Hi i need a system that has high availability and load balancing.. although im not really sure the kind of setup i would need and the specifications... It needs to be designed for being upgradable.. can anyone help?

Thanks
lewis

Your post does not detail much into what you're trying to do. A lot of those words are simply buzz words if you don't add context or applications you're trying to run. Can you provide more details?

Well at the moment i have a site which has a mysql database and customers files stored on the hard drives although what we want to provide is high availability to reduce any downtime where possible and load balancing to manage the load although i was told that HA clustering was the way to go.. i was told i need 2 masters. A server for storage. A server for database. And nodes. Although im looking for a way. If possible.. to spend less money and not have to buy lots of servers all at once. And also how will i know what specifications would the node masters be?

Here is a simple diagram of what i need it todo.. as i need it to cope with quite a few thousand customers..with a way to expand.http://s30.postimg.org/wnwf22l9t/IMAG0240.jpg

as sag47 said, you would need to be a bit more specific about what you want to do.
A RaspberryPi could be sufficient for what you need, a huge CDN might be underpowered.

I assume from your posts you might want to set up some kind of internet service, maybe a online shop, for a few thousand customers. In this case with your level of experience (remember, I'm just assuming since I don't have much to work with) I'd recommend to use a professional service provider, who runs the infrastructure and has experience and man power to fulfill your needs. There are many companies available around the world, who offer these services. Of course, you wouldn't get it for less than nothing, everything has it's price and value.

For MySQL database clustering check out galera clusters.

Regarding static files being served across a cluster of more than one file share you can refer to drbd. There's also gluster or ceph clusters for storage.

Apache can load balance http-based and ajp-based protocol web applications. See mod_proxy, mod_jk, and mod_balancer. The same could be done using nginx.

You might consider putting a GTM (global traffic manager) out front of a pair of load balancers. As well has host multiple clusters in multiple regions of which the GTM can manage traffic across geographically.

Really it depends on your budget and how nuts you want to go. There's also distributed clustering of web applications using docker containers and Apache mesos. Again, depends.

Also, you're not going to manage those clusters manually, right? Best to use configuration management such as cfengine, chef, puppet, ansible, or salt stack.

Store your configurations in SCM such as git.

Check out the hashicorp website. They have great open source software for testing your infrastructure as code.

Ah thanks.. well i have about £10,000 maybe more if i put in a request...alough obvously the less i spend the better.. the main question is what kind of hardware would i need for the main 2 servers .. i know its hard to tell but we need a base platform where we can upgrade...even if its only upgrading one at a time... Will 2x hp proliant dl360es work and what would we need to focus on more.. the cpu or ram .. as the nodes must. Do most of the processing ?

Thanks
lewis

I don't know what your client needs or limits are which are decisions that should be data driven. So I won't comment much on hardware. I will say this. Regarding the storage clusters it's more useful to have many really cheap machines than to have a few expensive ones.