[
https://issues.apache.org/jira/browse/HDFS-6982?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14118971#comment-14118971
]
Maysam Yabandeh commented on HDFS-6982:
---------------------------------------
Thanks [~philip]. I agree with you. I actually was planning to skip the audit
log tailing stuff altogether to keep the patch simple. If there was interest in
future I can submit a separate patch for that.
The metric key format is "operation.user". Here is a sample output from the jmx
interface:
{code}
[myabandeh@smf1-aro-39-sr1(hadoop-tst-nn) ~]$ curl localhost:12333/jmx | grep
Hadoop:service=nntop,name=topusers -B1 -A8
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
}, {
"name" : "Hadoop:service=nntop,name=topusers",
"modelerType" : "topusers",
"tag.Context" : "namenode",
"tag.ProcessName" : "DummyProcessName",
"tag.SessionId" : "DummySessionId",
"tag.Hostname" : "hhh",
"delete.xxx" : 1,
"setPermission.ALL" : 0,
"getfileinfo.ALL" : 3159,
{code}
> nntop: top-like tool for name node users
> -----------------------------------------
>
> Key: HDFS-6982
> URL: https://issues.apache.org/jira/browse/HDFS-6982
> Project: Hadoop HDFS
> Issue Type: New Feature
> Reporter: Maysam Yabandeh
> Attachments: nntop-design-v1.pdf
>
>
> In this jira we motivate the need for nntop, a tool that, similarly to what
> top does in Linux, gives the list of top users of the HDFS name node and
> gives insight about which users are sending majority of each traffic type to
> the name node. This information turns out to be the most critical when the
> name node is under pressure and the HDFS admin needs to know which user is
> hammering the name node and with what kind of requests. Here we present the
> design of nntop which has been in production at Twitter in the past 10
> months. nntop proved to have low cpu overhead (< 2% in a cluster of 4K
> nodes), low memory footprint (less than a few MB), and quite efficient for
> the write path (only two hash lookup for updating a metric).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
