Tecsisa/constructr-consul

This library enables to use Consul as cluster coordinator in a ConstructR based node

Arnout Engelen

@raboof

@gerson24 hi!

@gerson24 my trigger to add this logging is for the debug logging: to get some more background on when and how often Consul returns a 404 response when renewing a session

but the akka streams logger also takes care of logging errors, and I'm not sure we get notified of those in any other way when we don't add this logging, so that might be an additional advantage

Gerson Pozo

@gerson24

thanks! I also think it's a good idea to print this logs for better debugging. Let's merge it

Arnout Engelen

@raboof

This message was deleted

Gerson Pozo

@gerson24

@raboof
We have upgraded to version 0.2.0

Arnout Engelen

@raboof

That's great thanks!

charles adetiloye

@charlesa101

hi @gerson24: do you have any example of this implementation ?

Gerson Pozo

@gerson24

hi @charlesa101 , We are publishing a demo(so you can try it for real in your laptop) next week

charles adetiloye

@charlesa101

cool thanks @gerson24

so i we already using this in our dev….but this is the problem i am having right now

i start up all the docker nodes…to cluster but only the first docker connects

to consul others wont

This message was deleted

[ERROR] [05/26/2016 03:58:48.007] [TestCluster-akka.actor.default-dispatcher-6] [akka.tcp://TestCluster@172.18.0.6:2553/system/constructr/constructr-machine] Timeout in Joining!
[ERROR] [05/26/2016 03:58:48.012] [TestCluster-akka.actor.default-dispatcher-17] [akka.tcp://TestCluster@172.18.0.6:2553/system/constructr] Terminating the system, because constructr-machine has terminated!
[INFO] [05/26/2016 03:58:48.055] [TestCluster-akka.remote.default-remote-dispatcher-20] [akka.tcp://TestCluster@172.18.0.6:2553/system/remoting-terminator] Shutting down remote daemon.
[INFO] [05/26/2016 03:58:48.056] [TestCluster-akka.remote.default-remote-dispatcher-20] [akka.tcp://TestCluster@172.18.0.6:2553/system/remoting-terminator] Remote daemon shut down; proceeding with flushing remote transports.
[INFO] [05/26/2016 03:58:48.072] [TestCluster-akka.actor.default-dispatcher-18] [akka://TestCluster/deadLetters] Message [akka.cluster.InternalClusterAction$InitJoin$] from Actor[akka://TestCluster/system/cluster/core/daemon/joinSeedNodeProcess-1#-1175653885] to Actor[akka://TestCluster/deadLetters] was not delivered. [1] dead letters encountered. This logging can be turned off or adjusted with configuration settings 'akka.log-dead-letters' and 'akka.log-dead-letters-during-shutdown'.
[INFO] [05/26/2016 03:58:48.072] [TestCluster-akka.actor.default-dispatcher-18] [akka://TestCluster/deadLetters] Message [akka.cluster.InternalClusterAction$InitJoin$] from Actor[akka://TestCluster/system/cluster/core/daemon/joinSeedNodeProcess-1#-1175653885] to Actor[akka://TestCluster/deadLetters] was not delivered. [2] dead letters encountered. This logging can be turned off or adjusted with configuration settings 'akka.log-dead-letters' and 'akka.log-dead-letters-during-shutdown'.
[INFO] [05/26/2016 03:58:48.072] [TestCluster-akka.actor.default-dispatcher-18] [akka://TestCluster/deadLetters] Message [akka.cluster.InternalClusterAction$InitJoin$] from Actor[akka://TestCluster/system/cluster/core/daemon/joinSeedNodeProcess-1#-1175653885] to Actor[akka://TestCluster/deadLetters] was not delivered. [3] dead letters encountered. This logging can be turned off or adjusted with configuration settings 'akka.log-dead-letters' and 'akka.log-dead-letters-during-shutdown'.
[INFO] [05/26/2016 03:58:48.079] [TestCluster-akka.actor.default-dispatcher-18] [akka://TestCluster/system/endpointManager/reliableEndpointWriter-akka.tcp%3A%2F%2FTestCluster%40172.18.0.5%3A2553-0/endpointWriter] Message [akka.actor.Status$Failure] from Actor[akka://TestCluster/system/endpointManager/reliableEndpointWriter-akka.tcp%3A%2F%2FTestCluster%40172.18.0.5%3A2553-0/endpointWriter#913709218] to Actor[akka://TestCluster/system/endpointManager/reliableEndpointWriter-akka.tcp%3A%2F%2FTestCluster%40172.18.0.5%3A2553-0/endpointWriter#913709218] was not delivered. [4] dead letters encountered. This logging can be turned off or adjusted with configuration settings 'akka.log-dead-letters' and 'akka.log-dead-letters-during-shutdown'.

Arnout Engelen

@raboof

@charlesa101 hmm we'd need some more logging to know for sure

@charlesa101 did you configure this cluster successfully without constructr before?

if not, did you see http://doc.akka.io/docs/akka/snapshot/scala/remoting.html#Akka_behind_NAT_or_in_a_Docker_container ?

Sebastian Gavril

@sebigavril

hello guys!

hseeberger/constructr#83 has removed the possibility of setting a custom SupervisorStrategy for Constructr

any specific reason for this change?

any way around it, to still provide a SupervisorStrategy?

A. Alonso Dominguez

@alonsodomin

hi guys, I just noticed that your current version does not support Consul HTTP tokens, required for ACL checks in the Consul side

I’m working on a PR including that, if you’d like to have it

now, if someone could give me some light on how to change the project’s version to something like 0.6.1-SNAPSHOT, so I can upload into our corp repository it’d be great

not quite used to the project layout + configuration via plugins that you have put together

A. Alonso Dominguez

@alonsodomin

forget that, found out how, SbtGit changes the version when doing a commit

Juan José Vázquez

@juanjovazquez

@alonsodomin yep, that's right. We're versioning through sbt-git. The version is frozen as soon as you create a tag. Thanks for contributing. Please, if you don't mind, raise an issue in order to track it properly. Thanks!.

A. Alonso Dominguez

@alonsodomin

hi @juanjovazquez, I just created the issue: Tecsisa/constructr-consul#41, any feedback on it would be appreciated.

Open to get something in a working state by today

^ hope!

Juan José Vázquez

@juanjovazquez

Thanks!

Steffen Gebert

@StephenKing

Hey. I've tried constructr-consul (with consul 1.0.0). I end up with

[ERR] http: Request PUT /v1/session/create, error: Missing node registration from=10.0.22.17:47478

logged on the consul server. My application emits

[DEBUG] [10/30/2017 13:32:49.420] [number-actors-akka.actor.default-dispatcher-3] [akka.tcp://number-actors@10.0.22.17:55998/system/IO-TCP/selectors/$a/2] Connection established to [consul.service.consul:8500]
[DEBUG] [10/30/2017 13:32:49.428] [number-actors-akka.actor.default-dispatcher-18] [ConsulCoordination(akka://number-actors)] [constructr-coordination-consul-responses] Element: HttpResponse(500 Internal Server Error,List(Date: Mon, 30 Oct 2017 13:32:49 GMT),HttpEntity.Strict(text/plain; charset=UTF-8,Missing node registration),HttpProtocol(HTTP/1.1))
[DEBUG] [10/30/2017 13:32:49.429] [number-actors-akka.actor.default-dispatcher-18] [ConsulCoordination(akka://number-actors)] [constructr-coordination-consul-responses] Downstream finished.
[WARN] [10/30/2017 13:32:49.431] [number-actors-akka.actor.default-dispatcher-3] [akka.tcp://number-actors@10.0.22.17:55998/system/constructr/constructr-machine] Failure in Locking, going to Locking: com.tecsisa.constructr.coordination.consul.ConsulCoordination$UnexpectedStatusCode: Unexpected status code 500 Internal Server Error for URI /v1/session/create
[DEBUG] [10/30/2017 13:32:49.432] [number-actors-akka.actor.default-dispatcher-3] [akka.tcp://number-actors@10.0.22.17:55998/system/constructr/constructr-machine] Transitioning from Locking to RetryScheduled

any idea?

I found some reports related to the 0.8.0 release, but 1.0.0 works fine also in the demo

Fran Bermejo

@franbh

Hi @StephenKing, sorry for the late response. We haven't experienced any issues like the one you're describing, but on the other hand we are still using Consul 0.9.x. Are you still having this problem?

Steffen Gebert

@StephenKing

@franbh kind of. I didn't continue any further, but I had reproduced this locally as well (not only inside Nomad)

Fran Bermejo

@franbh

@StephenKing We'll let you know when we migrate to Consul 1.0.x. BTW have you seen this? https://github.com/hashicorp/consul/issues/819#issuecomment-319745456 Are you waiting for Consul servers to complete booting and then start your application?

Steffen Gebert

@StephenKing

thanks @franbh, no I haven't seen that issue

Chka Davaadorj

@bayarmunkh

hi, consul.agent-name = ${?CONSUL_NAME}, where env { CONSUL_NAME = "${meta.unique.consul.name}" is this the right way to configure agent-name?

Chka Davaadorj

@bayarmunkh

ah it was ${attr.unique.consul.name}, not ${meta.unique.consul.name}.

Chka Davaadorj

@bayarmunkh

hi, i get a lot of Remote connection to [null] failed with java.net.ConnectException: Connection refused: /10.3.2.1:30001 errors. What's realistic Akka Cluster's failure-detector config with Consul?

cluster {
  failure-detector {
    acceptable-heartbeat-pause = 10s
    hearbeat-interval          = 3s
    threshold                  = 12.0
  }
}

i have this config, but still getting the failures.

Chka Davaadorj

@bayarmunkh

with the default failure-detector config, it was a lot more.

Fran Bermejo

@franbh

Hi @bayarmunkh , when are you getting the Connection refused errors? Is it when a cluster node tries to communicate with another one? Or is it while using a cluster client?

Chka Davaadorj

@bayarmunkh

@franbh it is when cluster nodes tried to communicate with each other. I thought network failure causing the issue, but turned out the OOM killer was killing the processes. It's all good now. Thanks!

Chka Davaadorj

@bayarmunkh

when a process is terminated unexpectedly, Tecsisa/constructr-consul#53 happens.

Fran Bermejo

@franbh

Good to hear that, @bayarmunkh ! Yes, you're right about Tecsisa/constructr-consul#53, unfortunately it still depends on hseeberger/constructr#169 (or hseeberger/constructr#170).

Chka Davaadorj

@bayarmunkh

yea, also unfortunately Constructr is no longer supported/maintained.

Where communities thrive

People

Repo info

Activity