[
https://issues.apache.org/jira/browse/KAFKA-7111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16527577#comment-16527577
]
Stanislav Kozlovski commented on KAFKA-7111:
--------------------------------------------
Hey Luan,
After debugging for a while, I worked out that the problem is that your
producer (Spring Application) is not in the docker network where the Kafka
broker is running.
What's happening is your producer sends an `UpdateMetadataRequest` and gets
returned the brokers' urls. Since the broker runs inside the docker network, he
believes his address is `broker:9092` and hence returns that. Since your
producer is outside the docker network, he is unable to resolve that address.
This situation is a bit unusual and I believe that is why this hasn't been
caught yet.
> Review the NetworkClient log level used
> ---------------------------------------
>
> Key: KAFKA-7111
> URL: https://issues.apache.org/jira/browse/KAFKA-7111
> Project: Kafka
> Issue Type: Improvement
> Components: clients
> Reporter: Luan Cestari
> Assignee: Stanislav Kozlovski
> Priority: Trivial
>
> Hi,
>
> I was using Kafka on some projects and unfortunately I had to use Debug (and
> some times even Trace) log level to see some issues. One of the most recently
> cases was:
> [https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/clients/NetworkClient.java#L872]
>
> If I got the name of the broker and it is unreachable, the errors should be
> more severe than "DEBUG" level IMHO. I would at least put a INFO level for
> this case or ERROR level (which seems to fit better but I don't know the
> practices used in the project).
>
> Thank you in advance
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
