Mailing-List: contact dev-help@avro.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@avro.apache.org
Date: Thu, 6 Aug 2015 13:46:05 +0000 (UTC)
From: "Todd Baldwin (JIRA)" <jira@apache.org>
To: dev@avro.apache.org
Message-ID: <JIRA.12852578.1438868729000.389178.1438868765266@Atlassian.JIRA>
In-Reply-To: <JIRA.12852578.1438868729000@Atlassian.JIRA>
References: <JIRA.12852578.1438868729000@Atlassian.JIRA>
 <JIRA.12852578.1438868729183@arcas>
Subject: [jira] [Created] (AVRO-1712) NettyAvroRpcClient.configure()
 prevents timeouts less than one second
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Todd Baldwin created AVRO-1712:
----------------------------------

             Summary: NettyAvroRpcClient.configure() prevents timeouts less than one second
                 Key: AVRO-1712
                 URL: https://issues.apache.org/jira/browse/AVRO-1712
             Project: Avro
          Issue Type: Improvement
          Components: java
    Affects Versions: 1.7.7
            Reporter: Todd Baldwin
            Priority: Minor


The NettyAvroRpcClient prevents connection or request timeouts less than one second.  I would like to remove this restriction.  I'm hoping to specify values down to around the tens-of-millisecond ranges.

My actual use case is log4j/Flume with Load-Balancing Log4jAppender using Avro.  I have 50 clients hosts sending to 70 Flume collectors.  Under high load, log4j/Flume can spend a long time in timeouts to multiple destinations with the 1 second request/connect timeouts.  I would like to bring the NettyAvroRpcClient timeouts down into the 50 millisecond-or-so range to help avoid buffer overflows when one or more destinations are down.


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