[
https://issues.apache.org/jira/browse/FLINK-22284?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Zhilong Hong updated FLINK-22284:
---------------------------------
Description:
In NettyPartitionRequestClient#requestSubpartition, when a channel is closed,
the channel will throw a LocalTransportException with the error message
"Sending the partition request to 'null' failed.". The message is confusing
since we wouldn't know where the remote client connected to this channel
locates, and we couldn't track down to that TaskExecutor and find out what
happened.
Also I'm wondering that should we use TransportException instead of
LocalTransportException here, because it's a little confusing to see a
LocalTransportException is thrown out when a remote channel is closed.
was:
In NettyPartitionRequestClient#requestSubpartition, when channel is closed, the
channel will throw a LocalTransportException with the error message "Sending
the partition request to 'null' failed.". The message is confusing since we
wouldn't know where the remote client connected to this channel locates, and we
couldn't track down to that TaskExecutor and find out what happened.
Also I'm wondering that should we use TransportException instead of
LocalTransportException here, because it's a little confusing to see a
LocalTransportException is thrown out when a remote channel is closed.
> Null address will be logged when channel is closed in
> NettyPartitionRequestClient
> ---------------------------------------------------------------------------------
>
> Key: FLINK-22284
> URL: https://issues.apache.org/jira/browse/FLINK-22284
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Network
> Affects Versions: 1.13.0
> Reporter: Zhilong Hong
> Priority: Minor
> Fix For: 1.13.0
>
>
> In NettyPartitionRequestClient#requestSubpartition, when a channel is closed,
> the channel will throw a LocalTransportException with the error message
> "Sending the partition request to 'null' failed.". The message is confusing
> since we wouldn't know where the remote client connected to this channel
> locates, and we couldn't track down to that TaskExecutor and find out what
> happened.
>
> Also I'm wondering that should we use TransportException instead of
> LocalTransportException here, because it's a little confusing to see a
> LocalTransportException is thrown out when a remote channel is closed.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
