chukwa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jiaqi Tan <>
Subject Re: Exploring chukwa-initial questions!
Date Wed, 28 Apr 2010 00:40:47 GMT
Hi Arvind,

Welcome to the mailing list :)

I have been using Chukwa slightly differently from most; I have been
using it in an "offline" mode to process Hadoop logs and other system
metrics by using the backfill loader, which is a special mode for
loading logs previously generated.

I work on failure diagnosis and visualization of system behavior for
Hadoop, and I have been using Chukwa to help with that. Some of our
previous work has already been implemented and is available in Chukwa,
and I have been working on putting more of the stuff in (albeit at a
much slower rate than I would like).


On Wed, Apr 28, 2010 at 5:28 AM, Ariel Rabkin <> wrote:
> My lab-mates use chukwa regularly on EC2 to monitor test deployments.
> Nothing particularly distinctive about it.
> Our deployment model is typically to run the agents in the cloud, and
> the collector at Berkeley, pointing to a static HDFS cluster holding
> the collected logs.
> --Ari
> On Tue, Apr 27, 2010 at 1:40 PM, arvind subramanian
> <> wrote:
>> Hello,
>> Firstly, Congrats on the release of 0.4!
>> I am Arvind, and I  have just got started in exploring Chukwa
>> I have been reading  the docs from the release in 0.40 , issues on JIRA, and
>> some of the activity in the mailing list.
>> I am familiar with Hadoop/Hdfs having used them in pseudo-distributed mode
>> on laptop(Ubuntu 9.10), and  have also used hadoop AMI  on Ec2.
>> Has a Chukwa cluster been deployed on Ec2 ? If so, it would be great to see
>> some documentation/tutorials/pointers  of the same, so that I could try it
>> out.
>> Also,It would be nice to see more elaboration on setting up the development
>> environment for  Chukwa (pseudo-distributed mode)
>> Regards,
>> Arvind
> --
> Ari Rabkin
> UC Berkeley Computer Science Department

View raw message