Sorry, correction:
On Sun, Jun 7, 2015 at 12:27 PM, Justin du coeur jduco...@gmail.com wrote:
- Cluster Sharding -- built-in mechanisms to deal with spreading
heterogeneous Actors around the cluster, and letting clients communicate
with them without needing to know where they are
Interesting -- I haven't even thought about roles yet (for the moment,
Querki's architecture is pretty uniform), but that might also work.
Thanks -- at this point, it's clear that there are ways to deal with this
problem, and I just have to choose which one works best for my purposes...
On Sun,
On Sunday, June 7, 2015 at 6:38:51 PM UTC+3, drewhk wrote:
On Sun, Jun 7, 2015 at 5:32 PM, Andrey Kuznetsov fe...@loathing.in
javascript: wrote:
auth and session makes context.stop(self) inside Cancel handler of
their receive's.
umm, but they will not get a Cancel from merge if
Yeah, in Akka Cluster-Sharding you also have regions which I think you can
use to broadcast messages to actors living in such region, a region will
likely be spread between distinct nodes.
I use Akka Cluster roles because I have an admin node that keeps track of
each supervisor's actor and a
Hi Rowland,
In order to be able to remote deploy an actor, its Props must be
serializable. If you use constructor parameters, then this means the Actor
should be serializable, too. Instead of sending Connection, just can just
send a parameters object that allows the remote side to recreate the
From a pure Akka-user POV, I find the article not even so much slanted as a
bit blinkered -- it is focused on its approach and thinking, so of course
Quasar comes out looking like the right approach. (Context: I am just a
user, but I've been building a product centered on Akka for several years
No so automated but I do this with specific supervisor actors that are
subscribed to cluster events.
So I have broadcast routers that are kept updated, so when some event happens I
just use the router.
--
Read the docs: http://akka.io/docs/
Check the FAQ:
Cluster node has roles, and cluster events can keep your broadcast routers
updated, so you could broadcast a message to a role.
--
Read the docs: http://akka.io/docs/
Check the FAQ:
http://doc.akka.io/docs/akka/current/additional/faq.html
Search the archives:
Hi Endre,
Yes I think it will be very useful. Though I currently can easily send
messages in my akka cluster, I can't orchestrate a process that requires
multiple messages to be send via several actors and the receipts to be
collected. Doing so with just a series of messages becomes tedious
Hi,
I try but get error message:
java.lang.IllegalStateException: onNext is not allowed when the stream has
not requested elements, totalDemand was 0
My code is:
val jobManagerSource =
Source.actorPublisher[UserRegisterSource.RegisterUser](UserRegisterSource.props)
val ref =
Hi Guido,
You cannot import the release-2.3-dev in newest IntelliJ, I filed them a
bug last week. Can you still make this a proper PR? You can build Akka
using sbt only, or you can still use the old sbt-idea plugin to make a
project that IDEA can read (that is what I use now until they fix the
Hi Kostas,
Currently there is no such DSL, but this is actually something we already
started experimenting with. I hope that we can come up with something easy
to use in the near future.
-Endre
On Mon, Jun 1, 2015 at 5:09 PM, Kostas kougios
kostas.koug...@googlemail.com wrote:
Hi, I need to
Hi Leon,
The scan method of FJP will very likely be your top method unless you have
very favorable load patterns for FJP. I wouldn't worry about it, unless you
see a performance problem with your application.
-Endre
On Fri, Jun 5, 2015 at 10:54 PM, Justin Forder justin.for...@gmail.com
wrote:
Hi Jason,
On Mon, Jun 1, 2015 at 8:24 PM, Jason Martens m...@jasonmartens.com wrote:
Hello All,
I deal with a lot of large files on Amazon S3 and disk, which have been
indexed into chunks that can be used individually. What I would like to do
is have a Source of offsets and sizes which
Hi,
On Fri, Jun 5, 2015 at 7:53 PM, Владимир Морозов greenhos...@gmail.com
wrote:
Yes, I know about mapAsync, but my problem with Source. A want to use
single stream for processing some group of events
I am not sure if I understand your question properly, but if you want to
run the stream
Hi Andrey,
When do the auth and session sources finish? Do you see the onComplete log
message from broadcast? Since the connections you see are in CLOSE_WAIT,
that means that your server already received a close from the client, but
it has not yet closed its own side (half-closed). I suspect that
Hi Anders,
On Tue, Jun 2, 2015 at 3:15 PM, Anders Båtstrand ander...@gmail.com wrote:
I now encountered the problem again: The cluster (3 nodes) suddenly has
two leaders, and only one of the nodes reported all the other nodes to be
part of the cluster.
While it might have been triggered by
Hi Dana,
While Akka Http has really not yet optimized the numbers you have does not
match what we have seen in general. We are not very fast, but not *that*
slow. There is something fishy going on there. Unfortunately we have not
much time to look into large code, can you just create a very
Hi Eugene,
On Thu, Jun 4, 2015 at 1:25 AM, Eugene Dzhurinsky jdeve...@gmail.com
wrote:
Hello!
I wonder if there's something like a template to build pull task flow,
where
- defined some data source actor/actors
- data-source may produce either some data token, or EOD event (end of
data)
On Sun, Jun 7, 2015 at 5:32 PM, Andrey Kuznetsov f...@loathing.in wrote:
auth and session makes context.stop(self) inside Cancel handler of their
receive's.
umm, but they will not get a Cancel from merge if merge itself is not
cancelled, which it will not be in this case. So mtproto gets the
Hi Guido,
There is no overhead of calling Scala code from Java, whatever you can
access from Java, and looks like a method, then it is an ordinary JVM
method. In fact, sometimes calling the Scala API might be slightly faster,
since most of the time the Java API just calls the Scala API underneath
Hi Maatari,
On Wed, Jun 3, 2015 at 8:29 PM, Maatary Okouya maatarioko...@gmail.com
wrote:
Hi,
is there some more extensive doc and example on how to use akka http
client side.
Not yet, work on documentation is ongoing. We will have more docs in the
next RC release.
I'm switching to
Hi Richard,
A Flow is basically a thing with an input and output port. It does not
say though how those ports are connected internally: they can be unrelated
(like in your example), wired up to each other, or passing through a graph,
or piped through a TCP connection.
-Endre
On Sat, Jun 6, 2015
Hi,
On Thu, Jun 4, 2015 at 3:01 PM, Nweike Onwuyali nweikeonwuy...@gmail.com
wrote:
@Chanan.
if myActor.tell(Some Message or Some Class, ActorRef.noSender()) fails
due to network unavailability (could be due to other reasons), how do i
configure Akka to ensure it keeps trying to recover
So each Space actor would subscribe, and the admin monitor could send out
the request for updates when it needs it? That's plausible. I'll
investigate that approach -- thanks!
On Thu, Jun 4, 2015 at 8:54 AM, Anders Båtstrand ander...@gmail.com wrote:
Will the extension Distributed Pub Sub do
Hi,
Akka does not use TRACE level, only DEBUG, INFO, WARN and ERROR. If you use
akka logging facilities then there is no API to emit trace level (AFAIK).
-Endre
On Tue, Jun 2, 2015 at 12:58 AM, TS test.tester1...@gmail.com wrote:
I am using logback with slf4j and LoggingAdapter. Why is it
Hi Eax,
There is one pattern in contrib that does something similar to what you
need: http://doc.akka.io/docs/akka/2.3.11/contrib/aggregator.html
I recommend taking a look at that, and see if it what you want, or just
take the code and customize it as you want. In the future we plan to add
some
Hi,
This error message means that there were substreams that were not consumed
by anyone and therefore have been lingering around until that timeout
triggers. Can you pack up a small reproducer for this? Because if this is a
bug this can cause some weird behavior, as substreams constructed by
I am handling connections with the following code:
val connections = Tcp().bind(interface, port)
connections runForeach { conn ⇒
log.info(sClient connected from: ${conn.remoteAddress})
try {
val flow = MTProto.flow(nextConnId(), maxBufferSize, sessionRegion)
conn.handleWith(flow)
}
Hi Michael,
This is a good point, I think there should be a Directive that allows
matching on this, then user code can do the rest. Can you file a ticket for
this feature?
Thanks,
-Endre
On Tue, Jun 2, 2015 at 11:43 PM, Michael Hamrah mham...@gmail.com wrote:
I'd like to create a directive
Hi Biniam
Honestly, the Akka Team cannot answer these questions in a neutral and
unbiased way since we are all Akka developers. I recommend trying out these
frameworks and learn about them yourself. Experiences of other people might
match or not your own preferences, so it is always better to
Hi,
On Fri, Jun 5, 2015 at 10:20 AM, Giampaolo giampaolo.trapa...@gmail.com
wrote:
Many thanks Akka Team for the reply.
Do you think that this info should be added to TestKit doc to avoid this
mistake for Akka newbies like me?
Yes, I think so. Can you file a ticket for this?
-Endre
https://github.com/akka/akka/issues/17680
Thanks Endre.
--
Read the docs: http://akka.io/docs/
Check the FAQ:
http://doc.akka.io/docs/akka/current/additional/faq.html
Search the archives: https://groups.google.com/group/akka-user
---
You received this message because you
and mapResponse is a PushStage:
def mapResponse(system: ActorSystem) = new PushStage[MTProto, ByteString] {
private[this] var packageIndex: Int = -1
override def onPush(elem: MTProto, ctx: Context[ByteString]) = {
packageIndex += 1
val pkg = TransportPackage(packageIndex, elem)
If the entity is not a HttpEntity.Default, Akka can't provide a length
without buffering the stream first. Which is why one doesn't see the length
of the entity in the HttpRequest.
If there was Content-Length header but it was inaccurate, does that leave a
careless implementation around it
Hi Adam,
You cannot reuse the Builder instance, you can reuse the result from the
.build() call though. The builder has mutable state inside so it is not
meant to be reused.
-Endre
On Mon, Jun 1, 2015 at 11:45 AM, Adam Daines dainesa...@gmail.com wrote:
Hi,
I have been making use of the
Hi Peter,
Unfortunately this is one of the swampy areas where we have not good
support yet. This pattern is already implementable (Http client connection
pooling does something like this and more) but it is very error prone.
AsyncStage can work if you know what are you doing :) If you have
Hi Andrey,
I am not sure your problem is related to that ticket (
https://github.com/akka/akka/issues/17122). Do you have some code that
represents what you are trying to do? The TCP flows implement proper
half-close, so if it not used carefully it might end up being half-closed
instead of fully
auth and session makes context.stop(self) inside Cancel handler of their
receive's. And they see it if client closes connection (at least if it
is being closed by Connection reset by peer). I tried to send PoisonPill
to auth and session actors from Sink.onComplete and still observed
CLOSE_WAIT
Hi Akka Team,
I have a cluster sharding with 3 nodes (I, II, III). When the leader node
down (node I), after that the second node down (node II). The sharding
coordinator restart at another node (node III), it can not received the
terminated message of ShardRegion of node II. Can you tell me
I need to provide JMS topics for external non-Akka clients from inside of
my actor system. So far, I'm using the CamelExtension to do something like
..
val factory = new
ActiveMQConnectionFactory(vm:(broker:(tcp://localhost:61617)?persistent=false)
)
camelContext.addComponent(ext-jms,
I solve this problem, but found new one:
val ref = Flow[UserRegisterSource.RegisterUser]
.mapAsync(1)(callMe)
.to(Sink.ignore)
.runWith(jobManagerSource)
val t = (ref ask
UserRegisterSource.RegisterUser(test)).mapTo[UserRegisterSource.RegisterUser]
but t never complete, always end with Timeout
By HTTP (websocket) server I mean akka-http.
On Sunday, June 7, 2015 at 9:17:36 AM UTC+3, Andrey Kuznetsov wrote:
My application also has an HTTP (websocket) server on another port, it's
behind Elastic Load Balancer too, and it doesn't suffer from such problem -
there is no CLOSE_WAIT
My application also has an HTTP (websocket) server on another port, it's
behind Elastic Load Balancer too, and it doesn't suffer from such problem -
there is no CLOSE_WAIT connections to it.
On Sunday, June 7, 2015 at 7:17:44 AM UTC+3, Andrey Kuznetsov wrote:
Just found this issue:
44 matches
Mail list logo