thank you for share this:)
在 2016年3月9日星期三 UTC+8上午5:33:17，nkasvosve写道：
>
> So Lightbend is about to release Lagom ... we have already built a robhust
> framework based on Akka.
>
> http://beyondj.com
>
> BeyondJ is a JVM platform for deploying wars and fat jars. It provides,
> out of the box,
best wishes for all:)
--
>> 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
- *Solution A:* Migrate akka-remote from Netty 3 to 4 and gain 4 times
the performance with an effort equivalent to N hours.
- //this would not happen, and don't give you 4 times boost
- *Solution B:* Migrate akka-remote from Netty 3 to akka-streams and
gain 5 to 6 times the
yes,akka-stream based will give us more control,and I think that's is more
important than performance,because it means how stable akka-cluster will be.
在 2016年1月22日星期五 UTC+8下午9:23:03，Akka Team写道：
>
>
>
> On Fri, Jan 22, 2016 at 2:00 PM, 何品 <hepi...@gmail.com >
> wrote:
&
g like Vert.x, not sure what's going on here or why it has such
>> little importance when it is part of Akka selling speech: "Location
>> transparency" or is it only an advertisement but not used much in practice
>> because of its performance?
>>
>> I thin
congratulations,big time.
--
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 are subscribed to the Google
.x to 4.1.x
I'm sure you are doing that for a good reason, Netty guys tend improve by
a considerable factor the performance between versions.
Best regards,
Guido.
On Monday, June 15, 2015 at 2:02:59 PM UTC+1, 何品 wrote:
Hi,the code on github is what we used in the last production
plus,I think it would be great that i could create a dispatcher from the
eventloop,and using it for the actor creating,but currently it's impossible.
and for the flow control and thread usage,you could reuse the event loop
,control the register and deregister,control the read and write and
you need a frame based codec.
LV,L-Length,V-Value
在 2015年6月16日星期二 UTC+8下午7:17:26，Dmitrii Galagaev写道：
I am using TCP(http://doc.akka.io/docs/akka/snapshot/java/io-tcp.html)
actor as entry point to my java application:
public SupervisorActor() {
receive(ReceiveBuilder.
I don't have a dedicated dispatcher for the proxy actor,and for the server
and client side pools,in fact it's event loop group,a more nice one is
using reflection,create it via provider,then you have all the control ,but
I think the default one is good,and the code there,is not that good,more
as possible. It might be possible in August, but I can't
promise anything.
/Patrik
On Mon, Jun 15, 2015 at 7:12 AM, 何品 hepi...@gmail.com wrote:
I have build a new one via io extension,and maybe opensource when it is
more complete.
the code there is deprecated.
and I will try to build the io
I have build a new one via io extension,and maybe opensource when it is
more complete.
the code there is deprecated.
and I will try to build the io extension via akka stream too.
```
import java.net.InetSocketAddress
import akka.actor.{ Actor, ActorLogging, Props }
import akka.io.IO
import
We are using log4j2.
在 2015年2月23日星期一 UTC+8下午7:40:04，Martin Ford写道：
Hi,
In the Akka documentation
http://doc.akka.io/docs/akka/current/java/logging.html#logging-java
Typesafe recommend using Logback as the runtime SLF4J backend logger. I'm
just wondering what the reasons are for that
have add it to https://github.com/hepin1989/awesome-akka
nice to see it,I am writing akka-io-netty
在 2015年5月15日星期五 UTC+8上午11:27:33，Artur Opala写道：
Anyone interested in testing/using/developing Akka I/O driver for SCTP
protocol?
Just started working on this:
update to new version now.thanks
--
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 are subscribed to the
two link for the PDF
http://pages.cs.wisc.edu/~akella/CS838/F12/838-CloudPapers/FlumeJava.pdf
http://static.googleusercontent.com/media/research.google.com/zh-CN/us/pubs/archive/41378.pdf
--
Read the docs: http://akka.io/docs/
Check the FAQ:
you could just pause the current thread.not all of them.cursive debugger
could do that.
在 2015年5月7日星期四 UTC+8上午10:50:41，Jeff写道：
Every time I set a breakpoint in an actor involved with Actor Clustering,
the phi detector triggers and marks that node as Unreachable. Is there a
way to prevent
actor are object,just more complex than a simple runnable,and the actors
are running on threads(dispatchers),
just like the executeService runs runnables.
so it's not actors create threads.it's actors run on threads.
在 2015年4月3日星期五 UTC+8上午11:59:10，Jitendra写道：
Hi All,
I am new to akka. I
1,send some protocol level message to indicate that all file have been
handed ,like NoMoreWork,then the actor case at it,and then stop self via
context.stop(self)
2,send PoisonPill
3,tick a time out to self
I think the first one is OK
在 2015年4月2日星期四 UTC+8上午2:38:05，Syd Gillani写道：
Hi,
I have
. Thanks a lot hepin!
now just waiting on 2.3.10 ;)
On Tuesday, March 24, 2015 at 4:03:29 AM UTC+2, 何品 wrote:
https://github.com/akka/akka/pull/16669
在 2015年3月20日星期五 UTC+8下午11:03:07，Johannes Berg写道：
Hi!
I've been working with some monitoring of the cluster through the
ClusterMetrics
could we split akka-stream and akka-http as two separate release cycles?
在 2015年3月28日星期六 UTC+8上午5:05:12，rkuhn写道：
Dear hakkers,
it has been four weeks since the last milestone was published, so it is
high time for the next one. The user-visible changes we have been doing are
getting
it easy to using future to do something like MR work or simple computing.
I think actor is for some more complex work .
In practice you will mix using future/promise and actors in the same
application
在 2015年3月27日星期五 UTC+8上午12:57:03，akhil写道：
Hi,
I am trying to figure out the difference in
https://github.com/akka/akka/pull/16669
在 2015年3月20日星期五 UTC+8下午11:03:07，Johannes Berg写道：
Hi!
I've been working with some monitoring of the cluster through the
ClusterMetrics provided in Akka 2.3.9 using hyperic-sigar-1.6.4. I've been
comparing what is displayed in top and what I get by
have been fixed
在 2015年3月20日星期五 UTC+8下午11:03:07，Johannes Berg写道：
Hi!
I've been working with some monitoring of the cluster through the
ClusterMetrics provided in Akka 2.3.9 using hyperic-sigar-1.6.4. I've been
comparing what is displayed in top and what I get by listening to
the docker one?
在 2015年1月12日星期一 UTC+8下午4:01:35，Andrew James Ramirez写道：
Hi,
I just recently updated to Scala 2.11 and I need some features from
2.4-SNAPSHOT
but there is no _2.11-2.4-SNAPSHOT in the repository.
Specifically this feature:
https://github.com/akka/akka/pull/15610/files
means we may would want to see how busy between nodes.the message is send
to remote actor via remote transport.
在 2014年12月23日星期二UTC+8下午5时26分52秒，Akka Team写道：
Hi,
On Mon, Dec 22, 2014 at 7:12 PM, 何品 hepi...@gmail.com javascript:
wrote:
should cluster matrix add node to node throughput
very very nice work and words
在 2014年12月23日星期二UTC+8下午8时26分24秒，Akka Team写道：
*Dear Hakkers,*
We are happy to announce that the second milestone of the still brewing
Akka Streams and Akka Http modules is now available. This milestone
contains a new Streams and Http manual that smooths the
First,No.
In fact I think the problem is cause by leveldb.
the size is about 6.7GB,and I could not dump it to cassandra or so on.
在 2014年12月22日星期一UTC+8下午5时06分45秒，Akka Team写道：
Hi,
On Sun, Dec 21, 2014 at 6:54 PM, 何品 hepi...@gmail.com javascript:
wrote:
yes.only one process,in fact one
should cluster matrix add node to node throughput or network interface
throughput,so we could see how busy between nodes
--
Read the docs: http://akka.io/docs/
Check the FAQ:
http://doc.akka.io/docs/akka/current/additional/faq.html
Search the archives:
to make sure that there is only one process that is accessing
LevelDB storage files. Is that the case?
On Sat, Dec 13, 2014 at 11:33 AM, 何品 hepi...@gmail.com javascript:
wrote:
I am using akka persistent with netty to handle the statistic of game
server.
There are 1 to N statistic
I am using akka persistent with netty to handle the statistic of game
server.
There are 1 to N statistic client (with akka persistent ,ALOD netty ) and
1 server which receive the incoming protobuf encoded statistic data and
flush to database for later use.
I have some issue at the first
really remove parts
of a flow.
You could probably model this as a stream that does the
filtering/authorization that will produce a new stream that is the
filtered/authorized content if that is what you want.
B/
On 5 December 2014 at 13:22:54, 何品 (hepi...@gmail.com javascript:)
wrote
with actorOf(). In this case the name of the new
incarnation will be the same as the previous one but the UIDs will differ.
I hope someone could confirm this. I am gonna try making a test.
On Friday, December 5, 2014 8:15:05 PM UTC+8, 何品 wrote:
after restart the per actorref is still valid
First here
https://github.com/akka/akka/issues/15996
then
When we starting to kick out all ```await``` ,and make all things
```async``` ,and quickly going to the ```map``flatMap```
hole,especially badly on the java side(we have a wrapper for scala's future
/promise).I starting to think
yes,I am doing what just as you said.
I am stash what I received and flush it every 10 seconds
在 2014年12月3日星期三UTC+8上午6时07分48秒，Raymond Lau写道：
Let's assume that I'm running a social network where the amount of
peoples' likes coming in at a given time is very high (ex. a thousand likes
per
I only want to say 。
在 2014年11月27日星期四UTC+8下午1时48分20秒，Suman Adak写道：
Thanks to all,
Problem got solved.
mistake was transport = akka.remote.netty.NettyRemoteTransport
it should be akka.remote.netty.NettyTransport
Thanks a lot!!
On Tuesday, November 25, 2014 12:22:36 PM
Hi ,I just implement a remote transport via netty4.
the performance is good.
there is no ssl now,and no udp too.only tcp.
could more configuration is in the reference.conf
and not add docker support too
before
== It took 11515 ms to deliver 50 messages, throughtput 43421 msg/s, max
By the way ,we are using it.
在 2014年11月27日星期四UTC+8下午4时11分19秒，何品写道：
Hi ,I just implement a remote transport via netty4.
the performance is good.
there is no ssl now,and no udp too.only tcp.
could more configuration is in the reference.conf
and not add docker support too
before
== It took
https://github.com/hepin1989/akka-remote-transport-netty4
在 2014年11月27日星期四UTC+8下午4时11分19秒，何品写道：
Hi ,I just implement a remote transport via netty4.
the performance is good.
there is no ssl now,and no udp too.only tcp.
could more configuration is in the reference.conf
and not add docker
)
to make it easy for others to try it out?
Please add a section on the Community Projects page:
http://akka.io/community/
Regards,
Patrik
On Thu, Nov 27, 2014 at 9:40 AM, 何品 hepi...@gmail.com javascript:
wrote:
https://github.com/hepin1989/akka-remote-transport-netty4
在 2014年11月27日星期四UTC+8下午
So you want to set the akka's scheduler's base/start/`now` time?
在 2014年11月22日星期六UTC+8上午5时57分40秒，Bilal Günay写道：
Hi,
i am testing my application but i need to modify actual date. I am using
in whole application my custom now() method instead of
System.currentTimeMillis(), so i can test
I am asking will ,not how about the current status:P
在 2014年11月14日星期五UTC+8下午4时03分18秒，planetenkiller写道：
akka-io uses NIO (netty uses NIO too I think), see documentation here:
http://doc.akka.io/docs/akka/2.3.7/dev/io-layer.html#io-layer
On Friday, November 14, 2014 8:31:41 AM UTC+1, 何品 wrote
8:31:41 AM UTC+1, 何品 wrote:
will akka IO implement the AIO?
I found some snips
```
To reiterate we saw significant decrease in CPU Usage when using AIO
instead of NIO under high load. In our tests we saw CPU usage decrease from
75% to 15%. Many of our customers have confirmed
infact I have just implement one.
but I think you could find some nice thing here
https://github.com/netty/netty/blob/master/handler/src/main/java/io/netty/handler/ipfilter/AbstractRemoteAddressFilter.java
在 2014年11月14日星期五UTC+8下午4时27分22秒，JasonQ写道：
Hello,
I'm wondering whether AKKA has some
Will akka io using native implement like netty?
seems netty4's epoll implementation gains some performance on linux.
--
Read the docs: http://akka.io/docs/
Check the FAQ:
http://doc.akka.io/docs/akka/current/additional/faq.html
Search the archives:
, no netty for akka-io. But akka remoting uses it. Anyone up for a
swap and benchmark? :)
On Thu, Nov 13, 2014 at 11:55 AM, 何品 hepi...@gmail.com javascript:
wrote:
Will akka io using native implement like netty?
seems netty4's epoll implementation gains some performance on linux
will akka IO implement the AIO?
I found some snips
```
To reiterate we saw significant decrease in CPU Usage when using AIO
instead of NIO under high load. In our tests we saw CPU usage decrease from
75% to 15%. Many of our customers have confirmed this and their feedback is
that with AIO, the
DistributedPubSubExtension?
On Sun, Nov 9, 2014 at 7:45 AM, 何品 hepi...@gmail.com javascript:
wrote:
Currently we could implement this via DistributedPubSubExtension,and do
you guys think this is useful?
I think this is something like the vertx + vertx's eventbus.when using
The issue seems that you want using the AtLeastOnceDelivery Only,but not
the PersistentActor right?
I think it should be handy that only minxin the AtLeastOnceDelivery to the
normal actor,I haven't try that.
在 2014年11月9日星期日UTC+8下午7时59分41秒，Burak Emre Kabakcı写道：
I use persistent actors in order
Currently we could implement this via DistributedPubSubExtension,and do you
guys think this is useful?
I think this is something like the vertx + vertx's eventbus.when using it
with the Distributed Event Bus
http://vertx.io/core_manual_java.html#event-bus
--
Read the docs:
Hi,the code blow cause the StackOverFlow error,but if I want to do
something useful in an trait and let the actor who want to use mixin it,and
that trait need to do some initialize witch need to hook the actor's
lifecycle ,then what the best practice?
The problem of the code blow is that it
lifecycle hook,2
provide some trait level lifecycle hook for the user to call.
在 2014年11月9日星期日UTC+8下午3时15分30秒，何品写道：
Hi,the code blow cause the StackOverFlow error,but if I want to do
something useful in an trait and let the actor who want to use mixin it,and
that trait need to do some initialize
you guys may want a check
https://www.zenhub.io/
--
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 are
Thanks for you guys,I just encounter this problem,nice and clear.
在 2014年5月21日星期三UTC+8下午5时16分10秒，何品写道：
Hi could it be right to using Promise Inside an actor.
Actor received an message,and handle it to some other service ,which
return an promiseT,
and we chain an callback on the promiseT
I defined some message with protobuffer and using them for message,In java
,no case class,and hand writing nearly kill me
在 2014年5月20日星期二UTC+8下午10时38分38秒，Heiko Seeberger写道：
Akka users,
How do those of you who are using Akka with Java write immutable message
classes with collections?
I am using the persistent channel as a durable queue,and for it was not
recommend ,I switch to redis .
在 2014年5月21日星期三UTC+8上午2时32分38秒，rkuhn写道：
Hi Alex,
I have filed the ticket for Processor’s removal (
https://github.com/akka/akka/issues/15230), which also talks about
Channel since that
and ActorSystem, then I would use the EventBus for
broadcasting to all the ConnectionActor instances.
Can you let the clients keep track of their last received message number?
Then a ConnectionActor can be a plain Actor instead of an
EventsourcedProcessor.
B/
On 15 May 2014 at 18:49:12, 何品 (hepi
:17, 何品 (hepi...@gmail.com javascript:) wrote:
Hi,currently there is no war to retrieve one specified message from the
processor, the only way i think is using an hashmap to keep it on Recover
right?
and the EP's persist(message ,callback) ,I think the callback should
provide
the journal as a database. If you
want to use event sourcing to model the changes on the client, then I would
use a View to populate a query model (maybe in a SQL database) that fits
your queries.
B/
On 14 May 2014 at 13:15:51, 何品 (hepi...@gmail.com javascript:) wrote:
I want an actor which
Hi,currently there is no war to retrieve one specified message from the
processor, the only way i think is using an hashmap to keep it on Recover
right?
and the EP's persist(message ,callback) ,I think the callback should
provide the persistence message's seqNr too,currently I need to use the
yes, I am rewriting some code now,I am using play2.3 rc1 with akka 2.3.2
cluster and persistence,current ,6000~ qps now.
thanks roland.
在 2014年5月10日星期六UTC+8下午6时45分50秒，Akka Team写道：
On Sat, May 10, 2014 at 7:37 AM, 何品 hepi...@gmail.com javascript:wrote:
yes now I just using
than
updates).
Using a normal actor which creates children of known names sounds like a
good plan to me, the complexity of `context.child(name)` is O(log n) (it
uses a TreeMap).
Regards,
Roland
On Wed, May 7, 2014 at 4:26 AM, 何品 hepi...@gmail.com javascript:wrote:
Hi I was trying
Hi ,I have done this with right an Play plugin and api implement the api in
the plugin,
the implement just forward the request to the inner actor of the plugin.
在 2014年5月8日星期四UTC+8下午5时59分01秒，faisal...@gmail.com写道：
Hello all,
I am working on a fairly complicated system, and using Akka to
Hi I was trying to implement an router which could router a message to an
specified routee via the message's key.just as the ConsistentHashRouter.
when I look at the selected method ,the comes out routees are in an
immutable seq,but not a mapString,Routee,so if I want to selected one of
the
that what I just want.
I am still wondering should there be an deliveryConfirmedListener just
as redeliverFailureListener.by this way,I will make my life easier.
I don't think the application-level replay should be a answer for that
replay could be lost and there have one internal ACK in the
I am going to do something just like you,and I plan to implement an Router
and register the connectionActor as routee.
What do you think?
在 2014年4月9日星期三UTC+8下午7时24分25秒，Chanan Braunstein写道：
My use case:
I have an Actor per websocket called a User.
I have groups of user that need to be
66 matches
Mail list logo
