Mailing-List: contact hdfs-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-dev@hadoop.apache.org
Date: Sat, 23 Jan 2016 21:32:39 +0000 (UTC)
From: "Nikhil Mulley (JIRA)" <jira@apache.org>
To: hdfs-dev@hadoop.apache.org
Message-ID: <JIRA.12933627.1453584708000.191951.1453584759795@Atlassian.JIRA>
In-Reply-To: <JIRA.12933627.1453584708000@Atlassian.JIRA>
References: <JIRA.12933627.1453584708000@Atlassian.JIRA>
 <JIRA.12933627.1453584708627@arcas>
Subject: [jira] [Created] (HDFS-9692) Report top 10 Namenode rpc consumers
 through JMX
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Nikhil Mulley created HDFS-9692:
-----------------------------------

             Summary: Report top 10 Namenode rpc consumers through JMX
                 Key: HDFS-9692
                 URL: https://issues.apache.org/jira/browse/HDFS-9692
             Project: Hadoop HDFS
          Issue Type: Wish
            Reporter: Nikhil Mulley


Hi,

I think it would really help if namenode(s) through metrics/jmx report the top rpc consumers, so it will be really handy to look at the rogue clients in times of despair/troubleshooting times.
At times of rpc spikes on namenode, and callqueuelength increasing, it becomes tedious to figure out the offenders when there is a huge cluster (>1k nodes).
Having rpc client information(src_host:src_port)  in the top consumers list would help operators. Let me know if any other information is needed to make this feature added to namenode metrics system.

thank you

Nikhil


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