community-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Colin Patrick McCabe (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (COMDEV-192) Add a Kudu backend for Apache HTrace
Date Wed, 16 Mar 2016 18:08:33 GMT

    [ https://issues.apache.org/jira/browse/COMDEV-192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15197835#comment-15197835
] 

Colin Patrick McCabe commented on COMDEV-192:
---------------------------------------------

Hi [~ThamaliW],

It's great that you're interested in HTrace!  There are a lot of great ideas for improving
HTrace and its integration with the Hadoop ecosystem.

There has been a lot of interest in a Kudu receiver for HTrace!  Nisala Mendis just posted
to the HTrace development list about this.  I think a Kudu span receiver would be a great
idea and you should definitely file a proposal.

You may also be interested in some of the other HTrace proposals: there is one for integration
with YARN, and another for integration with the Hadoop s3 connector.  I actually feel like
these proposals would be very useful for users as well... and there is a lot to do and talk
about there.  (COMDEV-183 and COMDEV-191)

I definitely recommend subscribing to the HTrace development mailing list here: http://htrace.incubator.apache.org/mail-lists.html

As outlined on https://community.apache.org/gsoc.html, the next step is to upload a proposal
based on the template there.

> Add a Kudu backend for Apache HTrace
> ------------------------------------
>
>                 Key: COMDEV-192
>                 URL: https://issues.apache.org/jira/browse/COMDEV-192
>             Project: Community Development
>          Issue Type: New Feature
>            Reporter: Colin Patrick McCabe
>              Labels: gsoc, gsoc2016, mentor
>
> The Apache HTrace distributed tracing framework allows developers and system administrators
to get an end-to-end view of system performance, in a manner similar to XTrace or Dapper.
See http://htrace.incubator.apache.org/
> Kudu is a key/value store optimized for analytic workloads, with excellent performance
and latency on large datasets.  See https://blog.cloudera.com/blog/2015/09/kudu-new-apache-hadoop-storage-for-fast-analytics-on-fast-data/
 It is columnar and supports predicate push-down and many other optimizations.  Storing HTrace
traces in Kudu would allow us to run SQL queries on the data and get rapid responses (via
either Impala, Spark, or other execution engines).  This would enable near-realtime analysis
of HTrace data.



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

Mime
View raw message