[Discuss] easy clustering of applications

[tmetro+blu at gmail.com (Tom Metro)]

Rich Braun wrote:
> The problem with [most] services is they either don't have clustering
> capability or are a true pain to set up for clustering.  (Think postfix local
> delivery, think Jira, think MythTV's backend.)

True. This thread never really explored this aspect of Rich's question.

It does seem like every application has its own unique approach to
clustering. There is no generalized solution. We tend to gloss over this
with the prevalence of cloud services, but they're all web-based for
which there are countless clustering options.

In the same way that there are communities like turnkeylinux.org for
packing up popular open source applications as virtual appliances, there
should be a community that packages them up for HA clustering
environments. (In the same way that a package targets a particular Linux
distribution, these packages would need to target a specific clustering
environment.)

One potential answer to this is Canonical's Juju:
http://askubuntu.com/tags/juju/info

  Juju is apt-get for the cloud. Think of it as installing things onto a
  group of machines instead of one machine, and having them all
  automatically configure and talk to each other at the service level.
  Juju runs on HP Cloud, Amazon Web Services, OpenStack based private
  clouds, or on raw bare metal via MAAS

I'd rather see something a bit more removed from Canonical and in the
hand of the community, though it could still build on Juju. Also, more
focused on micro clusters. I suspect most Juju packages are designed
with the assumption you will be using 3+ node, with one acting as a load
balancer or similar. A collection of packages specifically designed to
work in 2-node setups would be ideal.

 -Tom

-- 
Tom Metro
The Perl Shop, Newton, MA, USA
"Predictable On-demand Perl Consulting."
http://www.theperlshop.com/