Chapter 12. Keeping Things Running with Virtual Clusters
Every piece of your hardwre will fail at some point. It might be 15 years down the road, or it might be today. Your hardware lifecycle might replace hardware before it fails, but chances are that you've experienced some sort of hardware failure since you started working with computers. Moving components fail more often than stationary ones, even if this happens less frequently. I've seen my share of hard drive and fan failures, but I have also seen RAM chips, motherboards, and network cards die.
In this chapter, we look at how to setup a fault tolerant virtual cluster to avoid any single point of failure Deciding where your single point of failure is (server, server room, site, and so on) as a function of budget is discussed, as is implementing fault tolerance at many levels. This chapter also discusses defining different types of disasters as well as their corresponding recovery time objectives (RTO). Using clusters is the most fault tolerant and redundant way to deploy VMware. I would say that again, but that would be redundant!
Reducing Single Points of Failure
If you think about it, no matter how careful you are, you always have a single point of failure. If you have a standalone server with no redundancy, the disk subsystem, the motherboard, CPU, and network cards all represent a single point of failure. Because disks fail more often than stationary components, you can remove a disk as a single point of failure by mirroring ...