[Discuss] Redundant array of inexpensive servers: clustering?

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

Rich Braun wrote:
> My goal is simple:  mash the power button or yank the network cable from
> either of these machines, and have all the apps still running.  Then plug the
> machine back in and have all state restored to full redundancy without having
> to type any commands.
> 
> For me, the use-case for HA technology really isn't just about designing
> around failure.  The more-important thing for me as a weekend-hobbyist user is
> being able to take something down, mess with it/upgrade it/overhaul it for a
> few hours or days, and put it back online without disrupting anything else.

I'm also very interested n the topic of micro clusters for home use. I'd
like to see this topic explored further here or elsewhere.

It would be great if Ken and others planning to build such systems
shared their proposed designs, and reported back on how well each aspect
worked.

As I suggested in the other posting on the packaging of applications, it
would similarly be nice to see a turn-key distribution of OpenStack that
is aimed at the micro cluster market.

If I recall correctly from Federico's earlier talk on OpenStack
(http://blu.org/cgi-bin/calendar/2013-aug), OpenStack may be ill suited
to a micro cluster. He ended up with, I think, a 9-node cluster, which
he felt was the minimum viable configuration.


> Is HA a dead technology?

Obviously not. It's more active than ever.


> What am I missing?  Did the commercial cloud services co-opt all
> development and horde this technology for their own overpriced business
> models?

Yes, I believe so.

It depends on what layer you are looking at. There are several open
source solutions for creating private clouds, with OpenStack being the
one most talked about these days. There are several clustering file
systems in use in production. You mentioned GlusterFS, and there is also
Ceph (http://ceph.newdream.net/, part of OpenStack), and others.

Lots has happened on the web side of things. Lots of off-the-shelf, open
source technologies to address that.

I think much of the reset ends up being carefully developed in-house
configurations that haven't been shared back with the community.

But I'm mostly speculating based on what I've read. This is outside my
areas of expertise.

 -Tom

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