You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@flume.apache.org by "Hari Shreedharan (JIRA)" <ji...@apache.org> on 2015/02/21 02:03:24 UTC

[jira] [Commented] (FLUME-2627) RpcClient implementations have inconsistent handling of unreachable hosts

    [ https://issues.apache.org/jira/browse/FLUME-2627?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14329874#comment-14329874 ] 

Hari Shreedharan commented on FLUME-2627:
-----------------------------------------

Lets make sure all of them have the same behavior. I think they should all throw an exception and complain. For compound clients, like LoadBalancing and Failover, we must throw only if we cannot connect to any of the hosts though.

> RpcClient implementations have inconsistent handling of unreachable hosts
> -------------------------------------------------------------------------
>
>                 Key: FLUME-2627
>                 URL: https://issues.apache.org/jira/browse/FLUME-2627
>             Project: Flume
>          Issue Type: Bug
>            Reporter: Santiago M. Mola
>            Priority: Minor
>
> NettyAvroRpcClient.configure will throw FlumeException if it cannot connect to host.
> ThriftRpcClient.configure will try to establish a connection but it will not throw if it cannot connect, it will set its internal connection state to DEAD and will return false for ThriftRpcClient.isActive().
> LoadBalancingRpcClient.configure and FailoverRpcClient.configure will not try to connect at all.



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