Below are example rules to demonstrate this configuration method. These rules are valid for both node and pod affinity and can be configured using the respective affinity fields.

preferredDuringSchedulingIgnoredDuringExecution.requiredDuringSchedulingIgnoredDuringExecution.You can configure these rules with the following fields in a pod manifest: Pod affinity and anti-affinity are also specified within the affinity section using the podAffinity and podAntiAffinity fields in the PodSpec.īoth node affinity and pod affinity can set required and preferred rules that act as hard and soft rules. For example, it can avoid placing two pods of the same application on the same Kubernetes node to offer redundancy in a failure scenario (more on this later). It prevents pods from being placed on the same node. Pod affinity provides administrators with a way to configure the scheduling process using labels configured on the existing pods within a node.Ī pod with a label key-value pair matching a condition can be scheduled on a node containing the matching labels.Īs the name suggests, pod anti-affinity simply offers the opposite functionality. Node affinity is specified in the PodSpec using the nodeAffinity field in the affinity section. Node affinity enables a conditional approach with logical operators in the matching process, while nodeSelector is limited to looking for exact label key-value pair matches. It is similar to nodeSelector but provides more granular control over the selection process. Kubernetes node affinity is a feature that enables administrators to match pods according to the labels on nodes. Node affinity attracts pods to nodes, and pod affinity attracts pods to pods. Since the naming may be misleading, it’s important to realize that both features are meant from the pod’s perspective. There are two types of affinity: Kubernetes node affinity and Kubernetes pod affinity. Match labels of pods running within nodes and determine the scheduling location of new pods.Create “preferred” and “required” rules for a greater variety of matching conditions.Define rules, including conditions with logical operators.With affinity and anti-affinity, administrators can: The affinity and anti-affinity features in Kubernetes provide administrators with more granular scheduling functionality. What are Kubernetes affinity and anti-affinity? With Kubernetes affinity, administrators gain greater control over the pod scheduling process. However, this method quickly becomes inadequate to facilitate complex use cases and larger K8s clusters. Using the nodeSelector is the recommended way to match pods with nodes for simple use cases in small Kubernetes clusters. Spec : containers : - name : nginx image : nginx nodeSelector : : You can use the following command to add labels to the nodes. If the key-value pair matches exactly the label defined in the node, the pod will get matched to the specific node. You can specify the nodeSelector in the PodSpec using a key-value pair. With nodeSelector, users can define label-key value pairs in nodes and use these labels to match when scheduling pods. The most basic approach to scheduling is through the nodeSelector available in Kubernetes since version 1.0.

Kube-scheduler is the default scheduler in K8s, but administrators can use custom schedulers, too. To understand Kubernetes node affinity, you first need to understand the basics of Kubernetes scheduling created to automate the process of pod placement. Kubernetes scheduling is the process of selecting a suitable node to run pods. This article will review scheduling basics, Kubernetes node affinity and anti-affinity, pod affinity and anti-affinity, and provide practical examples to help you get comfortable using this cluster scheduling feature. Kubernetes node affinity is an advanced scheduling feature that helps administrators optimize the distribution of pods across a cluster. How pods are distributed across nodes directly impacts performance and resource utilization. Pod scheduling is one of the most important aspects of Kubernetes cluster management. Kubernetes node affinity: examples & instructions

Affinity one