Debug School: rathi

K8-POD/Deploy/Service)

rathi — Wed, 19 Oct 2022 09:16:39 +0000

*POD *
POD is atomic unit of scheduling in K8 cluster which is instantiated by Kubelet
POD is an api resource in k8 cluster which can be created or deleted.
It is application or service.

Characteristics of pod
POD can contain one or more containers.
All pods can communicated with other pods in same cluster.
Same POD cannot be redeployed.

Deployment and its feature
Deployment is K8 resource kind.
Features
Deployment can be rolled out or rolled back.
Deployment can be versioned.
Deployment can be configured by replicas.
Deployment can be scaled up or down
Deployments can be created/modified or deleted.
Deployment strategy can be recreate or rollingupdate

Services and its types
Service is a load balancer and its types can be network or application.
Service load balances pod/network level.

Kubernetes Architecture and its components.

rathi — Tue, 18 Oct 2022 06:33:00 +0000

Kubernetes Architecture
1)K8 Architecture consists of Master(s) and Worker Nodes. Worker node can have multiple pods.
2)Worker Node contains PODs and PODS contain one or more containers.
3)Deployment Instruction is received by Master and it is in charge to instantiate the pod.
4)Master of Linux can be combined with worker nodes of windows too (but with recent version of windows) (Doubt)

Component of master and its role for each
1)Kubi-APIServer - Receives all instructions in the form of JSON.(REST)
2)Cluster Store . Key Value pair. Cluster information is persisted in this. powered by etcd.
3)Kube Controller manager - Controllers different type of controllers. Watches and maintains desired state.
4)Kube Scheduler - Watches APIServer for new pod requests.

Component of worker and its role for each....
1)Kubelet - Main agent in worker node. Registers node with the cluster. Instantiates pod. Reports to master.
Responsible for pod lifecycle.(Pending-running-Success).
No same pod can be received. But Similar pod can be maintained.
2)Container Engine - Pulls images. Usually docker. Recently it is rkt.
3)Kube Proxy - It takes care of networking relevant requirements of the newly created pod.

Problems solved by K8

rathi — Tue, 18 Oct 2022 04:42:17 +0000

1)Scalability challenges due to container based deployment.
2)Network Load Balancing issue due to container applications running on same port.
3)Auto discovery of new services.
4)Manages docker host(s) at any scale.
5)Workload placement - Placement of containers on appropriate clusters or nodes.
6)Auto Scale up or Scale down of container services.
7)Configuration Management and Secret managemen for larger cluster of container applications.
8)Self Healing capability.
9)Container monitoring capability and manages health of the cluster.
10)It also works well to manage micro service based applications , where applications need to talk to each other.
11)Avoids human intervention.
12)Reduces time and effort and cost and provides service to the customer with high performance.

Rathipriya Assigment - Docker Training Day1

rathi — Mon, 17 Oct 2022 11:27:17 +0000

** Difference between docker kill and docker stop?**

"docker stop" command stops all processes gracefully whereas docker kills the processes abruptly.

It is also noticed that docker kill works on running containers only. 
If it is executed on non running containers then below error is seen.
root@ip-172-31-22-5:/home/ubuntu# docker stop rathi
rathi
root@ip-172-31-22-5:/home/ubuntu# docker kill rathi
Error response from daemon: Cannot kill container: rathi: Container 5cbe35371e6122791a5504d6e4315241cc423895fc77580132a55663d8b108e9 is not running

docker pause and docker unpause difference
Docker pause puts all the threads of specified container threads to suspended state.
_Docker unpaus_e resumes all the threads of the specified container to running state.
**
URL of the images pushed**
https://hub.docker.com/repository/registry-1.docker.io/rathipriya28052020/rathidocimg/tags?page=1&ordering=last_updated