Summary

In this chapter, we learned about a few different controls that Kubernetes provides in order to enforce certain Pod placement rules via the scheduler. We learned that there are both "hard" requirements and "soft" rules, the latter of which are given the scheduler's best effort but do not necessarily prevent Pods that break the rules from being placed. We also learned a few reasons why you may want to implement scheduling controls – such as real-life failure domains and multitenancy.

We learned that there are simple ways to influence Pod placement, such as node selectors and node names – in addition to more advanced methods like taints and tolerations, which Kubernetes itself also uses by default. Finally, we discovered that there are some advanced tools that Kubernetes provides for node and Pod affinities and anti-affinities, which allow us to create complex rulesets for the scheduler to follow.

In the next chapter, we will discuss...