cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paco Trujillo <F.Truji...@genetwister.nl>
Subject RE: Centralized tool to install and manage C* from one machine
Date Tue, 28 Jan 2014 07:20:13 GMT
Hi

To manage the cluster we are using the OpsCenter, it is a free web tool from Datastax that
allow you basic execute basic operations in any node on the cluster. It also contain differents
metrics which you can add to a kind of dashboard and monitoring the most important parameters
in the cluster.

To install new nodes, we still do that manually.

From: Kumar Ranjan [mailto:winnerdood@gmail.com]
Sent: dinsdag 28 januari 2014 06:38
To: user@cassandra.apache.org
Subject: Re: Centralized tool to install and manage C* from one machine

Thank you Michael. I am trying out Priam as we speak and will post an update of my experience
with different tools. Again. Thank you. -- K

On Tue, Jan 28, 2014 at 12:29 AM, Michael Shuler <michael@pbandjelly.org<mailto:michael@pbandjelly.org>>
wrote:
On 01/27/2014 10:34 PM, Kumar Ranjan wrote:
I am used to working with CCM for testing. For production, I depend on
installing cassandra manually. Is there a proven tool to install and
manage multinode cassandra cluster? If you have any experience, please
let me know.

A good answer will depend on where you will be deploying your cluster, how many nodes you
plan, etc.  For bare metal installations, for instance with Debian or Ubuntu, you may want
to automate with a preseed [0] of your OS installations to set up the JVM, install your desired
cassandra version package, etc.  RedHat variants use similar OS automation with kickstart
[1]. Or, perhaps you want to use AWS and a preinstalled AMI [2] or create your own "golden
image" on AWS and save it as an AMI for booting your other machines.

I've done preseeds and kickstarts extensively and have set up a few custom AMIs - you may
want to consider that most production clusters are likely running Debian or Ubuntu.

Once you have machines installed, you may want some configuration management, or you could
use a config manager to aid in OS installation and setup at install time - it seems that chef
[3] has gained some traction over the once-hot puppet [4], but salt [5] is also quite mature,
if you like python better than ruby.

I've used all of these and they are great.  I've also been burned by them all.  There's nothing
quite like the complete control over configurations carefully checked into a VCS and parallel
ssh to pull them out, along with a few scripts to set up things just right - just my experience
 ;)

There is also priam [6].  I keep wanting to find some time to play with it, so I have no insight,
but it looks very interesting.

If you'd like cassandra cluster management beyond installation/configuration of the machine,
have a look at opscenter [7].

I'm sure there are a lot of other projects / cookbooks that others might be working on - chime
in!

Michael

[0] https://wiki.debian.org/DebianInstaller/Preseed
[1] https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Installation_Guide/ch-kickstart2.html
[2] https://aws.amazon.com/amis/datastax-auto-clustering-ami-2-2
[3] http://community.opscode.com/cookbooks/cassandra
[4] https://forge.puppetlabs.com/tags/cassandra
[5] http://docs.saltstack.com/ref/modules/all/salt.modules.cassandra.html
[6] https://github.com/Netflix/Priam
[7] http://www.datastax.com/what-we-offer/products-services/datastax-opscenter


Mime
View raw message