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

Posted to common-issues@hadoop.apache.org by "tom liu (JIRA)" <ji...@apache.org> on 2010/10/28 11:58:23 UTC

[jira] Updated: (HADOOP-7012) IPC.Client may have some connections for each given host/port

     [ https://issues.apache.org/jira/browse/HADOOP-7012?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

tom liu updated HADOOP-7012:
----------------------------

    Summary: IPC.Client may have some connections for each given host/port  (was: IPC.Client may have some connection for each a given host/port)

> IPC.Client may have some connections for each given host/port
> -------------------------------------------------------------
>
>                 Key: HADOOP-7012
>                 URL: https://issues.apache.org/jira/browse/HADOOP-7012
>             Project: Hadoop Common
>          Issue Type: Improvement
>          Components: ipc
>         Environment: JDK1.6.0_17/Tomcat6
>            Reporter: tom liu
>
> ipc.Client class holds connection pool, that connection would be reused for a given host/port.
> and WritableRpcEngine holds Client pool, that client would be reused for a given SocketFactory.
> so, RPC.getProxy method returns proxy object, that only hold one connection or socket connected to a host/port, if or not, we create many proxy objects.
> if large requests are sent to RPC Server through proxy, the request would be wait long time to be sent.
> i think, for each host/port, ipc.Client would hold many connections.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.