Something like downing all members if there's no quorum? I was thinking
about that, but there are two problems:
- under the current cluster specification, is there a way to resurrect a
node? Wouldn't smarter downing require you to add a new state, something
like suspended (waiting for enough
On Tue, Jan 6, 2015 at 11:01 AM, Adam Warski a...@warski.org wrote:
Something like downing all members if there's no quorum?
Yes, something like that. From the larger healthy side of the partition
mark all unreachable as down, and from the other side shutdown yourself.
That will not work well
Ah, I did not understand that.
I have solved that by starting a new Akka system and a new JavaTestKit in
each test, and doing the test code in a static initializer. Before, I
switch to the in-mem journal using system properties.
In my abstract super-class:
*private *ActorSystem *system*;
Hi,
Notice the naming mismatch - FlowGraphBuilder builds FlowGraph's when
there's also the case class GraphFlow. Are the types (supposed to be)
working together? What's the reason behind the naming difference?
Just ran across it this morning while reviewing the source code and worth
to point
Hi Kane,
Exploring the different options, one limitation that I can see is that
ClusterPoolRouter requires the class of the actor that's going to be
remotely deployed to the cluster to be *present on the class path of the
router. *That is, if our front-ends are to create a worker on a
On Mon, Jan 5, 2015 at 9:18 PM, Pradeep Gollakota pradeep...@gmail.com
wrote:
Hi Patrik,
Thanks for the response. Is there a work around for this I can employ? Is
it possible to use a custom mailbox with the old balancing dispatcher
(from pre 2.3)?
Sorry, not that I know from the top of my
Hi Jacek,
The public API consist of FlowGraph and FlowGraphBuilder. GraphFlow is an
internal implementation detail.
Regards,
Patrik
PS. In my opinion we should rename GraphFlow to something else. It's rather
confusing for anyone reading the source code.
On Tue, Jan 6, 2015 at 12:04 PM, Jacek
I would guess it is a classpath ordering issue. To be sure you load the
right config from tests you can use another file name for the test config
name. In the tests you load that config and pass to the factory method of
the ActorSystem.
/Patrik
On Mon, Jan 5, 2015 at 1:45 PM, Anders Båtstrand
That's exactly what I needed.
Thanks Anders.
(I put it in my akka.conf)
On Tue, Jan 6, 2015 at 2:41 AM, Anders Båtstrand ander...@gmail.com wrote:
Ah, I did not understand that.
I have solved that by starting a new Akka system and a new JavaTestKit in
each test, and doing the test code in a
Hi,
It is not possible to simulate OversizedPayloadException in test using
local actor system. You have to start two remote/cluster actor systems and
send over the configured port.
The send-buffer-size and receive-buffer-size attempt to inform the OS how
much buffering it should do on our
Hi Muthu,
On Mon, Jan 5, 2015 at 7:00 AM, Muthukumaran Kothandaraman
muthu.kmk2...@gmail.com wrote:
I have following requirements for designing a clusterwide messaging system
1. There can be many to many relationship between topics and subscribers.
ie. one module (OSGi) in each cluster
Further studying convinces me that this is not yet implemented. I'm going
to pursue the ActorPublisher / ActorSubscriber route; thank you for this
documentation page
http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M2/scala/stream-integrations.html,
Akka-Stream folks! I'll post
Ok, if there's only one ball, that answers the question. Thanks! :)
Adam
On Thursday, January 1, 2015 10:00:27 AM UTC+1, Akka Team wrote:
Hi Adam,
PushPull stage callbacks are never concurrent. Also, if a stage calls
ctx.Push then it will receive eventually an onPull, and whenever it calls
Hello,
I've got this code:
def httpGet( uri:String )(implicit s:ActorSystem) = {
implicit val materializer = FlowMaterializer()
var r:HttpResponse = null
val req = HttpRequest(HttpMethods.GET, Uri(uri))
val host:String = req.uri.authority.host.toString
val port:Int = req.uri.effectivePort
val
We're also deploying an Akka cluster on CoreOS with Docker. We deploy the
same fat JAR for every node (in fact, the exact same docker image) and then
change their behavior by setting the role via environment variables.
In our case, each role has a bootstrap class which sets up whatever
See also -Dconfig.trace=loads (documented at
https://github.com/typesafehub/config#debugging-your-configuration )
That should tell you what's happening.
Havoc
On Tue, Jan 6, 2015 at 8:34 AM, Patrik Nordwall
patrik.nordw...@gmail.com wrote:
I would guess it is a classpath ordering issue. To be
Thanks Patrik.
By failure-zones, I mean to assign different dispatchers to different
combinations of topic + subscribers.
For example, Subscriber A and Subscriber B both handle messages from topic
C and hence there would be two subscribers - actor sub_a_topic_c and actor
sub_b_topic_c (I
I am modeling a data processing pipeline where there are several decision
points, where the system may refer to a human being to make a decision.
Since earlier decisions change the state of the system and affect later
decisions, another goal of mine is to process a narrow sliding view of the
18 matches
Mail list logo