Chapter 8. Handling Machine Failure
Machines fail, and there are diminishing returns in reducing the failure rate of individual machines. And sometimes, even if a machine is operating correctly, you might not be able to reach it due to a networking issue. But machines can come and go as part of normal operations too. If we broaden our definition of a machine beyond just the physical computer and also think about virtual machines or containers, normal deployment activities can involve machines shutting down and new machines, potentially located in different places, spinning up in their place.
We have to accept that machines can come and go, so we need to work out how to deal with machine failure when it happens. A theme running through this book is “making sure that as many things go right as possible,” so with that in mind, we should turn our attention to how we make sure things go right even if a machine becomes unreachable.
In this chapter, we’ll explore how clients can find servers, as this can be key in recovering from machines being unavailable. We’ll also look at the topic of load balancing as a mechanism for reducing the impact of a single server instance failing. Finally, we’ll look at the challenges that stateful processing causes when also trying to deal with failure.
Service Discovery
If the machine a client is talking to suddenly becomes unavailable, the client may need to be pointed to another machine. With that in mind, let’s take a look at some of the options around ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access