[ 
https://issues.apache.org/jira/browse/FLINK-2821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15661367#comment-15661367
 ] 

Andrey Melentyev commented on FLINK-2821:
-----------------------------------------

The custom akka 3 build of Flink 1.1.3 also resolves the problem with running 
Flink on Kubernetes: with vanilla 1.1.3 build the JobManager's Akka system 
would drop the messages from Task Managers because of the IP mismatch. I 
couldn't find a working combination of job manager.rpc.address values without 
making Kubernetes use fixed IP address for Job Manager which kind of ruins the 
whole idea. 

With the custom 1.1.3 build the following setup works.

h5. Job Manager flink-conf.yaml

{noformat}
jobmanager.rpc.address: jobmanager.flink.svc.cluster.local
jobmanager.rpc.port: 6123
jobmanager.rpc.bind-address: 0.0.0.0
jobmanager.rpc.bind-port: 6123
{noformat}

h5. Task Manager flink-conf.yaml

{noformat}
jobmanager.rpc.address: jobmanager.flink.svc.cluster.local
jobmanager.rpc.port: 6123
jobmanager.rpc.bind-address: localhost # is this necessary on Task Manager side?
jobmanager.rpc.bind-port: 6123 # is this necessary on Task Manager side?
{noformat}

where jobmanager.flink.svc.cluster.local is the Kubernetes DNS record for the 
Job Manager service.

Sources: https://github.com/melentye/flink-kubernetes and 
https://github.com/melentye/flink-docker/tree/custom-akka3 which is published 
as melentye/flink:1.1.3-custom-akka3 on https://hub.docker.com/r/melentye/flink/

> Change Akka configuration to allow accessing actors from different URLs
> -----------------------------------------------------------------------
>
>                 Key: FLINK-2821
>                 URL: https://issues.apache.org/jira/browse/FLINK-2821
>             Project: Flink
>          Issue Type: Bug
>          Components: Distributed Coordination
>            Reporter: Robert Metzger
>            Assignee: Maximilian Michels
>
> Akka expects the actor's URL to be exactly matching.
> As pointed out here, cases where users were complaining about this: 
> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Error-trying-to-access-JM-through-proxy-td3018.html
>   - Proxy routing (as described here, send to the proxy URL, receiver 
> recognizes only original URL)
>   - Using hostname / IP interchangeably does not work (we solved this by 
> always putting IP addresses into URLs, never hostnames)
>   - Binding to multiple interfaces (any local 0.0.0.0) does not work. Still 
> no solution to that (but seems not too much of a restriction)
> I am aware that this is not possible due to Akka, so it is actually not a 
> Flink bug. But I think we should track the resolution of the issue here 
> anyways because its affecting our user's satisfaction.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to