Flink on Kubernetes - Hostname resolution between job/tasks-managers

[bastien dine]

Hello,

I am trying to install Flink on Kube, it's almost working.. 

I am using the kube files on flink 1.7.1 doc

My cluster is starting well, my 2 tasksmanagers are registering successfully to job manager

On webUI, i see them :

akka.tcp://flink@dev-flink-taskmanager-3717639837-gvwh4:37057/user/taskmanager_0

I can submit a job too..

But when I am going in job detail, or try to load the logs.. I have nothing.. and log on jobmanager give me plenty of error like :

2019-01-15 14:12:40.111 [flink-metrics-96] WARN akka.remote.ReliableDeliverySupervisor flink-metrics-akka.remote.default-remote-dispatcher-113 - Association with remote system [akka.tcp://flink-metrics@dev-flink-taskmanager-3717639837-gvwh4:40508] has failed, address is now gated for [50] ms. Reason: [Association failed with [akka.tcp://flink-metrics@dev-flink-taskmanager-3717639837-gvwh4:40508]] Caused by: [dev-flink-taskmanager-3717639837-gvwh4: Name does not resolve]

-> Name does not resolve.. 

So trying to ping on the pod hostname and it's not working

Thus, ping on the pod's IP is working

So, my question is :

- Can we force usage of IPv4 over hostname resolution ? (will be better for perf also)

- If no, do I need to had a service or something to make it work ?

Best Regards,

Bastien

------------------

[bastien dine]

Nevermind.. 

Problem already discussed in thread : 

Flink 1.7 jobmanager tries to lookup taskmanager by its hostname in k8s environment"

------------------

Le mar. 15 janv. 2019 à 15:16, bastien dine <[hidden email]> a écrit :

Hello,

I am trying to install Flink on Kube, it's almost working.. 

I am using the kube files on flink 1.7.1 doc

My cluster is starting well, my 2 tasksmanagers are registering successfully to job manager

On webUI, i see them :

akka.tcp://flink@dev-flink-taskmanager-3717639837-gvwh4:37057/user/taskmanager_0

I can submit a job too..

But when I am going in job detail, or try to load the logs.. I have nothing.. and log on jobmanager give me plenty of error like :

2019-01-15 14:12:40.111 [flink-metrics-96] WARN akka.remote.ReliableDeliverySupervisor flink-metrics-akka.remote.default-remote-dispatcher-113 - Association with remote system [akka.tcp://flink-metrics@dev-flink-taskmanager-3717639837-gvwh4:40508] has failed, address is now gated for [50] ms. Reason: [Association failed with [akka.tcp://flink-metrics@dev-flink-taskmanager-3717639837-gvwh4:40508]] Caused by: [dev-flink-taskmanager-3717639837-gvwh4: Name does not resolve]

-> Name does not resolve.. 

So trying to ping on the pod hostname and it's not working

Thus, ping on the pod's IP is working

So, my question is :

- Can we force usage of IPv4 over hostname resolution ? (will be better for perf also)

- If no, do I need to had a service or something to make it work ?

Best Regards,

Bastien

------------------