k8s
Kubernetes > Create a cluster on Google cloud platform
Launch a Kubernetes Cluster using Kubeadm
Kubernetes > Pods > Node Selector > Labels
Kubernetes > Pods > Containers > Resources > Memory
Kubernetes > Pods > Containers > Resources > CPU
Kubernetes > Pods > Volumes > Empty Dir
Kubernetes > Pods > Volumes > GCE Persistent Disk
Kubernetes > HorizontalPodAutoscaler
Some notes:
Kubernetes has 2 different layers: 1. Master, 2. Nodes
Master is responsible for scheduling, provisioning, controlling and exposing the API to the clients which can come via API, UI or CLI
Kubernetes understands declarative artifacts which are defined using YAML. These YAML definitions are submitted to the master.
Kubernetes is like a distributed computing architecture with a master/slave kind of a scenario. Nodes form the cluster.
The registry is where the docker images are going to be centrally stored. This can be a private registry or a public registry like docker hub, google container registry etc.
There are 3 main components in the master: 1. API server, which exposes various APIs, in-fact Kubernetes is highly API centered, 2. Scheduler, 3. Controller
etcd – lightweight key value database, central database to store the current cluster state, single source of truth
Kubectl is a tiny golang binary
Every pod is completely autonomous and independent
Alias:
pod(s) = po
replicaset(s) = rs
namespace(s) = ns
service(s) = svc
apiVersion:
Pod: v1
ServiceAccount: v1
ReplicaSet: apps/v1
HorizontalPodAutoScaler: autoscaling/v1
--end-of-post--