[
https://issues.apache.org/jira/browse/FLINK-4547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15455158#comment-15455158
]
ASF GitHub Bot commented on FLINK-4547:
---------------------------------------
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2455
In order to reuse the same gateway, one would need to cache them in the RPC
service. When would they be cleaned up? There needs to be a good solution to
that, otherwise it may create a slow but constant memory leak in the JobManager.
> Return same object when call connect method in AkkaRpcService using same
> address and same rpc gateway class
> -----------------------------------------------------------------------------------------------------------
>
> Key: FLINK-4547
> URL: https://issues.apache.org/jira/browse/FLINK-4547
> Project: Flink
> Issue Type: Sub-task
> Components: Distributed Coordination
> Reporter: zhangjing
> Assignee: zhangjing
>
> Now every time call connect method in AkkaRpcService class using same address
> and same rpc gateway class, the return gateway object is totally different
> with each other which equals and hashcode are not same.
> Maybe it’s reasonable to have the same result (equals return true, and
> hashcode is same) when using the same address and same Gateway class.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
