[EN] Kubernetes Notes - Part 1

Saturday. October 09, 2021 - 3 mins

Some kubernetes studies notes 🎉

Kubernetes

To provide portable and isolated virtual environment
Need to be managed and connected to external world
Must be scheduled, distribuited and load balanced and the data persisted somewhere

“Kubernetes” is the Greek word which means helmsman or ship pilot. With this analogy in mind, we can think of Kubernetes as the pilot on a ship of containers.
Referred to as k8s
The Kubernetes project focuses on building a robust platform (a fault-tolerant and scalable solution) for running thousands of containers in production.
Can be achieved by creating a single controller/management unit, after connecting multiple nodes together. This controller/management unit is generally referred to as a container orchestrator.

Automatic bin packing
Self-healing: Kubernetes automatically replaces and reschedules containers from failed nodes
Horizontal scaling: scaled manually or automatically based on CPU or custom metrics utilization.
Service discovery and Load balancing
Automated rollouts and rollbacks: to prevent any downtime
Secret and configuration management
Storage orchestration
Batch execution: supports batch execution, long-running jobs, and replaces failed containers.

Can be deployed in many environments such as local or remote Virtual Machines, bare metal, or in public/private/hybrid/multi-cloud setups
Kubernetes’ architecture is modular and pluggable
Kubernetes’ functionality can be extended

Node: or workers, it runs proccess. It contains the app, proxy and kubelet, it can has one or more pods.
Kubelet (Node Agent): responsible to inform to control plane how many cpu, memory the node has.
Control plane: virtual machine, it run a lot of services, such as scheduler service, API server, node controller and state storage (etcd).
etcd: it is a distributed key-value store which only holds cluster state related data, no client workload data.
Kubectl: Kubernetes CLI, point of control for all apps based in Kubernetes, based in configuration files.
Workloads: everthing you can put inside of kubernetes that generate a work rules, like apps and DNS. The workloads are:
- Pod
- Deployment
- ReplicaSet
- Secret
- Configmap
- HorizontalPodAutoscaler
- Service
- Ingress
- …
Pods: the smallest unit in Kubernetes, you can run one or more containers. Each pod contains an internal IP, it runs the application in the container and it can has or not a volume. Volume is a pre-existing data structure to use in the pod to share files or to use a temporary files.
Labels: as “nametags” to identify things, it can be used to query based on the labels.
Deployments: deployments manages and adds “intelligence” to pods, and it can group pods by an label. Also, you can define the number of replicas inside of the deployment

Deployment

Deployment File example

Replicaset: it created by a deployment. It manage the number of replicas inside of the cluster. Each one contains a different image version, it is able to create versions of images. It always keep the desired number state.
Replication Controllers: responsible to manage the numbers of pods, pods’ lifecycle, scaling up and down, rolling deployments, and monitoring.
Services: Responsible to tell to the rest of the Kubernetes environment what services your application provides, the service IP addresses and ports remain the same.
Volume: Location where containers can access and store information.
Namespaces: group of services, pods, replication controllers, and volumes.

Some Kubernetes terms

Software Engineer