I am trying to build a cluster of HP DL380 servers, and I am having a lot of trouble getting them to do what I want. I have tried Windows Server 2003 and 2008, and am now exploring the possibility of a Linux solution and was hoping somebody could point me in the right direction.

I am trying to make 1 powerful system that will let me throw a large amount of services on it. I manage a network of 1000+ computers, and will need things like DNS, DHCP, Mail Serving, and Web Serving. I am currently doing these tasks on individual servers, but am trying to achieve better redundancy and better utilization of my available power.

I have looked at Linux clustering but it seems to be similar to Windows clustering, where you are still only running 1 service on 1 machine. When that 1 fails, then another takes over. But I want to run all services on all machines all at once. If my utilization get to high, then I want to be able to just buy more servers and add them in.

If people could pass on their wisdom in this matter, I would greatly appreciate it. I have a basic knowledge of Linux, but am no expert. Thanks for any help anyone can provide...

I think the best way to do what you are trying to do would be to have 1 server to run ha lvs as a load balancer and then 2 high power servers as your cluster servers, you could if you wish have a secondary load balancer to make things more redundant.

This works well and I have it running in large scale on a global client. They are running DL380's for the servers.

Check out:

http://www.linuxvirtualserver.org/

I should also mention that I have a HP MSA500 capable of connecting 4 nodes, to use as a shared storage.

Thanks kentyler for your response, I have taken a look at the Load Balancing solution. But I was leaning away from that, as it needs to have 1 or 2 servers sitting basically unused running the cluster. I was hoping to be able to run all my servers at 100% utilization if need be...

I'm not sure if the config is clear enough about how you can configure it.

I have the mentioned servers doing a 1 to 1 ratio so one request goes to one server the next request goes to the other server.

You 'can' have all requests go to one and have the other in failover standby if you like but that's not how I configured this setup.

The software even shows how requests are handled with counts per server/service and then an overall cluster count for each service.

It's easy to configure and stable.

System uptime now is 500+ days since install.