Mailing-List: contact dev-help@reef.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@reef.apache.org
Date: Wed, 6 Apr 2016 22:54:25 +0000 (UTC)
From: "Julia (JIRA)" <jira@apache.org>
To: dev@reef.apache.org
Message-ID: <JIRA.12954135.1459208159000.156700.1459983265691@Atlassian.JIRA>
In-Reply-To: <JIRA.12954135.1459208159000@Atlassian.JIRA>
References: <JIRA.12954135.1459208159000@Atlassian.JIRA>
 <JIRA.12954135.1459208159214@arcas>
Subject: [jira] [Updated] (REEF-1292) Link and StreamingLink classes should
 have retry logic for setting up remote connection
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/REEF-1292?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Julia updated REEF-1292:
------------------------
    Labels: FT  (was: )

> Link and StreamingLink classes should have retry logic for setting up remote connection
> ---------------------------------------------------------------------------------------
>
>                 Key: REEF-1292
>                 URL: https://issues.apache.org/jira/browse/REEF-1292
>             Project: REEF
>          Issue Type: Improvement
>          Components: Wake.NET
>         Environment: C#
>            Reporter: Dhruv Mahajan
>            Assignee: Dhruv Mahajan
>              Labels: FT
>             Fix For: 0.15
>
>
> Currently {{Org.Apache.REEF.Wake.Remote.Impl.Link}} and {{Org.Apache.REEF.Wake.Remote.Impl.StreamingLink}} try only once to establish connection to remote endpoint. However, if remote server is very busy or loaded, it might reject the connection. This can happen for example, when we ask for say 1000 evaluators and NameClient tries to open connection with NameServer which is overwhelmed because of huge traffic. 
> The aim of JIRA is to introduce this re-try logic


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