You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Stephan Ewen (JIRA)" <ji...@apache.org> on 2016/08/11 15:31:20 UTC

[jira] [Commented] (FLINK-4382) Buffer rpc calls until RpcEndpoint is properly started

    [ https://issues.apache.org/jira/browse/FLINK-4382?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15417432#comment-15417432 ] 

Stephan Ewen commented on FLINK-4382:
-------------------------------------

Would it buffer the messages, or simply discard them while the endpoint is not ready?
I think discarding is fine. It would be equivalent to the behavior that the endpoint started later (was unavailable for some time), which all senders have to be able to deal with anyways.

> Buffer rpc calls until RpcEndpoint is properly started
> ------------------------------------------------------
>
>                 Key: FLINK-4382
>                 URL: https://issues.apache.org/jira/browse/FLINK-4382
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Distributed Coordination
>            Reporter: Till Rohrmann
>            Assignee: Till Rohrmann
>
> When creating a {{RpcEndpoint}} it starts a rpc server. The server should wait to dispatch incoming rpc calls until the {{RpcEndpoint}} signals that it's ready.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)