As a consultant I’ve been really diving deep into an industry hot topic: Clustering. It seems to be the big topic for so many companies. Hi availability need not be an issue for you ever again. Here’s the simple way you can scale like the big boys do it. The long and short of it are simple. You need “Heartbeat.”
No, I don’t mean a biological function I mean Heartbeat from the Linux-HA project. Heartbeat is a service for Linux systems that will make a process highly available. This way if the main machine that is servicing clients goes down then you’ll have another that will take it’s place without anyone noticing.
Oddly enough, there seem to not be enough resources on the internet to help people with this subject. However, to get started there are so few resources to learn with. I actually wanted to go a little deeper in my own knowledge of the subject but couldn’t find many books and even fewer websites. What is this world coming to when I can’t find websites to teach me all the information I needed. The thing is… I looked everywhere but the source. The Linux-HA website. There’s a screencast video link here. It’s about 10 minutes and the program designer shows how to make a basic cluster. You can even go deeper with more examples on the site that teach you how to cluster different programs.
Just as a note from me, if you are truly considering learning all it takes to make your own cluster please consider learning more about high availability. The core to high availability is redundancy. Make sure that you have all of your bases covered. Have no single points of failure. Redundant hard disks (RAID), power, internet, and network links.
By the way, I had started this article a little while ago and noticed a great link on Howtoforge that really pulls this together quickly and in one page.
Hope these are very useful to you.