Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 17:26
    akka-ci commented #28338
  • 17:26
    akka-ci unlabeled #28338
  • 17:26
    akka-ci labeled #28338
  • 16:56
    akka-ci unlabeled #28338
  • 16:56
    akka-ci labeled #28338
  • 16:55
    raboof synchronize #28338
  • 16:55

    raboof on documentSourceCompletionStage

    Fix filename (compare)

  • 16:41
    akka-ci commented #28338
  • 16:41
    akka-ci unlabeled #28338
  • 16:41
    akka-ci labeled #28338
  • 16:38
    raboof commented #28328
  • 16:14
    raboof edited #25468
  • 16:11
    akka-ci labeled #28338
  • 16:11

    raboof on documentSourceCompletionStage

    Add source.completionStage exam… (compare)

  • 16:11
    raboof opened #28338
  • 16:07
    TisonKun commented #28328
  • 14:55
    akka-ci commented #28337
  • 14:55
    akka-ci unlabeled #28337
  • 14:55
    akka-ci labeled #28337
  • 14:54
    dataoSun starred akka/akka
mahengyang
@mahengyang
I have read Event Stream, but it was about classic akka
Arnout Engelen
@raboof
@mahengyang indeed we don't have a specific typed API for the eventstream yet, but it is fine to use the new and the classic API's together
Patrik Nordwall
@patriknw
@mahengyang @raboof There is eventStream API for Typed. Accessed via eventStream ActorRef in akka.actor.typed.ActorSystem. See apidoc of akka.actor.typed.eventstream.EventStream.
It might be missing from reference docs. I'll add a ticket for that.
Arnout Engelen
@raboof
oh :smile:
Patrik Nordwall
@patriknw
Renārs Kudiņš
@retriku
Hi! Is there an estimate of when akka 2.5.27 could be released? https://github.com/akka/akka/milestone/153
Sean Kwak
@cosmir17

I have a question. Why 'akka-discovery-aws-api' doesn't require 'akka.remote.artery' configs but 'akka.discovery' does? Can I ask where the omitting part happens in 'https://github.com/akka/akka-management'?

akka-discovery-aws-api works with just

  discovery {    
   method = aws-api-ec2-tag-based
    aws-api-ec2-tag-based {
      filters = "tag:mycomponent=my service name"
    }}
  management {
    cluster.bootstrap {
      contact-point-discovery {
        service-name = "my service name"
        discovery-method = aws-api-ec2-tag-based
      }  } }

https://mvnrepository.com/artifact/com.lightbend.akka.discovery/akka-discovery-aws-api_2.12/1.0.5

Sean Kwak
@cosmir17
because I see that 'akka-discovery-aws-api' uses 'akka.discovery' under the hood (as a dependency).
Johan Andrén
@johanandren
@retriku we’ll try to get that out in the coming weeks
Brendan McAdams
@bwmcadams
@patriknw that typed EventStream is only in 2.6.x, right? (not seeing it in 2.5 which I’m stuck on for time being)
Arnout Engelen
@raboof
Anand Singh
@Anand_Singh20_twitter
Hi , I'm facing issue in akka cluster
[backend-cluster-ingestion-7fbcb5497f-qnkw7] 15:49:30.233 [backend-cluster-akka.actor.default-dispatcher-76] WARN akka.cluster.sharding.ShardRegion - scheduler-27: Trying to register to coordinator at [ActorSelection[Anchor(akka.tcp://backend-cluster@10.2.136.10:2551/), Path(/system/sharding/scheduler-27-Coordinator/singleton/coordinator)]], but no acknowledgement. Total [30750] buffered messages. [Coordinator [Member(address = akka.tcp://backend-cluster@10.2.136.10:2551, status = Up)] is reachable.]
Need help
Arnout Engelen
@raboof
@Anand_Singh20_twitter are you starting sharding on all nodes? specifically, did it successfully start on the machine with address akka.tcp://backend-cluster@10.2.136.10:2551 ?
Anand Singh
@Anand_Singh20_twitter
@raboof Thanks to address, Yes, I'm facing this when deploying kubernetes, initially the cluster is start and running fine. But when some spot goes down, then it unable to form cluster
cluster setting is as follow:
cluster {
seed-nodes = []
seed-nodes = ${?SEED_NODES}
roles = ["seed", "worker", "S3Sharding", "GCSSharding"]
role {
seed.min-nr-of-members = 1
}
sharding.state-store-mode = ddata
sharding.remember-entities = on
sharding.distributed-data.durable.keys = []
sharding.passivate-idle-entity-after = off
downing-provider-class = "tanukki.akka.cluster.autodown.MajorityLeaderAutoDowning"
use-dispatcher = akka.cluster-dispatcher
failure-detector {
heartbeat-interval = 1 s
threshold = 15.0
min-std-deviation = 200 ms
acceptable-heartbeat-pause = 300 s
expected-response-after = 300 s
monitored-by-nr-of-members = 3
}
}
on role S3Sharding i'm getting issue
Anand Singh
@Anand_Singh20_twitter
ERROR akka.cluster.sharding.DDataShardCoordinator - The ShardCoordinator was unable to update a distributed state within 'updating-state-timeout': 5000 millis (retrying). Perhaps the ShardRegion has not started on all active nodes yet?
Arnout Engelen
@raboof
do the nodes that go down leave gracefully or just go unreachable?
Sean Kwak
@cosmir17
2.6.1 is out :tada: https://akka.io/blog/news/2019/12/06/akka-2.6.1-released
Thank you for releasing a new version !
Anand Singh
@Anand_Singh20_twitter

do the nodes that go down leave gracefully or just go unreachable?

Nope, it up and running

Anand Singh
@Anand_Singh20_twitter

I will try to explain the scenario, when application start on one port (say 2551) it run successfully. But when new port (say 2552) join the cluster then cluster with role, start failing and showing up with this

[backend-cluster-ingestion-7fbcb5497f-qnkw7] 15:49:30.233 [backend-cluster-akka.actor.default-dispatcher-76] WARN akka.cluster.sharding.ShardRegion - scheduler-27: Trying to register to coordinator at [ActorSelection[Anchor(akka.tcp://backend-cluster@10.2.136.10:2551/), Path(/system/sharding/scheduler-27-Coordinator/singleton/coordinator)]], but no acknowledgement. Total [30750] buffered messages. [Coordinator [Member(address = akka.tcp://backend-cluster@10.2.136.10:2551, status = Up)] is reachable.]

Hope this will help to understand the problem @raboof
Patrik Nordwall
@patriknw
@bwmcadams EventStream is probably only in 2.6. Underneath it’s the same eventStream so with adapters it should be possible to subscribe anyway, even though the api isn’t perfect
Brendan McAdams
@bwmcadams
fair enough. Had a few thoughts it might be useful to have eventstream for, I can deal with the slightly non-typed API for it I think. thanks.
Patrik Nordwall
@patriknw
@bwmcadams i’m interested in the “message dissapears” if sent directly after spawn. That’s not expected. Could you open an issue with more description?
Brendan McAdams
@bwmcadams
ah, so that one turned out to be as far as I can tell “I unintentionally mixed AbstractBehavior and Functional-style code"
Patrik Nordwall
@patriknw
Ok
Brendan McAdams
@bwmcadams
because the docs (at least for 2.5.x) often gloss over the OO-style syntax I made some mistakes. Removing the AbstractBehavior trait from my class fixed things (and I was able to organize the behaviors in a non-static definition as I wanted to, anyway)
Patrik Nordwall
@patriknw
Yeah, u should read 2.6 docs even though some api changed. Much improved.
Sven Ludwig
@sourcekick
https://twitter.com/ktosopl/status/1202604960111706112 this made my day in the sense that my most beloved technology keeps thriving :-)
Sean Kwak
@cosmir17

Hi, I have been using 'akka distributed' data with 'akka-discovery-aws-api' (akka 2.5.23 & management 1.0.3). It has been running fine in production with minimum setting. I upgraded to akka 2.6.0. It seems I need to specify akka.remote.artery {enabled, port} explicitly. I then encountered the following error. I have no idea what to do.

Caused by: com.typesafe.config.ConfigException$Missing: application.test.conf @ jar:file:/usr/lib/..........jar!/application.test.conf: 70: No configuration setting found for key 'large-message-destinations''

I guess I now need to specify 'large-message-destinations'. I don't understand why I need to manually set these.
I read 'https://doc.akka.io/docs/akka/current/remoting-artery.html#dedicated-subchannel-for-large-messages' but I have no idea of granualty of this setting (given that I use akka-distributed-data)
I expect that 'akka-management-clusterbootstrap' & 'akka-discovery-aws-api' would handle such settings automatically. It doesn't look like it does.
Any advice?

nikhilaroratgo
@nikhilaroratgo
@ihostage Hi,I need your help.I created a graphstage which throws Timeout from it and this stops the graph to which it is attached.The graph materializer has supervisor and decider but it seems TimeOut exception never reaches the supervisor.Please suggest how do I deal with this problem
DouglasGray
@DouglasGray
Hello, I've another question. If I have a supervised actor and it fails while in state X, if it restarts in state Y should any messages it was sent in state X that haven't been processed by state X's Behaviors. receiveMessage be received in state Y's Behaviors.receiveMessage?
Patrik Nordwall
@patriknw
@DouglasGray the actor has a mailbox (queue) where pending messages stay until they are processed one-by-one. The mailbox survive restarts so after restart it will continue processing pending messages.
DouglasGray
@DouglasGray
@patriknw Hi Patrik, thanks for the reply. I must be getting something wrong because I'm seeing what appear to be dropped messages on restart. I've put together a little example at https://scalafiddle.io/sf/OkAKenX/0 with some comments explaining what I was trying to do, i.e on restart sending the actor a message with its previous state, and in the meantime stashing further incoming messages. Although all the messages are stashed, if I fail halfway through processing the unstashed messages, the second half don't seem to be received.
Patrik Nordwall
@patriknw
Can't see any code there. The stashBuffer isn't part of the mailbox. It should be created outside of the setup if it should survive the restart. https://doc.akka.io/docs/akka/current/typed/stash.html
and maybe unstashAll is needed after the restart
Aleksandar Skrbic
@aleksandarskrbic
Hello, I have a problem with KafkaConsumerActor when asking for EndingOffset, I'm getting offset that is not correct. Does anyone have similar problem?
DouglasGray
@DouglasGray
@patriknw Ah indeed, you are correct - I should have defined my stash outside of the state the actor restarts in...apologies, I should have consulted the docs first. Thanks for the help!
Patrik Nordwall
@patriknw
Np, glad you found it
Khagay Nagdimov
@KhagayN
hi all, looking for a way to extract the sub path directory into a variable. For e.g. when a user hits /metadata/cars, saving that cars string into a variable somehow. Atm I have path("metadata") { get( etc...) } which only responds to /metadata calls
Vasyl Osadchyi
@kbks
Hi) Does anyone runs akka clusters with nomad?
Arnout Engelen
@raboof
@KhagayN like path("metadata" / Segment) { segment => get( etc...) }?
we need some nicer introduction and examples at https://doc.akka.io/docs/akka-http/current/routing-dsl/path-matchers.html :)
Khagay Nagdimov
@KhagayN
@raboof that works only if /metadata/cars is given. Can I extend this to work with /metadata as well