[
https://issues.apache.org/jira/browse/FLINK-20743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17254795#comment-17254795
]
Zhijiang commented on FLINK-20743:
----------------------------------
Thanks for creating this issue [~清月].
I remembered I also encountered the same concern while debugging some failover
issues before. I guess we might rely on the port info to trace the other
required infos such as container ID from job manager log. But indeed it is not
very convinenet and efficient. So in general I supoort your proposal for
improving the debugging process, but I am a bit worried that it might not be
easy to pass container ID into the network stack which might touch many
components.
Anyway, you can try out your way and I can assign this ticket to you if you
desire to contribute it. :)
> Print ContainerId For RemoteTransportException
> ----------------------------------------------
>
> Key: FLINK-20743
> URL: https://issues.apache.org/jira/browse/FLINK-20743
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Network
> Affects Versions: 1.10.0, 1.11.1, 1.12.1
> Reporter: yang gang
> Priority: Major
> Attachments: image-2020-12-23-15-13-21-226.png
>
>
> !image-2020-12-23-15-13-21-226.png|width=970,height=291!
> RemoteTransportException, this exception reminds the user which service has
> a problem by means of Ip/Port.
> When we troubleshoot the problem, the information is not accurate enough.
> Usually at this time we need to look at the running log of the container that
> has the problem, but when we see this log, it also shows that the container
> has died, so pass Ip/ The port method can no longer quickly locate a specific
> container.
> So I hope that when such an exception occurs, I hope to print the
> containerId。
> E.g:
> Connection unexpectedly closed by remote task manager
> 'hostName/ip:port/containerId'. This might indicate that the remote task
> manager was lost.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
