EndpointWriter error message in logs


(seemanto barua) #1

Is there way to turn off these kind of logs. Because it occurs very frequently. I understand what is the reason for the complain that appears on the log. But if I dont have the option to stop the process that is trying to get an actor ref of an actor in my cluster, I end up with a lot of these on my logs.

2018-07-25 19:13:18.210 [stage-akka.actor.default-dispatcher-23] **akka.remote.EndpointWriter**
ERROR: dropping message [class akka.actor.ActorSelectionMessage] for non-local recipient [Actor[akka.tcp://sbarua@127.0.0.1.nomura.com:2551/]] arriving at [akka.tcp://sbarua@127.0.0.1.nomura.com:2551] inbound addresses are [akka.tcp://stage@127.0.0.1.nomura.com:2551]

(Johan Andrén) #2

This error message means that you are receiving messages destined for 127.0.0.1 but you have bound your node to 127.0.0.1.nomura.com. That looks like a configuration error to me and could very likely be an actual problem, having a node that you cannot send messages to at all, I think you should look into that rather than try to silence the error.

In general you need to make sure that the exact value of hostname (or canonical.hostname in case of using artery) is what other hosts use to connect to your node.


(seemanto barua) #3

hi

it is not because of the hostname. We can see on the log message that the bound hostname and the hostname on the message is same. The issue is with the ActorSystem names which are different. I don’t want those messages to come to my cluster or appear in the logs at all.
thanks