In Kubernetes, controlling where a pod runs is essential for optimizing performance, ensuring high availability, and meeting specific hardware or compliance requirements. You can achieve this using nodeSelector<\/strong> and nodeAffinity<\/strong>.<\/p>\n\n\n\n In this blog, we will cover:<\/p>\n\n\n\n The To schedule a pod only on nodes with an SSD disk type, you first label a node:<\/p>\n\n\n\n Now, create a pod specification using \ud83d\udd39 Limitations:<\/strong> \ud83d\udd39 Key Benefits of nodeAffinity:<\/strong><\/p>\n\n\n\n A machine learning application requires a GPU for training models efficiently. We must ensure that ML workloads run only on nodes equipped with GPUs.<\/p>\n\n\n\n Pod scheduling in Kubernetes allows precise control over where workloads run. Whether using nodeSelector<\/strong> for simple cases or nodeAffinity<\/strong> for advanced scheduling, these methods help optimize performance and ensure efficient resource utilization.<\/p>\n\n\n\n Would you like a guide on using taints and tolerations alongside affinity? Let us know in the comments!<\/p>\n","protected":false},"excerpt":{"rendered":" Introduction In Kubernetes, controlling where a pod runs is essential for optimizing performance, ensuring high availability, and meeting specific hardware or compliance requirements. You can achieve this using nodeSelector and nodeAffinity. In this blog, we will cover: Using nodeSelector for Simple Scheduling The nodeSelector field allows you to specify a key-value pair that a node […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[219,9,216,232,218],"tags":[536,537,538],"class_list":["post-4079","post","type-post","status-publish","format-standard","hentry","category-deployment","category-kubernetes","category-kubernetes-basic","category-namespaces","category-pod","tag-nodeaffinity","tag-pod-on-a-specific-node","tag-pod-scheduling"],"yoast_head":"\n\n
Using nodeSelector for Simple Scheduling<\/h2>\n\n\n\n
nodeSelector<\/code> field allows you to specify a key-value pair that a node must have for the pod to be scheduled on it.<\/p>\n\n\n\nExample: Scheduling a Pod on an SSD-backed Node<\/strong><\/h3>\n\n\n\n
kubectl label nodes node-1 disktype=ssd<\/code><\/pre>\n\n\n\nnodeSelector<\/code>:<\/p>\n\n\n\napiVersion: v1\nkind: Pod\nmetadata:\n name: specific-node-pod\nspec:\n nodeSelector:\n disktype: ssd\n containers:\n - name: my-container\n image: nginx<\/code><\/pre>\n\n\n\nnodeSelector<\/code> provides a simple key-value matching but lacks advanced filtering capabilities.<\/p>\n\n\n\nUsing nodeAffinity for More Advanced Scheduling<\/h2>\n\n\n\n
nodeAffinity<\/code> offers more flexibility, allowing complex rules and conditions.<\/p>\n\n\n\nExample: Using nodeAffinity for Scheduling<\/strong><\/h3>\n\n\n\n
apiVersion: v1\nkind: Pod\nmetadata:\n name: affinity-pod\nspec:\n affinity:\n nodeAffinity:\n requiredDuringSchedulingIgnoredDuringExecution:\n nodeSelectorTerms:\n - matchExpressions:\n - key: disktype\n operator: In\n values:\n - ssd\n containers:\n - name: my-container\n image: nginx<\/code><\/pre>\n\n\n\n\n
In<\/code>, NotIn<\/code>, Exists<\/code>, etc.<\/li>\n\n\n\nReal-World Scenario: Scheduling Workloads Based on GPU Availability<\/h2>\n\n\n\n
Problem:<\/strong><\/h3>\n\n\n\n
Solution:<\/strong><\/h3>\n\n\n\n
\n
kubectl label nodes gpu-node gpu=true<\/code><\/pre>\n\n\n\n\n
apiVersion: apps\/v1\nkind: Deployment\nmetadata:\n name: ml-workload\nspec:\n replicas: 2\n template:\n metadata:\n labels:\n app: ml-app\n spec:\n affinity:\n nodeAffinity:\n requiredDuringSchedulingIgnoredDuringExecution:\n nodeSelectorTerms:\n - matchExpressions:\n - key: gpu\n operator: In\n values:\n - \"true\"\n containers:\n - name: ml-container\n image: tensorflow\/tensorflow:latest-gpu<\/code><\/pre>\n\n\n\nBest Practices for Pod Scheduling<\/h2>\n\n\n\n
\n
ssd=true<\/code>, gpu=true<\/code>).<\/li>\n\n\n\nkubectl describe node <node-name><\/code> to verify labels before scheduling.<\/li>\n<\/ul>\n\n\n\nConclusion<\/h2>\n\n\n\n