connect to seed node 10.0.0.19:4040. |
187 |
2012-03-04 19:41:10 <9> 10.0.0.19:4040 connected as rain |
188 |
2012-03-04 19:41:10 <7> rain is up () |
189 |
|
190 |
A lot of info, but at least you can see that it does something. |
191 |
|
192 |
The other environment variable that can be useful is |
193 |
C<PERL_ANYEVENT_MP_TRACE>, which, when set to a true value, will cause |
194 |
most messages that are sent or received to be printed. For example, F<aemp |
195 |
restart rijk> might output these message exchanges: |
196 |
|
197 |
SND rijk <- [null,"eval","AnyEvent::Watchdog::Util::restart; ()","aemp/cerebro/z4kUPp2JT4#b"] |
198 |
SND rain <- [null,"g_slave",{"'l":{"aemp/cerebro/z4kUPp2JT4":["10.0.0.1:48168"]}}] |
199 |
SND rain <- [null,"g_find","rijk"] |
200 |
RCV rain -> ["","g_found","rijk",["10.0.0.23:4040"]] |
201 |
RCV rijk -> ["b",""] |
202 |
|
203 |
=head1 PART 1: Passing Messages Between Processes |
204 |
|
205 |
=head2 The Receiver |
206 |
|
207 |
Lets split the previous example up into two programs: one that contains |
208 |
the sender and one for the receiver. First the receiver application, in |
209 |
full: |
210 |
|
211 |
use AnyEvent; |
212 |
use AnyEvent::MP; |
213 |
|
214 |
configure nodeid => "eg_receiver/%u", binds => ["*:4040"]; |
215 |
|
216 |
my $port = port; |
217 |
db_set eg_receivers => $port; |
218 |
|
219 |
rcv $port, test => sub { |
220 |
my ($data, $reply_port) = @_; |
221 |
|
222 |
print "Received data: " . $data . "\n"; |
223 |
}; |
224 |
|
225 |
AnyEvent->condvar->recv; |
226 |
|
227 |
Now, that wasn't too bad, was it? OK, let's go through the new functions |
228 |
that have been used. |
229 |
|
230 |
=head3 C<configure> and Joining and Maintaining the Network |
231 |
|
232 |
First let's have a look at C<configure>: |
233 |
|
234 |
configure nodeid => "eg_receiver/%u", binds => ["*:4040"]; |
235 |
|
236 |
Before we are able to send messages to other nodes we have to initialise |
237 |
ourself to become a "distributed node". Initialising a node means naming |
238 |
the node and binding some TCP listeners so that other nodes can |
239 |
contact it. |
240 |
|
241 |
Additionally, to actually link all nodes in a network together, you can |
242 |
specify a number of seed addresses, which will be used by the node to |
243 |
connect itself into an existing network, as we will see shortly. |
244 |
|
245 |
All of this (and more) can be passed to the C<configure> function - later |
246 |
we will see how we can do all this without even passing anything to |
247 |
C<configure>! |
248 |
|
249 |
The first parameter, C<nodeid>, specified the node ID (in this case |
250 |
C<eg_receiver/%u> - the default is to use the node name of the current |
251 |
host plus C</%u>, which goves the node a name with a random suffix to |
252 |
make it unique, but for this example we want the node to have a bit more |
253 |
personality, and name it C<eg_receiver> with a random suffix. |
254 |
|
255 |
Why the random suffix? Node IDs need to be unique within the network and |
256 |
appending a random suffix is the easiest way to do that. |
257 |
|
258 |
The second parameter, C<binds>, specifies a list of C<address:port> pairs |
259 |
to bind TCP listeners on. The special "address" of C<*> means to bind on |
260 |
every local IP address (this might not work on every OS, so explicit IP |
261 |
addresses are best). |
262 |
|
263 |
The reason to bind on a TCP port is not just that other nodes can connect |
264 |
to us: if no binds are specified, the node will still bind on a dynamic |
265 |
port on all local addresses - but in this case we won't know the port, and |
266 |
cannot tell other nodes to connect to it as seed node. |
267 |
|
268 |
Now, a I<seed> is simply the TCP address of some other node in the |
269 |
network, often the same string as used for the C<binds> parameter of the |
270 |
other node. The need for seeds is easy to explain: I<somehow> the nodes |
271 |
of an aemp network have to find each other, and often this means over the |
272 |
internet. So broadcasts are out. |
273 |
|
274 |
Instead, a node usually specifies the addresses of a few (for redundancy) |
275 |
other nodes, some of which should be up. Two nodes can set each other as |
276 |
seeds without any issues. You could even specify all nodes as seeds for |
277 |
all nodes, for total redundancy. But the common case is to have some more |
278 |
or less central, stable servers running seed services for other nodes. |
279 |
|
280 |
All you need to do to ensure that an AnyEvent::MP network connects |
281 |
together is to make sure that all connections from nodes to their seed |
282 |
nodes I<somehow> span the whole network. The simplest way to do that would |
283 |
be for all nodes to specify a single node as seed node, and you would get |
284 |
a star topology. If you specify all nodes as seed nodes, you get a fully |
285 |
meshed network (that's what previous releases of AnyEvent::MP actually |
286 |
did). |
287 |
|
288 |
A node tries to keep connections open to all of it's seed nodes at all |
289 |
times, while other connections are made on demand only. |
290 |
|
291 |
All of this ensures that the network stays one network - even if all the |
292 |
nodes in one half of the net are separated from the nodes in the other |
293 |
half by some network problem, once that is over, they will eventually |
294 |
become a single network again. |
295 |
|
296 |
In addition to creating the network, a node also expects the seed nodes to |
297 |
run the shared database service - if need be, by automatically starting it, |
298 |
so you don't normally need to configure this explicitly. |
299 |
|
300 |
#TODO# later?#d# |
301 |
The process of joining a network takes time, during which the node |
302 |
is already running. This means it takes time until the node is |
303 |
fully connected, and information about services in the network are |
304 |
available. This is why most AnyEvent::MP programs start by waiting a while |
305 |
until the information they need is available. |
306 |
|
307 |
We will see how this is done later, in the sender program. |
308 |
|
309 |
=head3 Registering the Receiver |
310 |
|
311 |
Coming back to our example, after the node has been configured for network |
312 |
access, it is time to publish some service, namely the receive service. |
313 |
|
314 |
For that, let's look at the next lines: |
315 |
|
316 |
my $port = port; |
317 |
db_set eg_receivers => $port; |
318 |
|
319 |
The C<port> function has already been discussed. It simply creates a new |
320 |
I<port> and returns the I<port ID>. The C<db_set> function, however, is |
321 |
new: The first argument is the name of a I<database family> and the second |
322 |
argument is the name of a I<subkey> within that family. The third argument |
323 |
would be the I<value> to be associated with the family and subkey, but, |
324 |
since it is missing, it will simply be C<undef>. |
325 |
|
326 |
What is a "family" you wonder? Well, AnyEvent::MP comes with a distributed |
327 |
database. This database runs on so-called "global" nodes, which usually |
328 |
are the seed nodes of your network. The database structure is "simply" a |
329 |
hash of hashes of values. |
330 |
|
331 |
To illustrate this with Perl syntax, assume the database was stored in |
332 |
C<%DB>, then the C<db_set> function more or less would do this: |
333 |
|
334 |
$DB{eg_receivers}{$port} = undef; |
335 |
|
336 |
So the ominous "family" selects a hash in the database, and the "subkey" |
337 |
is simply the key in this hash - C<db_set> very much works like this |
338 |
assignment. |
339 |
|
340 |
The family namespace is shared by all nodes in a network, so the names |
341 |
should be reasonably unique, for example, they could start with the name |
342 |
of your module, or the name of the program, using your port name or node |
343 |
name as subkey. |
344 |
|
345 |
The purpose behind adding this key to the database is that the sender can |
346 |
look it up and find our port. We will shortly see how. |
347 |
|
348 |
The last step in the example is to set up a receiver callback for those |
349 |
messages, just as was discussed in the first example. We again match |
350 |
for the tag C<test>. The difference is that this time we don't exit the |
351 |
application after receiving the first message. Instead we continue to wait |
352 |
for new messages indefinitely. |
353 |
|
354 |
=head2 The Sender |
355 |
|
356 |
OK, now let's take a look at the sender code: |
357 |
|
358 |
use AnyEvent; |
359 |
use AnyEvent::MP; |
360 |
|
361 |
configure nodeid => "eg_sender/%u", seeds => ["*:4040"]; |
362 |
|
363 |
my $guard = db_mon eg_receivers => sub { |
364 |
my ($family, $a, $c, $d) = @_; |
365 |
return unless %$family; |
366 |
|
367 |
# now there are some receivers, send them a message |
368 |
snd $_ => test => time |
369 |
for keys %$family; |
370 |
}; |
371 |
|
372 |
AnyEvent->condvar->recv; |
373 |
|
374 |
It's even less code. The C<configure> serves the same purpose as in the |
375 |
receiver, but instead of specifying binds we specify a list of seeds - the |
376 |
only seed happens to be the same as the bind used by the receiver, which |
377 |
therefore becomes our seed node. |
378 |
|
379 |
Remember the part about having to wait till things become available? Well, |
380 |
after configure returns, nothing has been done yet - the node is not |
381 |
connected to the network, knows nothing about the database contents, and |
382 |
it can take ages (for a computer :) for this situation to change. |
383 |
|
384 |
Therefore, the sender waits, in this case by using the C<db_mon> |
385 |
function. This function registers an interest in a specific database |
386 |
family (in this case C<eg_receivers>). Each time something inside the |
387 |
family changes (a key is added, changed or deleted), it will call our |
388 |
callback with the family hash as first argument, and the list of keys as |
389 |
second argument. |
390 |
|
391 |
The callback only checks whether the C<%$family> has is empty - if it is, |
392 |
then it doesn't do anything. But eventually the family will contain the |
393 |
port subkey we set in the sender. Then it will send a message to it (and |
394 |
any other receiver in the same family). Likewise, should the receiver go |
395 |
away and come back, or should another receiver come up, it will again send |
396 |
a message to all of them. |
397 |
|
398 |
You can experiment by having multiple receivers - you have to change the |
399 |
"binds" parameter in the receiver to the seeds used in the sender to start |
400 |
up additional receivers, but then you can start as many as you like. If |
401 |
you specify proper IP addresses for the seeds, you can even run them on |
402 |
different computers. |
403 |
|
404 |
Each time you start the sender, it will send a message to all receivers it |
405 |
finds (you have to interrupt it manually afterwards). |
406 |
|
407 |
Additional experiments you could try include using |
408 |
C<PERL_ANYEVENT_MP_TRACE=1> to see which messages are exchanged, or |
409 |
starting the sender before the receiver and see how long it then takes to |
410 |
find the receiver. |
411 |
|
412 |
=head3 Splitting Network Configuration and Application Code |
413 |
|
414 |
OK, so far, this works reasonably. In the real world, however, the person |
415 |
configuring your application to run on a specific network (the end user |
416 |
or network administrator) is often different to the person coding the |
417 |
application. |
418 |
|
419 |
Or to put it differently: the arguments passed to configure are usually |
420 |
provided not by the programmer, but by whoever is deploying the program - |
421 |
even in the example above, we would like to be able to just start senders |
422 |
and receivers without having to patch the programs. |
423 |
|
424 |
To make this easy, AnyEvent::MP supports a simple configuration database, |
425 |
using profiles, which can be managed using the F<aemp> command-line |
426 |
utility (yes, this section is about the advanced tinkering mentioned |
427 |
before). |
428 |
|
429 |
When you change both programs above to simply call |
430 |
|
431 |
configure; |
432 |
|
433 |
then AnyEvent::MP tries to look up a profile using the current node name |
434 |
in its configuration database, falling back to some global default. |
435 |
|
436 |
You can run "generic" nodes using the F<aemp> utility as well, and we will |
437 |
exploit this in the following way: we configure a profile "seed" and run |
438 |
a node using it, whose sole purpose is to be a seed node for our example |
439 |
programs. |
440 |
|
441 |
We bind the seed node to port 4040 on all interfaces: |
442 |
|
443 |
aemp profile seed binds "*:4040" |
444 |
|
445 |
And we configure all nodes to use this as seed node (this only works when |
446 |
running on the same host, for multiple machines you would replace the C<*> |
447 |
by the IP address or hostname of the node running the seed), by changing |
448 |
the global settings shared between all profiles: |
449 |
|
450 |
aemp seeds "*:4040" |
451 |
|
452 |
Then we run the seed node: |
453 |
|
454 |
aemp run profile seed |
455 |
|
456 |
After that, we can start as many other nodes as we want, and they will |
457 |
all use our generic seed node to discover each other. The reason we can |
458 |
start our existing programs even though they specify "incompatible" |
459 |
parameters to C<configure> is that the configuration file (by default) |
460 |
takes precedence over any arguments passed to C<configure>. |
461 |
|
462 |
That's all for now - next we will teach you about monitoring by writing a |
463 |
simple chat client and server :) |
464 |
|
465 |
=head1 PART 2: Monitoring, Supervising, Exception Handling and Recovery |
466 |
|
467 |
That's a mouthful, so what does it mean? Our previous example is what one |
468 |
could call "very loosely coupled" - the sender doesn't care about whether |
469 |
there are any receivers, and the receivers do not care if there is any |
470 |
sender. |
471 |
|
472 |
This can work fine for simple services, but most real-world applications |
473 |
want to ensure that the side they are expecting to be there is actually |
474 |
there. Going one step further: most bigger real-world applications even |
475 |
want to ensure that if some component is missing, or has crashed, it will |
476 |
still be there, by recovering and restarting the service. |
477 |
|
478 |
AnyEvent::MP supports this by catching exceptions and network problems, |
479 |
and notifying interested parties of these. |
480 |
|
481 |
=head2 Exceptions, Port Context, Network Errors and Monitors |
482 |
|
483 |
=head3 Exceptions |
484 |
|
485 |
Exceptions are handled on a per-port basis: all receive callbacks are |
486 |
executed in a special context, the so-called I<port-context>: code |
487 |
that throws an otherwise uncaught exception will cause the port to be |
488 |
C<kil>led. Killed ports are destroyed automatically (killing ports is |
489 |
actually the only way to free ports). |
490 |
|
491 |
Ports can be monitored, even from a different node and host, and when a |
492 |
port is killed, any entity monitoring it will be notified. |
493 |
|
494 |
Here is a simple example: |
495 |
|
496 |
use AnyEvent::MP; |
497 |
|
498 |
# create a port, it always dies |
499 |
my $port = port { die "oops" }; |
500 |
|
501 |
# monitor it |
502 |
mon $port, sub { |
503 |
warn "$port was killed (with reason @_)"; |
504 |
}; |
505 |
|
506 |
# now send it some message, causing it to die: |
507 |
snd $port; |
508 |
|
509 |
AnyEvent->condvar->recv; |
510 |
|
511 |
It first creates a port whose only action is to throw an exception, |
512 |
and the monitors it with the C<mon> function. Afterwards it sends it a |
513 |
message, causing it to die and call the monitoring callback: |
514 |
|
515 |
anon/6WmIpj.a was killed (with reason die oops at xxx line 5.) at xxx line 9. |
516 |
|
517 |
The callback was actually passed two arguments: C<die>, to indicate it |
518 |
did throw an I<exception> as opposed to, say, a network error, and the |
519 |
exception message itself. |
520 |
|
521 |
What happens when a port is killed before we have a chance to monitor |
522 |
it? Granted, this is highly unlikely in our example, but when you program |
523 |
in a network this can easily happen due to races between nodes. |
524 |
|
525 |
use AnyEvent::MP; |
526 |
|
527 |
my $port = port { die "oops" }; |
528 |
|
529 |
snd $port; |
530 |
|
531 |
mon $port, sub { |
532 |
warn "$port was killed (with reason @_)"; |
533 |
}; |
534 |
|
535 |
AnyEvent->condvar->recv; |
536 |
|
537 |
This time we will get something else: |
538 |
|
539 |
2012-03-21 00:50:36 <2> unmonitored local port fADb died with reason: die oops at - line 3. |
540 |
anon/fADb was killed (with reason no_such_port cannot monitor nonexistent port) |
541 |
|
542 |
The first line is a warning that is printed when a port dies that isn't |
543 |
being monitored, because that is normally a bug. When later a C<mon> is |
544 |
attempted, it is immediately killed, because the port is already gone. The |
545 |
kill reason is now C<no_such_port> with some descriptive (we hope) error |
546 |
message. |
547 |
|
548 |
As you probably suspect from these examples, the kill reason is usually |
549 |
some identifier as first argument and a human-readable error message as |
550 |
second argument - all kill reasons by AnyEvent::MP itself follow this |
551 |
pattern. But the kill reason can be anything: it is simply a list of |
552 |
values you can choose yourself. It can even be nothing (an empty list) - |
553 |
this is called a "normal" kill. |
554 |
|
555 |
Apart from die'ing, you can kill ports manually using the C<kil> |
556 |
function. Using the C<kil> function will be treated like an error when a |
557 |
non-empty reason is specified: |
558 |
|
559 |
kil $port, custom_error => "don't like your steenking face"; |
560 |
|
561 |
And a I<normal> kill without any reason arguments: |
562 |
|
563 |
kil $port; |
564 |
|
565 |
By now you probably wonder what this "normal" kill business is: A common |
566 |
idiom is to not specify a callback to C<mon>, but another port, such as |
567 |
C<$SELF>: |
568 |
|
569 |
mon $port, $SELF; |
570 |
|
571 |
This basically means "monitor $port and kill me when it crashes" - and |
572 |
the thing is, a "normal" kill does not count as a crash. This way you can |
573 |
easily link ports together and make them crash together on errors, while |
574 |
allowing you to remove a port silently when it has done it's job properly. |
575 |
|
576 |
=head3 Port Context |
577 |
|
578 |
Code runs in the so-called "port context". That means C<$SELF> contains |
579 |
its own port ID and exceptions that the code throws will be caught. |
580 |
|
581 |
Since AnyEvent::MP is event-based, it is not uncommon to register |
582 |
callbacks from within C<rcv> handlers. As example, assume that the |
583 |
following port receive handler wants to C<die> a second later, using |
584 |
C<after>: |
585 |
|
586 |
my $port = port { |
587 |
after 1, sub { die "oops" }; |
588 |
}; |
589 |
|
590 |
If you try this out, you would find it does not work - when the C<after> |
591 |
callback is executed, it does not run in the port context anymore, so |
592 |
exceptions will not be caught. |
593 |
|
594 |
For these cases, AnyEvent::MP exports a special "closure constructor" |
595 |
called C<psub>, which works mostly like perl's built-in C<sub>: |
596 |
|
597 |
my $port = port { |
598 |
after 1, psub { die "oops" }; |
599 |
}; |
600 |
|
601 |
C<psub> remembers the port context and returns a code reference. When the |
602 |
code reference is invoked, it will run the code block within the context |
603 |
that it was created in, so exception handling once more works as expected. |
604 |
|
605 |
There is even a way to temporarily execute code in the context of some |
606 |
port, namely C<peval>: |
607 |
|
608 |
peval $port, sub { |
609 |
# die'ing here will kil $port |
610 |
}; |
611 |
|
612 |
The C<peval> function temporarily replaces C<$SELF> by the given C<$port> |
613 |
and then executes the given sub in a port context. |
614 |
|
615 |
=head3 Network Errors and the AEMP Guarantee |
616 |
|
617 |
Earlier we mentioned another important source of monitoring failures: |
618 |
network problems. When a node loses connection to another node, it will |
619 |
invoke all monitoring actions, just as if the port was killed, I<even if |
620 |
it is possible that the port is still happily alive on another node> (not |
621 |
being able to talk to a node means we have no clue what's going on with |
622 |
it, it could be crashed, but also still running without knowing we lost |
623 |
the connection). |
624 |
|
625 |
So another way to view monitors is: "notify me when some of my messages |
626 |
couldn't be delivered". AEMP has a guarantee about message delivery to a |
627 |
port: After starting a monitor, any message sent to a port will either |
628 |
be delivered, or, when it is lost, any further messages will also be lost |
629 |
until the monitoring action is invoked. After that, further messages |
630 |
I<might> get delivered again. |
631 |
|
632 |
This doesn't sound like a very big guarantee, but it is kind of the best |
633 |
you can get while staying sane: Specifically, it means that there will be |
634 |
no "holes" in the message sequence: all messages sent are delivered in |
635 |
order, without any of them missing in between, and when some were lost, |
636 |
you I<will> be notified of that, so you can take recovery action. |
637 |
|
638 |
And, obviously, the guarantee only works in the presence of |
639 |
correctly-working hardware, and no relevant bugs inside AEMP itself. |
640 |
|
641 |
=head3 Supervising |
642 |
|
643 |
OK, so how is this crashing-everything-stuff going to make applications |
644 |
I<more> stable? Well, in fact, the goal is not really to make them |
645 |
more stable, but to make them more resilient against actual errors |
646 |
and crashes. And this is not done by crashing I<everything>, but by |
647 |
crashing everything except a I<supervisor> that then cleans up and sgtarts |
648 |
everything again. |
649 |
|
650 |
A supervisor is simply some code that ensures that an application (or a |
651 |
part of it) is running, and if it crashes, is restarted properly. That is, |
652 |
it supervises a service by starting and restarting it, as necessary. |
653 |
|
654 |
To show how to do all this we will create a simple chat server that can |
655 |
handle many chat clients. Both server and clients can be killed and |
656 |
restarted, and even crash, to some extent, without disturbing the chat |
657 |
functionality. |
658 |
|
659 |
=head2 Chatting, the Resilient Way |
660 |
|
661 |
Without further ado, here is the chat server (to run it, we assume the |
662 |
set-up explained earlier, with a separate F<aemp run seed> node): |
663 |
|
664 |
use common::sense; |
665 |
use AnyEvent::MP; |
666 |
use AnyEvent::MP::Global; |
667 |
|
668 |
configure; |
669 |
|
670 |
my %clients; |
671 |
|
672 |
sub msg { |
673 |
print "relaying: $_[0]\n"; |
674 |
snd $_, $_[0] |
675 |
for values %clients; |
676 |
} |
677 |
|
678 |
our $server = port; |
679 |
|
680 |
rcv $server, join => sub { |
681 |
my ($client, $nick) = @_; |
682 |
|
683 |
$clients{$client} = $client; |
684 |
|
685 |
mon $client, sub { |
686 |
delete $clients{$client}; |
687 |
msg "$nick (quits, @_)"; |
688 |
}; |
689 |
msg "$nick (joins)"; |
690 |
}; |
691 |
|
692 |
rcv $server, privmsg => sub { |
693 |
my ($nick, $msg) = @_; |
694 |
msg "$nick: $msg"; |
695 |
}; |
696 |
|
697 |
db_set eg_chat_server => $server; |
698 |
|
699 |
warn "server ready.\n"; |
700 |
|
701 |
AnyEvent->condvar->recv; |
702 |
|
703 |
Looks like a lot, but it is actually quite simple: after your usual |
704 |
preamble (this time we use common sense), we define a helper function that |
705 |
sends some message to every registered chat client: |
706 |
|
707 |
sub msg { |
708 |
print "relaying: $_[0]\n"; |
709 |
snd $_, $_[0] |
710 |
for values %clients; |
711 |
} |
712 |
|
713 |
The clients are stored in the hash C<%client>. Then we define a server |
714 |
port and install two receivers on it, C<join>, which is sent by clients |
715 |
to join the chat, and C<privmsg>, that clients use to send actual chat |
716 |
messages. |
717 |
|
718 |
C<join> is most complicated. It expects the client port and the nickname |
719 |
to be passed in the message, and registers the client in C<%clients>. |
720 |
|
721 |
rcv $server, join => sub { |
722 |
my ($client, $nick) = @_; |
723 |
|
724 |
$clients{$client} = $client; |
725 |
|
726 |
The next step is to monitor the client. The monitoring action removes the |
727 |
client and sends a quit message with the error to all remaining clients. |
728 |
|
729 |
mon $client, sub { |
730 |
delete $clients{$client}; |
731 |
msg "$nick (quits, @_)"; |
732 |
}; |
733 |
|
734 |
And finally, it creates a join message and sends it to all clients. |
735 |
|
736 |
msg "$nick (joins)"; |
737 |
}; |
738 |
|
739 |
The C<privmsg> callback simply broadcasts the message to all clients: |
740 |
|
741 |
rcv $server, privmsg => sub { |
742 |
my ($nick, $msg) = @_; |
743 |
msg "$nick: $msg"; |
744 |
}; |
745 |
|
746 |
And finally, the server registers itself in the server group, so that |
747 |
clients can find it: |
748 |
|
749 |
db_set eg_chat_server => $server; |
750 |
|
751 |
Well, well... and where is this supervisor stuff? Well... we cheated, |
752 |
it's not there. To not overcomplicate the example, we only put it into |
753 |
the..... CLIENT! |
754 |
|
755 |
=head3 The Client, and a Supervisor! |
756 |
|
757 |
Again, here is the client, including supervisor, which makes it a bit |
758 |
longer: |
759 |
|
760 |
use common::sense; |
761 |
use AnyEvent::MP; |
762 |
|
763 |
my $nick = shift || "anonymous"; |
764 |
|
765 |
configure; |
766 |
|
767 |
my ($client, $server); |
768 |
|
769 |
sub server_connect { |
770 |
my $db_mon; |
771 |
$db_mon = db_mon eg_chat_server => sub { |
772 |
return unless %{ $_[0] }; |
773 |
undef $db_mon; |
774 |
|
775 |
print "\rconnecting...\n"; |
776 |
|
777 |
$client = port { print "\r \r@_\n> " }; |
778 |
mon $client, sub { |
779 |
print "\rdisconnected @_\n"; |
780 |
&server_connect; |
781 |
}; |
782 |
|
783 |
$server = (keys %{ $_[0] })[0]; |
784 |
|
785 |
snd $server, join => $client, $nick; |
786 |
mon $server, $client; |
787 |
}; |
788 |
} |
789 |
|
790 |
server_connect; |
791 |
|
792 |
my $w = AnyEvent->io (fh => 0, poll => 'r', cb => sub { |
793 |
chomp (my $line = <STDIN>); |
794 |
print "> "; |
795 |
snd $server, privmsg => $nick, $line |
796 |
if $server; |
797 |
}); |
798 |
|
799 |
$| = 1; |
800 |
print "> "; |
801 |
AnyEvent->condvar->recv; |
802 |
|
803 |
The first thing the client does is to store the nick name (which is |
804 |
expected as the only command line argument) in C<$nick>, for further |
805 |
usage. |
806 |
|
807 |
The next relevant thing is... finally... the supervisor: |
808 |
|
809 |
sub server_connect { |
810 |
my $db_mon; |
811 |
$db_mon = db_mon eg_chat_server => sub { |
812 |
return unless %{ $_[0] }; |
813 |
undef $db_mon; # stop monitoring |
814 |
|
815 |
This monitors the C<eg_chat_server> database family. It waits until a |
816 |
chat server becomes available. When that happens, it "connects" to it |
817 |
by creating a client port that receives and prints chat messages, and |
818 |
monitoring it: |
819 |
|
820 |
$client = port { print "\r \r@_\n> " }; |
821 |
mon $client, sub { |
822 |
print "\rdisconnected @_\n"; |
823 |
&server_connect; |
824 |
}; |
825 |
|
826 |
If the client port dies (for whatever reason), the "supervisor" will start |
827 |
looking for a server again - the semantics of C<db_mon> ensure that it |
828 |
will immediately find it if there is a server port. |
829 |
|
830 |
After this, everything is ready: the client will send a C<join> message |
831 |
with its local port to the server, and start monitoring it: |
832 |
|
833 |
$server = (keys %{ $_[0] })[0]; |
834 |
|
835 |
snd $server, join => $client, $nick; |
836 |
mon $server, $client; |
837 |
} |
838 |
|
839 |
This second monitor will ensure that, when the server port crashes or goes |
840 |
away (e.g. due to network problems), the client port will be killed as |
841 |
well. This tells the user that the client was disconnected, and will then |
842 |
start to connect the server again. |
843 |
|
844 |
The rest of the program deals with the boring details of actually invoking |
845 |
the supervisor function to start the whole client process and handle the |
846 |
actual terminal input, sending it to the server. |
847 |
|
848 |
Now... the "supervisor" in this example is a bit of a cheat - it doesn't |
849 |
really clean up much (because the cleanup done by AnyEvent::MP suffices), |
850 |
and there isn't much of a restarting action either - if the server isn't |
851 |
there because it crashed, well, it isn't there. |
852 |
|
853 |
In the real world, one would often add a timeout that would trigger when |
854 |
the server couldn't be found within some time limit, and then complain, |
855 |
or even try to start a new server. Or the supervisor would have to do |
856 |
some real cleanups, such as rolling back database transactions when the |
857 |
database thread crashes. For this simple chat server, however, this simple |
858 |
supervisor works fine. Hopefully future versions of AnyEvent::MP will |
859 |
offer some predefined supervisors, for now you will have to code it on |
860 |
your own. |
861 |
|
862 |
You should now try to start the server and one or more clients in different |
863 |
terminal windows (and the seed node): |
864 |
|
865 |
perl eg/chat_client nick1 |
866 |
perl eg/chat_client nick2 |
867 |
perl eg/chat_server |
868 |
aemp run profile seed |
869 |
|
870 |
And then you can experiment with chatting, killing one or more clients, or |
871 |
stopping and restarting the server, to see the monitoring in action. |
872 |
|
873 |
The crucial point you should understand from this example is that |
874 |
monitoring is usually symmetric: when you monitor some other port, |
875 |
potentially on another node, that other port usually should monitor you, |
876 |
too, so when the connection dies, both ports get killed, or at least both |
877 |
sides can take corrective action. Exceptions are "servers" that serve |
878 |
multiple clients at once and might only wish to clean up, and supervisors, |
879 |
who of course should not normally get killed (unless they, too, have a |
880 |
supervisor). |
881 |
|
882 |
If you often think in object-oriented terms, then you can think of a port |
883 |
as an object: C<port> is the constructor, the receive callbacks set by |
884 |
C<rcv> act as methods, the C<kil> function becomes the explicit destructor |
885 |
and C<mon> installs a destructor hook. Unlike conventional object oriented |
886 |
programming, it can make sense to exchange port IDs more freely (for |
887 |
example, to monitor one port from another), because it is cheap to send |
888 |
port IDs over the network, and AnyEvent::MP blurs the distinction between |
889 |
local and remote ports. |
890 |
|
891 |
Lastly, there is ample room for improvement in this example: the server |
892 |
should probably remember the nickname in the C<join> handler instead of |
893 |
expecting it in every chat message, it should probably monitor itself, and |
894 |
the client should not try to send any messages unless a server is actually |
895 |
connected. |
896 |
|
897 |
=head1 PART 3: TIMTOWTDI: Virtual Connections |
898 |
|
899 |
The chat system developed in the previous sections is very "traditional" |
900 |
in a way: you start some server(s) and some clients statically and they |
901 |
start talking to each other. |
902 |
|
903 |
Sometimes applications work more like "services": They can run on almost |
904 |
any node and even talk to copies of themselves on other nodes in case they |
905 |
are distributed. The L<AnyEvent::MP::Global> service for example monitors |
906 |
nodes joining the network and sometimes even starts itself on other nodes. |
907 |
|
908 |
One good way to design such services is to put them into a module and |
909 |
create "virtual connections" to other nodes. We call this the "bridge |
910 |
head" method, because you start by I<creating a remote port> (the bridge |
911 |
head) and from that you start to bootstrap your application. |
912 |
|
913 |
Since that sounds rather theoretical, let us redesign the chat server and |
914 |
client using this design method. |
915 |
|
916 |
As usual, we start with the full program - here is the server: |
917 |
|
918 |
use common::sense; |
919 |
use AnyEvent::MP; |
920 |
|
921 |
configure; |
922 |
|
923 |
db_set eg_chat_server2 => $NODE; |
924 |
|
925 |
my %clients; |
926 |
|
927 |
sub msg { |
928 |
print "relaying: $_[0]\n"; |
929 |
snd $_, $_[0] |
930 |
for values %clients; |
931 |
} |
932 |
|
933 |
sub client_connect { |
934 |
my ($client, $nick) = @_; |
935 |
|
936 |
mon $client; |
937 |
mon $client, psub { |
938 |
delete $clients{$client}; |
939 |
msg "$nick (quits, @_)"; |
940 |
}; |
941 |
|
942 |
$clients{$client} = $client; |
943 |
|
944 |
msg "$nick (joins)"; |
945 |
|
946 |
rcv $SELF, sub { msg "$nick: $_[0]" }; |
947 |
} |
948 |
|
949 |
warn "server ready.\n"; |
950 |
|
951 |
AnyEvent->condvar->recv; |
952 |
|
953 |
It starts out not much different then the previous example, except that |
954 |
this time, we register the node port in the database and not a port we |
955 |
created - the clients only want to know which node the server should |
956 |
be running on, and there can only be one such server (or service) per |
957 |
node. In fact, the clients could also use some kind of election mechanism, |
958 |
to find the node with lowest node ID, or lowest load, or something like |
959 |
that. |
960 |
|
961 |
The much more interesting difference to the previous server is that |
962 |
indeed no server port is created - the server consists only of code, |
963 |
and "does" nothing by itself. All it "does" is to define a function |
964 |
named C<client_connect>, which expects a client port and a nick name as |
965 |
arguments. It then monitors the client port and binds a receive callback |
966 |
on C<$SELF>, which expects messages that in turn are broadcast to all |
967 |
clients. |
968 |
|
969 |
The two C<mon> calls are a bit tricky - the first C<mon> is a shorthand |
970 |
for C<mon $client, $SELF>. The second does the normal "client has gone |
971 |
away" clean-up action. |
972 |
|
973 |
The last line, the C<rcv $SELF>, is a good hint that something interesting |
974 |
is going on. And indeed, when looking at the client code, you can see a |
975 |
new function, C<spawn>: |
976 |
#todo# |
977 |
|
978 |
use common::sense; |
979 |
use AnyEvent::MP; |
980 |
use AnyEvent::MP::Global; |
981 |
|
982 |
my $nick = shift; |
983 |
|
984 |
configure; |
985 |
|
986 |
$| = 1; |
987 |
|
988 |
my $port = port; |
989 |
|
990 |
my ($client, $server); |
991 |
|
992 |
sub server_connect { |
993 |
my $servernodes = grp_get "eg_chat_server2" |
994 |
or return after 1, \&server_connect; |
995 |
|
996 |
print "\rconnecting...\n"; |
997 |
|
998 |
$client = port { print "\r \r@_\n> " }; |
999 |
mon $client, sub { |
1000 |
print "\rdisconnected @_\n"; |
1001 |
&server_connect; |
1002 |
}; |
1003 |
|
1004 |
$server = spawn $servernodes->[0], "::client_connect", $client, $nick; |
1005 |
mon $server, $client; |
1006 |
} |
1007 |
|
1008 |
server_connect; |
1009 |
|
1010 |
my $w = AnyEvent->io (fh => 0, poll => 'r', cb => sub { |
1011 |
chomp (my $line = <STDIN>); |
1012 |
print "> "; |
1013 |
snd $server, $line |
1014 |
if $server; |
1015 |
}); |
1016 |
|
1017 |
print "> "; |
1018 |
AnyEvent->condvar->recv; |
1019 |
|
1020 |
The client is quite similar to the previous one, but instead of contacting |
1021 |
the server I<port> (which no longer exists), it C<spawn>s (creates) a new |
1022 |
the server I<port on node>: |
1023 |
|
1024 |
$server = spawn $servernodes->[0], "::client_connect", $client, $nick; |
1025 |
mon $server, $client; |
1026 |
|
1027 |
And of course the first thing after creating it is monitoring it. |
1028 |
|
1029 |
Phew, let's go through this in slow motion: the C<spawn> function creates |
1030 |
a new port on a remote node and returns its port ID. After creating |
1031 |
the port it calls a function on the remote node, passing any remaining |
1032 |
arguments to it, and - most importantly - executes the function within |
1033 |
the context of the new port, so it can be manipulated by referring to |
1034 |
C<$SELF>. The init function can reside in a module (actually it normally |
1035 |
I<should> reside in a module) - AnyEvent::MP will automatically load the |
1036 |
module if the function isn't defined. |
1037 |
|
1038 |
The C<spawn> function returns immediately, which means you can instantly |
1039 |
send messages to the port, long before the remote node has even heard |
1040 |
of our request to create a port on it. In fact, the remote node might |
1041 |
not even be running. Despite these troubling facts, everything should |
1042 |
work just fine: if the node isn't running (or the init function throws an |
1043 |
exception), then the monitor will trigger because the port doesn't exist. |
1044 |
|
1045 |
If the spawn message gets delivered, but the monitoring message is not |
1046 |
because of network problems (extremely unlikely, but monitoring, after |
1047 |
all, is implemented by passing a message, and messages can get lost), then |
1048 |
this connection loss will eventually trigger the monitoring action. On the |
1049 |
remote node (which in return monitors the client) the port will also be |
1050 |
cleaned up on connection loss. When the remote node comes up again and our |
1051 |
monitoring message can be delivered, it will instantly fail because the |
1052 |
port has been cleaned up in the meantime. |
1053 |
|
1054 |
If your head is spinning by now, that's fine - just keep in mind, after |
1055 |
creating a port using C<spawn>, monitor it on the local node, and monitor |
1056 |
"the other side" from the remote node, and all will be cleaned up just |
1057 |
fine. |
1058 |
|
1059 |
=head2 Services |
1060 |
|
1061 |
Above it was mentioned that C<spawn> automatically loads modules. This can |
1062 |
be exploited in various useful ways. |
1063 |
|
1064 |
Assume for a moment you put the server into a file called |
1065 |
F<mymod/chatserver.pm> reachable from the current directory. Then you |
1066 |
could run a node there with: |
1067 |
|
1068 |
aemp run |
1069 |
|
1070 |
The other nodes could C<spawn> the server by using |
1071 |
C<mymod::chatserver::client_connect> as init function - without any other |
1072 |
configuration. |
1073 |
|
1074 |
Likewise, when you have some service that starts automatically when loaded |
1075 |
(similar to AnyEvent::MP::Global), then you can configure this service |
1076 |
statically: |
1077 |
|
1078 |
aemp profile mysrvnode services mymod::service:: |
1079 |
aemp run profile mysrvnode |
1080 |
|
1081 |
And the module will automatically be loaded in the node, as specifying a |
1082 |
module name (with C<::>-suffix) will simply load the module, which is then |
1083 |
free to do whatever it wants. |
1084 |
|
1085 |
Of course, you can also do it in the much more standard way by writing |
1086 |
a module (e.g. C<BK::Backend::IRC>), installing it as part of a module |
1087 |
distribution and then configure nodes. For example, if I wanted to run the |
1088 |
Bummskraut IRC backend on a machine named "ruth", I could do this: |
1089 |
|
1090 |
aemp profile ruth addservice BK::Backend::IRC:: |
1091 |
|
1092 |
And any F<aemp run> on that host will automatically have the Bummskraut |
1093 |
IRC backend running. |
1094 |
|
1095 |
There are plenty of possibilities you can use - it's all up to you how you |
1096 |
structure your application. |
1097 |
|
1098 |
=head1 PART 4: Coro::MP - selective receive |
1099 |
|
1100 |
Not all problems lend themselves naturally to an event-based solution: |
1101 |
sometimes things are easier if you can decide in what order you want to |
1102 |
receive messages, regardless of the order in which they were sent. |
1103 |
|
1104 |
In these cases, L<Coro::MP> can provide a nice solution: instead of |
1105 |
registering callbacks for each message type, C<Coro::MP> attaches a |
1106 |
(coro-) thread to a port. The thread can then opt to selectively receive |
1107 |
messages it is interested in. Other messages are not lost, but queued, and |
1108 |
can be received at a later time. |
1109 |
|
1110 |
The C<Coro::MP> module is not part of L<AnyEvent::MP>, but a separate |
1111 |
module. It is, however, tightly integrated into C<AnyEvent::MP> - the |
1112 |
ports it creates are fully compatible to C<AnyEvent::MP> ports. |
1113 |
|
1114 |
In fact, C<Coro::MP> is more of an extension than a separate module: all |
1115 |
functions exported by C<AnyEvent::MP> are exported by it as well. |
1116 |
|
1117 |
To illustrate how programing with C<Coro::MP> looks like, consider the |
1118 |
following (slightly contrived) example: Let's implement a server that |
1119 |
accepts a C<< (write_file =>, $port, $path) >> message with a (source) |
1120 |
port and a filename, followed by as many C<< (data => $port, $data) >> |
1121 |
messages as required to fill the file, followed by an empty C<< (data => |
1122 |
$port) >> message. |
1123 |
|
1124 |
The server only writes a single file at a time, other requests will stay |
1125 |
in the queue until the current file has been finished. |
1126 |
|
1127 |
Here is an example implementation that uses L<Coro::AIO> and largely |
1128 |
ignores error handling: |
1129 |
|
1130 |
my $ioserver = port_async { |
1131 |
while () { |
1132 |
my ($tag, $port, $path) = get_cond; |
1133 |
|
1134 |
$tag eq "write_file" |
1135 |
or die "only write_file messages expected"; |
1136 |
|
1137 |
my $fh = aio_open $path, O_WRONLY|O_CREAT, 0666 |
1138 |
or die "$path: $!"; |
1139 |
|
1140 |
while () { |
1141 |
my (undef, undef, $data) = get_cond { |
1142 |
$_[0] eq "data" && $_[1] eq $port |
1143 |
} 5 |
1144 |
or die "timeout waiting for data message from $port\n"; |
1145 |
|
1146 |
length $data or last; |
1147 |
|
1148 |
aio_write $fh, undef, undef, $data, 0; |
1149 |
}; |
1150 |
} |
1151 |
}; |
1152 |
|
1153 |
mon $ioserver, sub { |
1154 |
warn "ioserver was killed: @_\n"; |
1155 |
}; |
1156 |
|
1157 |
Let's go through it, section by section. |
1158 |
|
1159 |
my $ioserver = port_async { |
1160 |
|
1161 |
Ports can be created by attaching a thread to an existing port via |
1162 |
C<rcv_async>, or as in this example, by calling C<port_async> with the |
1163 |
code to execute as a thread. The C<async> component comes from the fact |
1164 |
that threads are created using the C<Coro::async> function. |
1165 |
|
1166 |
The thread runs in a normal port context (so C<$SELF> is set). In |
1167 |
addition, when the thread returns, it will be C<kil> I<normally>, i.e. |
1168 |
without a reason argument. |
1169 |
|
1170 |
while () { |
1171 |
my ($tag, $port, $path) = get_cond; |
1172 |
or die "only write_file messages expected"; |
1173 |
|
1174 |
The thread is supposed to serve many file writes, which is why it |
1175 |
executes in a loop. The first thing it does is fetch the next message, |
1176 |
using C<get_cond>, the "conditional message get". Without arguments, it |
1177 |
merely fetches the I<next> message from the queue, which I<must> be a |
1178 |
C<write_file> message. |
1179 |
|
1180 |
The message contains the C<$path> to the file, which is then created: |
1181 |
|
1182 |
my $fh = aio_open $path, O_WRONLY|O_CREAT, 0666 |
1183 |
or die "$path: $!"; |
1184 |
|
1185 |
Then we enter a loop again, to serve as many C<data> messages as |
1186 |
necessary: |
1187 |
|
1188 |
while () { |
1189 |
my (undef, undef, $data) = get_cond { |
1190 |
$_[0] eq "data" && $_[1] eq $port |
1191 |
} 5 |
1192 |
or die "timeout waiting for data message from $port\n"; |
1193 |
|
1194 |
This time, the condition is not empty, but instead a code block: similarly |
1195 |
to grep, the code block will be called with C<@_> set to each message in |
1196 |
the queue, and it has to return whether it wants to receive the message or |
1197 |
not. |
1198 |
|
1199 |
In this case we are interested in C<data> messages (C<< $_[0] eq "data" |
1200 |
>>), whose first element is the source port (C<< $_[1] eq $port >>). |
1201 |
|
1202 |
The condition must be this strict, as it is possible to receive both |
1203 |
C<write_file> messages and C<data> messages from other ports while we |
1204 |
handle the file writing. |
1205 |
|
1206 |
The lone C<5> argument at the end is a timeout - when no matching message |
1207 |
is received within C<5> seconds, we assume an error and C<die>. |
1208 |
|
1209 |
When an empty C<data> message is received we are done and can close the |
1210 |
file (which is done automatically as C<$fh> goes out of scope): |
1211 |
|
1212 |
length $data or last; |
1213 |
|
1214 |
Otherwise we need to write the data: |
1215 |
|
1216 |
aio_write $fh, undef, undef, $data, 0; |
1217 |
|
1218 |
And that's basically it. Note that every port thread should have some |
1219 |
kind of supervisor. In our case, the supervisor simply prints any error |
1220 |
message: |
1221 |
|
1222 |
mon $ioserver, sub { |
1223 |
warn "ioserver was killed: @_\n"; |
1224 |
}; |
1225 |
|
1226 |
Here is a usage example: |
1227 |
|
1228 |
port_async { |
1229 |
snd $ioserver, write_file => $SELF, "/tmp/unsafe"; |
1230 |
snd $ioserver, data => $SELF, "abc\n"; |
1231 |
snd $ioserver, data => $SELF, "def\n"; |
1232 |
snd $ioserver, data => $SELF; |
1233 |
}; |
1234 |
|
1235 |
The messages are sent without any flow control or acknowledgement (feel |
1236 |
free to improve). Also, the source port does not actually need to be a |
1237 |
port - any unique ID will do - but port identifiers happen to be a simple |
1238 |
source of network-wide unique IDs. |
1239 |
|
1240 |
Apart from C<get_cond> as seen above, there are other ways to receive |
1241 |
messages. The C<write_file> message above could also selectively be |
1242 |
received using a C<get> call: |
1243 |
|
1244 |
my ($port, $path) = get "write_file"; |
1245 |
|
1246 |
This is simpler, but when some other code part sends an unexpected message |
1247 |
to the C<$ioserver> it will stay in the queue forever. As a rule of thumb, |
1248 |
every threaded port should have a "fetch next message unconditionally" |
1249 |
somewhere, to avoid filling up the queue. |
1250 |
|
1251 |
Finally, it is also possible to use more switch-like C<get_conds>: |
1252 |
|
1253 |
get_cond { |
1254 |
$_[0] eq "msg1" and return sub { |
1255 |
my (undef, @msg1_data) = @_; |
1256 |
...; |
1257 |
}; |
1258 |
|
1259 |
$_[0] eq "msg2" and return sub { |
1260 |
my (undef, @msg2_data) = @_; |
1261 |
...; |
1262 |
}; |
1263 |
|
1264 |
die "unexpected message $_[0] received"; |
1265 |
}; |
1266 |
|
1267 |
=head1 THE END |
1268 |
|
1269 |
This is the end of this introduction, but hopefully not the end of |
1270 |
your career as AEMP user. I hope the tutorial was enough to make the |
1271 |
basic concepts clear. Keep in mind that distributed programming is not |
1272 |
completely trivial, in fact, it's pretty complicated. We hope AEMP makes |
1273 |
it simpler and will be useful to create exciting new applications. |
1274 |
|
1275 |
=head1 SEE ALSO |
1276 |
|
1277 |
L<AnyEvent::MP> |
1278 |
|
1279 |
L<AnyEvent::MP::Global> |
1280 |
|
1281 |
L<Coro::MP> |
1282 |
|
1283 |
L<AnyEvent> |
1284 |
|
1285 |
=head1 AUTHOR |
1286 |
|
1287 |
Robin Redeker <elmex@ta-sa.org> |
1288 |
Marc Lehmann <schmorp@schmorp.de> |
1289 |
|