6.

1) Kubernetes Detailed notes

Topic:Overview of Containers :-

Why do we need Kubernetes

Question or Concerns about containers

● Containers are a wonderful way of bundling and running your applications, But are they
production ready?
● What would happen if the container or the Docker Host goes down?
● How to make containers available 24*7 ?
● How to handle loads during peak time for the applications ?
● How to replace containers without having downtime with new Docker Image based
containers ?
● How to monitor containers?

Solution to above Questions or Concerns

Wouldn’t it be good if there is some system which can help for handling all the
questions/concerns raised in the above section. That is exactly what Kubernetes does.

● Kubernetes takes care of

○ Scaling requirements
○ failover
○ deployment patterns
 ● Kubernetes Provides
○ Service Discovery & Load Balancing
○ Storage Orchestration
○ Automated rollouts and rollbacks
○ Automated bin packing
○ Self-Healing
○ Secret & Configuration Management
● Kubernetes is not only for the open source community embraced containers, It is deeply
embraced by the Cloud Providers.
○ Amazon Web Service offers Elastic Kubernetes Services(EKS)
○ Google Cloud platform offers Google Kubernetes Engine(GKE)
○ Microsoft Azure offers Azure Kubernetes Services(AKS)

Topic:What is Kubernetes:-
● Kubernetes is a platform that manages container-based applications, their networking
and storage components.
● In Kubernetes, we focus on the application workloads rather than the underlying
infrastructure.
● Kubernetes provides a declarative approach to deployments, backed by a rich set of APIs
for management operations.
● Cluster: Cluster is a collection of compute, storage and networking resources that
Kubernetes uses to run workloads.
● Node: It is a single host. Now we can put the cluster as a collection of nodes.

Kubernetes has two kinds of Nodes

● Master:
○ Provides core Kubernetes Services and orchestration to application workloads
● Node:
○ run your application workloads

Kubernetes Cluster Architecture

What would be our approach with workloads in Kubernetes?

● We would be describing our application as a YAML manifest and pass it to Kubernetes

master from some client, and Cluster Master does the rest.
● k8s is a platform that is used with a huge number of services & capabilities that keeps
growing.
● The core functionality of k8s is scheduled workloads in containers across your infra. In
addition to this there are many functionalities that k8s brings in
○ Managing storage systems
○ Secret Management
○ Application health & readiness check
○ Replicating application instances
○ Horizontal pod Autoscaling
○ Cluster Autoscaling
○ Naming & Service Discovery
○ Balancing loads
○ Rolling updates
○ Resource Monitoring
○ Log Management
○ Application debugging
○ Authentication & Authorization
Topic:k8s components:-
https://kubernetes.io/docs/concepts/overview/components/

Kubernetes Master
Cluster:
● It is collection of hosts(nodes) that provide compute, memory, storage and networking
resources
● k8s uses these resources to run application workloads.

Master Node:
● The master node is a single host in k8s cluster.
● This can be a physical server or virtual machine
● This is the control plane of k8s. This master is responsible for assigning work (container
workloads) to be scheduled on nodes
● Kubernetes master runs on Linux node
● Master Components has i) kube-apiserver ii) etct iii) kube-scheduler iv) Controller
manager v) Cloud-controller-manager

kube-apiserver
● This component is Central to Kubernetes. All communications between all components
goes through the kube-apiserver
● This component is frontend of the Kubernetes control plane.
● This component exposes a REST API.
● We would interact with this component using kubectl by using the YAML files, which
are also referred as manifests

etcd
● etcd stores the entire configuration and the state of the cluster.
● etcd is a consistent and highly available distributed key-value store.

kube-scheduler
● It watches for new work tasks and assigns them to healthy nodes in the cluster

Controller-Manager
● It is responsible for maintaining desired states mentioned in the manifest.
● It looks like single component, but with in it has
○ Node Controller: for noticing & responding when node goes down
 ○ Replication Controller: for maintaining the correct number of pods for every
replication controller object.
○ Endpoints Controller: Populates the Endpoints object

Cloud-Controller-manager
● It is responsible for underlying cloud specific controllers.

● If you run the Kubernetes on a supported cloud platform such as AWS, Azure or Google,
your control plane runs the Cloud-Controller-Manager.

Node Components
Nodes:
● It is a single host in k8s cluster
● This can be physical server or virtual machine
 ● These are managed by k8s master and they do the work of executing workloads in k8s
(They were referred as minions)
● Node is responsible for doing the work assigned.

kubelet

● This is an agent which runs on each node in the cluster.

● It watches for the instructions from API Server for new work assignments.
● If it can’t run the task assigned, it reports back to the master and lets the control plane
decide on the actions.
● It is responsible for the node registration process

Container Runtime
● This is software which is responsible for running containers.
● Some of them are Docker, containerd, cri-o, rktlet

kube-proxy
● Maintains the network rules on nodes
● This is responsible for networking on nodes.

Cluster DNS
● Every Kubernetes Cluster has an internal DNS service
● This has static IP address that is hardcoded into every Pod on the cluster i.e. all Pods
now how to find the DNS Server
● Services, Stateful Sets and Pods are registered with Cluster DNS.
● Cluster DNS is based on CoreDNS
Interesting ADD ONS

Web UI (Dashboard)
 ● Web based Kubernetes user interface
● Capabilities:
○ Deployment
○ View cluster resources
○ Troubleshooting

Container Resource Monitoring

● We will discuss about it in next series of articles

Cluster Level Logging

● We will discuss about it in next series of articles

How to Package Apps To Run on Kubernetes

 ● Package your application as a container image.
○ We would create a Image for the application
○ Push the image to Docker Registry
● Pod will Wrap container
● Pod is deployed into the Kubernetes cluster via Declarative Manifest file.

Pods:-
● A pod is a unit of work in k8s.
● Each pod contains one or more containers.
● Every pod gets a unique ip address and the containers in the pod have the same ip
address.

Topic: K8s Installations:-

● Refer Here for devops blog installation steps
● Refer Here for official documentation
● Install docker and kubeadm,kubectl on all the three servers as mentioned in above
documentation. Now login into master and execute as Refereed Over Here
● We need to execute “kubeadm init” it will generate a script,we need to follow according
to that and need to create a folder for k8s and configure the network driver for that.
● Now login to kubernetes master and execute “kubeadm init”
● Make node kubeadm init commands login as normal user and create .kube folder and
config as below
Your Kubernetes control-plane has initialized successfully!

To start using your cluster, you need to run the following as a regular user:

mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

Alternatively, if you are the root user, you can run:

export KUBECONFIG=/etc/kubernetes/admin.conf

You should now deploy a pod network to the cluster.

Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
https://kubernetes.io/docs/concepts/cluster-administration/addons/

Then you can join any number of worker nodes by running the following on each as root:

kubeadm join 172.31.41.237:6443 --token 9g3h4c.oxuxjih0ee4f98g6 \

--discovery-token-ca-cert-hash
sha256:95646829cf695d5829e9391e1df0453041d797800118f2c6b344e16c3fce041a

● As mentioned above if we execute “kubeadm init” it will display like above we need to
execute below commands as normal users

mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

● After that as mentioned in the lines go to the url. This url contain network drivers and
that drivers installation commands
{https://kubernetes.io/docs/concepts/cluster-administration/addons/}
● Generally we have network drivers like Flannel,kniter,Romana..etc in above we have
selected weave-network
● I wish to install “weave-network” so we will go to the installation they will give the
command
● For install weave-wetwork -network driver we need to execute {kubectl apply -f
"https://cloud.weave.works/k8s/net?k8s-version=$(kubectl version | base64 | tr -d '\n')"}
● Then join the worker node by following command for ex

{kubeadm join 172.31.78.198:6443 --token qbhg9d.v46u4elnoi65bzpp \

--discovery-token-ca-cert-hash
sha256:ba79a470521e5f13c4460e7def9a23936890c6bb8a725d86b088d83b334dbd44}
 ● Now join two nodes to cluster by executing kubeadm join command shown above
● After joining nodes, login into master and execute kubectl get nodes

Topic: Kubernetes commands:-

1) Kubectl version (or) kubectl version -o json {to know the version of k8s}
2) kubectl apply -f <yamal file> {To create a POD}
3) kubectl get pods {To know number of pods running}

4) kubectl get pods -o wide {It will display pod clearly}

5) kubectl describe pod gol-pod {It will describe entire pod details}
 6) kubectl port-forward --address 0.0.0.0 <pod-name> source port(the port where the pod is
running):destination port(port where we have to expose the pod)

NOTE:- To know the pod is running on which port etc details execute
“kubectl describe <pod-name>” . When ever we do port-forward, default it will try to map on
loopback adapter which will work when we are in local machine but whenever you are
working on remote machines especially in cloud don’t forget to put ‘0.0.0.0’

7) kubectl logs <pod-name> {to view the logs of the pod}

8) kubectl exec <pod-name> -- command {to execute commands in pod using exec }

9) kubectl exec -it <pod-name> -- /bin/bash {TO get inside the docker container }

10) kubectl top nodes {to see the metrics of nodes} && kubectl top pods {to see the
metrics of pods}
11) kubectl get rs {to view the replica-set}

12) kubectl describe rs replication-set {to describe the replication set}

13) kubectl get rs -o wide {to see rs in detailed}

14) kubectl get pods -o wide {to display pods in wide}

15) kubectl get rs -w {to watch the replication set like what is happening}

16) kubectl scale replicaset <rs-name> --replicas=<number>

17) kubectl autoscale rs <name of replicaset> --min=<number> --max=<number>

--cpu-percent=<percentage> {to configure replicaset for auto scaling using imperative
commands}

18) kubectl get hpa {horizontal pod autoscaler}

19) kubectl get pods --show-labels {To view names of the labels}

20) kubectl get pods --selector=”app=<name>” {to view the particular pods with selected
name/label}

21) Kubectl get pods --selectors in (appname,app-name)

22) kubectl api-resources {To know available api-resources}

23) kubectl get svc {to get service object nodeport, etc}

24) kubectl get namespaces (or) kubectl get ns {To view all name-spaces}
25) kubectl get sts {to get statefuset}

26) kubectl get ds {to get demonset}

27) kubectl get deploy {to get deployment set}

28) Kubectl describe rs <name of rs> {to describe replication set}

Topic:concepts in kubernetes:-
kubectl
● This is a command-line utility which we would be using to interact with
k8s cluster.
● This utility creates api-request internally and forwards them to kube-api
server
● Refer Here for kubectl cheat sheet.

Kubernetes Objects
● K8s object is anything which can be stored in etcd, As a key-value pair.
● k8s objects are persistent entities in the k8s system.
● These objects are used to represent the state of your cluster
● Every k8s object has
 ○ Specification: This is also referred as object Spec
○ Status: This is also referred as object status
● How can we describe a k8s object (Specification):
○ K8s objects can be describe in the json format using k8s rest api or in yaml format
via kubectl
○ For api version Refer Here
○ Every k8s object can be described using .yaml (specification) or json when using a
rest api and passed across to the kube-api server via kubectl or k8s REST API.
The cluster responds back with status
○ While describing object spec, we will have some required fields
■ apiVersion: Which version of k8s api are you using to create this object
■ kind: what kind of k8s object do you want to create
■ metadata: Data which is helpful in uniquely identifying the object, here we
will be have name, UID and optional namespace
■ spec: What is your desired state

K8s Workloads

● Workload is an application running on k8s cluster.

● A Pod has a defined lifecycle. If the pod fails, we need to create the new
pod running the same application, K8s has workload resources which can
maintain desired state of pods
● Workload resources are
○ Deployments
○ Replica Sets
○ Stateful Sets
○ Daemon Sets
○ Job and Cron Job

The Pod Manifest (pod yaml file)

● Pods are described in a Pod Manifest which is a text-file representation of k8s API
Object.
● Creating a Pod in a imperative kubectl run command
 ● Creating a Pod using a manifest file.
● Refer Here for the pod manifest

name: gameoflife
spec:
containers:
- image: qualitythought.in/gameoflife:07112020
- image: qualitythought/gameoflife:07112020
name: gol
ports:
- containerPort: 8080
 ● Accessing your pod: {if we can observe the pod ip address and node ip address is
different also those are in two different networks. So to access that pod we need to map
the ip to port “kubectl port-forward --address 0.0.0.0 gameoflife 8080:8080”}

* Logs information of your pod:

 ● Running commands in your container with exec

K8s Pods
● Pods are the atomic unit of computing that you create in k8s
● A pod represents an application/microservice running in k8s cluster.
● Pods will have containers in it.
 ● Each container within a pod runs its own cgroup, but they share a number of Linux
namespaces like network, mnt, UTS, IPC, etc.
● These namespaces are for pods. Not for containers.

In k8s to create any object we have two ways:-

Imperative:
● We can create objects by executing commands Refer Here

Declarative:

● We create an object specification in a yaml file.

 ● Send/apply this yaml which has a specification & k8s master will do the rest to
create/maintain the specification sent
● To create this specification we will be using k8s api reference

Resource Management:
● Resource Requests for Minimum required resources and maximum limits Refer Here
● https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/
● Refer Here for resource limits in a pod.

containers:
- image: qualitythought/gameoflife:07112020
name: gol
resources:
requests:
cpu: "500m"
memory: "128Mi"
limits:
cpu: "1000m"
memory: "512Mi"
ports:
- containerPort: 8080
name: golhttpport
protocol: TCP
protocol: TCP

● Refer Here for multi container pod

---
apiVersion: v1
kind: Pod
metadata:
name: gameoflife
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol
resources:
requests:
cpu: "500m"
memory: "128Mi"
limits:
cpu: "1000m"
 memory: "512Mi"
ports:
- containerPort: 8080
name: golhttpport
protocol: TCP
- image: httpd
name: httpd
ports:
- containerPort: 80
name: httpport
protocol: TCP

K8s API Groups

● K8s groups api’s as api groups.
● The major purpose is to extend k8s
● Refer Here for api groups
● To use resources from api groups the naming convention is
<GROUP_NAME>/<VERSION> in yaml file
○ batch/v1
○ extensions/v1beta1
○ rbac.authorization.k8s.io/v1
● To use resource from api groups in the REST api, convention is
/apis/<GROUP_NAME>/<VERSION>
○ /apis/batch/v1
○ /apis/extensions/v1beta1
○ /apis/rbac.authorization.k8s.io/v1
● If the resources are part of core group we can directly write the version apiVersion:
v1

Topic:- K8s Service

● We have created a replica set with n pods. How do we access the application running
inside the pods.
● In normal scenarios for these kinds of problems, we use a load balancer.
● In K8s we use Service Object. A Service is a way to create a named label selector. This
Service Object will recieve an IP Address and also DNS Name
 ● So now let's create two replica-sets and two services.
● Lets understand service types
a. Cluster IP
b. Node Port
c. load balancer
d. external name

EX:-

● now we create 4 objects 2 for replica set in dev environment and test environment. And
2 for services like dev environment and test environment.

---
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: gol-dev
spec:
minReadySeconds: 10
replicas: 2
selector:
matchLabels:
app: gol
env: dev
 template:
metadata:
labels:
app: gol
env: dev
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol-dev
ports:
- containerPort: 8080
protocol: TCP
################################################################

---
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: jenkins-qa
spec:
minReadySeconds: 10
replicas: 2
selector:
matchLabels:
app: jenkins
env: qa
template:
metadata:
labels:
app: jenkins
env: qa
spec:
containers:
- image: jenkins:2.60.3
name: jenkins-qa
ports:
- containerPort: 8080
protocol: TCP

#################################################################
 ● In the below example in the field of ports.

port belongs to ‘port of your service’

targetPort belongs to ‘port of your pod’
nodePort is ‘port of your node. By using which we will access this service’

---
apiVersion: v1
kind: Service
metadata:
name: gol-dev
spec:
type: NodePort
selector:
app: gol
env: dev
ports:
- targetPort: 8080
port: 8080
nodePort: 32766

##########################################################

---
apiVersion: v1
kind: Service
metadata:
name: jenkins-test
spec:
selector:
app: jenkins
env: qa
type: NodePort
ports:
- targetPort: 8080
port: 8080
nodePort: 32765
#############################################################
Topic: Kubernetes Networking Model (KNM)
● KNM requirements are
○ all pods can communicate with all other pods without using NAT
○ all Nodes can communicate with all Pods without NAT
○ the IP that Pod sees itself is the same IP that other see it as
● Lets understand the following
○ Container to Container networking
○ Pod-to-Pod Networking
○ Pod-to-svc networking
○ Internet to SVC networking
Container to Container Networking

Pod to Pod Networking

● Both Pods in Same Node, will communicate like below one container in pod1 is
communicating to another container in pod2. For that node it has a root network, each
pod has its own network namespace.
● Because of this network namespace each of this pod will get ethernet interface like
‘eth0,eth0’
● This ‘etc’ is connected to virtual-eather net device ‘veth0’
● These all virtual ethernets which are created in our node n/w name spaces are connected
to each other via a bridge ‘cbr0’
● Now we need to send a pocket from pod1 to pod2
● Now what will happen is pocket will go to eth interface and reached to pod1’ eth0
● Your eth0 is connected to veth0 and it will reach over there and forward to ‘cbr0’
● Then the container bridge will send an ‘arp’ packet to all other networks in that node
except source n/w.
● ‘veth1’ will know that this is pod2’s ip address which we are trying to get and it will
responds back to container bridge
● So this bridge source will send an ‘ARP’ Address resolution protocol to all other
networks and any network will respond to that then it will take that and will give to
source

● Pods on different nodes, as shown in the below pic, pod1 was trying to communicate
with pod4 in node-2.
● For that it will send a packet to eth0 and request will reach to ‘container bridge’ then the
bridge will call out for destination ip. If it doesn’t get any response from its own node it
will communicate to the external network.
● Basically all nodes are connected via a network. And all the node ethernet will consist of
its container's CIDR range. So that it is easy to find out ip of nodes.
● In that pool the matched CIDR node will respond and the packet will reach the
destination node. The node-bridge will take that packet to destination pod in this case
pod4
● Pod to Service IP Address , Hear pod1 wants to communicate to service1
● As same as above concepts the packet is reached to eth0 → cb0→ here it will
communicate to ‘iptables’ entries
● Because service is an iptable entry that's why service doesn’t fail. So here this entry will
tell that svc1 means pod4
● Then the packet will reach to destination
 ● From external world to service, request comes to eth0 then it will check in iptables and
as we know all service entries are in iptables and then the request is forwarded to that
particular service

Health Checks of Pod

● When we run our application in k8s, it is kept alive using process health checks.
● However in most cases a simple process might not be enough
○ process has deadlocked and unable to serve requests (process check will return as
application is health)
● To address this, k8s introduced health checks for application liveness
● Liveness Probe: This probe checks for application specific logic to verify that the
application is still running
● Containers that fail liveness probes will be restarted.
● Readiness probe describes when a container is ready to serve user requests. Containers
that fail readiness check are removed from service (load balancers)
● Refer Here for the liveness probe.

---
apiVersion: v1
kind: Pod
metadata:
name: gameoflife
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol
livenessProbe:
 httpGet:
path: /gameoflife
port: 8080
initialDelaySeconds: 5
timeoutSeconds: 1
periodSeconds: 10
failureThreshold: 3
resources:
requests:
cpu: "500m"
memory: "128Mi"
limits:
cpu: "1000m"
memory: "512Mi"
ports:
- containerPort: 8080
name: golhttpport
protocol: TCP

Topic:- Activities
Activity: 1:-
Lets create a pod with the below yaml file.

---
apiVersion: v1
kind: Pod
metadata:
name: gol-pod
spec:
containers:
- name: my-gol-cont
image: qualitythought/gameoflife:07112020
ports:
- containerPort: 8080
Activity:2:-

Topic: k8s yaml file

---
apiVersion: v1
kind: Pod
metadata:
name: gol-pod
spec:
containers:
- name: my-gol-cont
image: qualitythought/gameoflife:07112020
ports:
- containerPort: 8080
Topic:Azure Kubernetes Services (AKS)
● AKS is a managed k8s cluster in Azure.
● Azure manages master node
● Managed vs UnManage k8s cluster
 ● Lets install AKS Cluster. AKS Cluster can be installed using
○ Portal Refer Here
{https://docs.microsoft.com/en-us/azure/aks/kubernetes-walkthrough-
portal}
○ CLI Refer Here
{https://docs.microsoft.com/en-us/azure/aks/kubernetes-walkthrough
}
○ Powershell Refer Here
{https://docs.microsoft.com/en-us/azure/aks/kubernetes-walkthrough-
powershell}
● For installing Azure CLI/Powershell Refer Here

Topic:Amazon Elastic K8s Service (EKS)

● Lets create a ubuntu ec2 instance
● Install aws cli in the ubuntu
● Refer Here for setting up eks cluster
{https://docs.aws.amazon.com/eks/latest/userguide/getting-started.html}
● Install eks ctl by following docs Refer Here
{https://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.html}
● After installing the eks, aws cli and kubectl commands, create k8s cluster by executing
the following command
{
eksctl create cluster \
--name my-cluster \
--version 1.18 \
--region us-west-2 \
--nodegroup-name linux-nodes \
--nodes 1 \
--nodes-min 1 \
--nodes-max 2 \
--with-oidc
--ssh-access
--ssh-public-key dockerclassroom
--managed
}

1) Create a vm and install aws cli on that

2) Create a user callead k8s in IAM and give him admin previlises
3) Configure aws cli with the provision of k8s keys
4) Install EKS by aws official documentation
5) Install Kubectl by using official documents
6) Configure managed nodes using docs

Topic: Namespaces
● K8s is one physical cluster with k8s master and cluster nodes & in this cluster if you
want to create multiple virtual clusters (like one for dev, staging, qa..) we can use
namespaces.
● Execute kubectl api-resources. The resource with value namespaced
 ○ true => will be created per namespace {It means when we create two namespaces
that will be different. }
○ false => will be same/shared across namespaces {if it is false across name-spaces
or virtual clusters you can’t have different name spaces or nodes. Nodes are same
across virtual clusters}

● Generally kubectl command will pull the values from namespace default. If you want
k8s to fetch the information from other namespaces use –namespace flag
NOTE:- If we need to k8s will pull the values from some particular name-space, Then you can
set something called as context (context is nothing but some values stored in kubectl config file
~/.kube/config)

Upgrading
● Upgrading to published versions is easier in the cloud.
○ In AKS, we can upgrade the nodepools and k8s seperately
○ In the case of eksctl
Topic: Storage in K8s
● Both containers & Pods are ephemeral (when we delete the Pods/Containers data
generated is deleted)
● Docker Volume separates storage from the life cycle of Container, So even if the
container is deleted, data will still be available.
● Since Docker containers are created in the Pod, and the problem of storage has to be
solved at Pod level in k8s as Pod provides kernel namespace to docker container.
● Pod has a container. The container in the pod stores the data and let's assume it is
crashes, Pod will try to restart the container and in this case data will be lost. The new
container will start with empty disk space allocated.

● We cannot rely on containers themselves even for temporary storage of data.

● K8s Volume abstraction solves the problem

● Volume is exposed to the applications which eventually store data on any type of storage
(node/cloud).
● Lifetime of k8s volume is the same as the Pod’s lifetime. As long as Pod is alive the
containers within POD will get the same volume across restarts or new container
Creation.
● How to use K8s Volume
 ○ Volume is defined in Pod Spec
○ Apply the k8s pod manifest with empty volume

---
apiVersion: v1
kind: Pod
metadata:
name: pod-with-emptydir
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol-container
volumeMounts:
- mountPath: /tmp
name: tmp-volume
volumes:
- name: tmp-volume
emptyDir: {}
 ● Refer Here for official docs
{https://kubernetes.io/docs/concepts/storage/volumes/}
● Lets use one more example of hostPath (Type of volume) Refer Here
---
apiVersion: v1
kind: Pod
metadata:
name: pod-with-hostpath
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol-container
volumeMounts:
- mountPath: /tmp
name: tmp-volume-host
volumes:
- name: tmp-volume-host
hostPath:
path: /tmp
type: Directory
● Volumes such as emptyDir and hostPath get deleted when the pod using them is deleted
or the pod is restarted. So we need to find a solution Where the data/storage lifetime is
different than Pod’s lifetime
● To Solve this Problem, K8s has Persistent Storage in the form of Persistent Volume
(PV).
Topic: Persistent Volumes & Persistent Volume Claims
● A PV is a K8s object that represents a block of storage in the cluster. This can be
provisioned before hand by cluster admins or be dynamically provisioned.
● In Order to use PV, a Persistent Volume Claim (PVC) needs to be created. A PVC is a
request for storage by user or by Pod. This request can be of specific size or specific
access mode.
 ● Each PV belongs to a certain storage class. A Storage Class is a K8s Object that provides
a way for administrators to describe different types of Storage.
● Volume mode has two Values Block, FileSystem
● Access Modes:
○ ReadWriteOnce (RWO): Mounted as read-write by a single node
○ ReadOnlyMany (ROX): Mounted as read-only by many nodes
○ ReadWriteMany (RWX): Mounted as Read-write by many nodes
● persistentVolume Reclaim Policy
○ Retain (Even if u delete your persistent volume the data will still be there )
○ Recycle (when we delete your persistent volume the store will be there but the
data in the storage will be cleaned )
○ Delete (when ever we delete the persistence volume the storage/volume and data
in that volume will be deleted)

PVC in Azure

● Dynamic Provision Azure Disk Refer Here Azure Files Refer Here
● Static Provision Azure Disk Refer Here Azure Files Refer Here
● NFS in Azure Refer Here
● Azure Netapp Refer Here
● AWS Refer Here
● Refer Here for AWS Storage Classes Documentation

Storage Classes in Azure (In azure we have 4 different classes )

PVC in AWS

Storage Class in AWS (we have 1 storage class in aws)

Activity:

● Now let's create a ‘my-sql’ container, my-sql is a database so it require some kind of
storage, so we will create a pod in that we will create a my-sql,
● Lets understand how to use mysql container in a Pod with Persistent Volume (in AWS
and Azure)
● Refer Here to mysql Dockerfile which has volume at /var/lib/mysql

In azure:

● Lets use existing storage classes in Azure to create mysql-pod Refer Here

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: mysql-azure-pvc
spec:
accessModes:
- ReadWriteOnce
storageClassName: managed-premium
resources:
requests:
storage: 1Gi
---
apiVersion: v1
kind: Pod
metadata:
name: mysql-azure
spec:
containers:
- name: mysql-cont
image: mysql:8
volumeMounts:
- mountPath: /var/lib/mysql
 name: mysql-volume
env:
- name: MYSQL_ROOT_PASSWORD
value: 'rootroot'
- name: MYSQL_DATABASE
value: 'openmrs'
- name: MYSQL_USER
value: qtdevops
- name: MYSQL_PASSWORD
value: qtdevops
ports:
- name: mysqlport
containerPort: 3306
protocol: TCP
volumes:
- name: mysql-volume
persistentVolumeClaim:
claimName: mysql-azure-pvc
In AWS:

● for the manifest Refer Here

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: mysql-pvc-aws
spec:
accessModes:
- ReadWriteOnce
storageClassName: gp2
resources:
requests:
storage: 1Gi
---
apiVersion: v1
kind: Pod
metadata:
name: mysql-pvc-aws
spec:
containers:
- name: mysql-cont
image: mysql:8
volumeMounts:
- mountPath: /var/lib/mysql
name: mysql-volume
env:
- name: MYSQL_ROOT_PASSWORD
value: 'rootroot'
- name: MYSQL_DATABASE
value: 'openmrs'
- name: MYSQL_USER
value: qtdevops
- name: MYSQL_PASSWORD
value: qtdevops
ports:
 - name: mysqlport
containerPort: 3306
protocol: TCP
volumes:
- name: mysql-volume
persistentVolumeClaim:
claimName: mysql-pvc-aws

K8s Volume Snapshots & Clones

● K8s Volumesnapshot represents the snapshot of volume on the storage system.
● Source can be specified as PVC Refer Here
● CSI Volume Cloning features add support for existing PVCs to clone the volume. Refer
Here

Topic: Replica set (K8s controller type-1)

● Replica Set will ensure that the number of replicas specified in the spec is maintained by
k8s cluster.
 ● K8s will try to maintain the desired state which is the number of replicas. If some pod
gets deleted/crashed etc, k8s will create one more pod to maintain the desired state.

● Pods might be allocated on different cluster nodes.

● So lets create a replica set with 2 replicas for the game of life. Refer Here for the
changeset

---
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: gol-rs
spec:
minReadySeconds: 5
replicas: 2
template:
metadata:
labels:
app: gol
version: "07112020"
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol
ports:
- containerPort: 8080
● Now lets add selector for the labels

---
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: replocatio-set
spec:
replicas: 2
minReadySeconds: 10
template:
metadata:
labels:
app: gol
version: "1.1"
 spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol
ports:
- containerPort: 8080
selector:
matchLabels:
app: gol
version: "1.1"
● apply the replica set and explore
● kubectl get rs - to view the replica-set
● kubectl describe rs replication-set - to describe the replication set

Now lets shutdown one node and see the result

We can even scale the pods in replica set with out modifying yml file
kubectl scale replica set replication-set --replicas=5
NOTE:- it is not a recommended way

● when ever we need to scale the replicas is in two ways i) chaining the yaml file ii) using
auto scaling imperative commands

EX:- kubectl autoscale rs replocatio-set --min=2 --max=8 --cpu-percent=80

Labels:-
● Basic workflow of creating replicas

● Giving any pod ip address to access the application is not good caz any pod can be killed
any time and any pod can create at any time.
● As k8s is used and application microservices run in the cluster, the number of pods
running of nodes might grow in size and complexity.
● As shown in the below pic orange circled pods are belongs to one service for ex web
server, Black circled pods are belongs to another service for ex app server..etc.,
 ● K8s has added foundational concepts of Labels and annotations to organization, mark
and cross-index all of your resources to represent the groups.
● Labels are key/value pairs that can be attached to k8s objects such as pods and
ReplicaSets. Labels are used by k8s cluster to group pods/other objects
● _Annotations provide storage mechanism similar to labels which information that can be
used by tools & libraries
● Labels provide identifying metadata

---
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: gol-rs-label
spec:
minReadySeconds: 10
replicas: 5
selector:
 matchLabels:
app: gol
template:
metadata:
labels:
app: gol
version: "07112020"
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol
ports:
- containerPort: 8080
protocol: TCP

● Pods can be selected by writing selectors

Some selector examples:

---
selector:
matchLabels:
app: jenkins
---
selector:
matchExpressions:
- {key: app, operator: In, values: ['jenkins', 'gol']}
---
selector:
matchExpressions:
- {key: ver, operator: NotIn, values: ['1.0', '2.0']}

● If you want to all the api-resources in k8s cluster

Service Discovery
● We are able to run our applications in pods
● We can dynamically scale pods by using autoscale command
● How do i access my application which is running in pods
● Service-discovery helps solving the problem of finding which processes are listening at
which address for each service.
● We need to expose our services to the external world. As of now w.r.t networking we
know each pod is having its own ip address.

Topic: StatefulSet (K8s Controller type-2)

● Generally we map same storage disk/location to a application running in different
pos/containers
● But if we each pod/container running for same application can maintain separate
pvc/storage as shown in the below fig then we need stateful-set
● Stateful sets are used to stateful replicas
Stateful sets are used to stateful replicas Refer Here
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: stateful-demo
spec:
replicas: 3
serviceName: jenkins-svc
selector:
matchLabels:
app: jenkins
template:
metadata:
labels:
app: jenkins
version: "2.60.3"
spec:
containers:
- image: jenkins:2.60.3
name: jenkins
ports:
- containerPort: 8080
protocol: TCP
Topic: DaemonSet (K8s Controller type-3)
● Daemonsets are used to manage creation of a Particular Pod on all or selected nodes in
k8s cluster
● Use Cases:
○ Logging agent
○ Monitoring agent
Apply the spec Refer Here
---
apiVersion: apps/v1
kind: DaemonSet
metadata:
name: daemonset-demo
spec:
selector:
matchLabels:
app: alpine
template:
metadata:
labels:
app: alpine
spec:
containers:
- name: alpine-cont
image: alpine
args:
- /bin/sh
- -c
- sleep 1d
Topic: K8S Deployments (K8s Controller type-4)
Deployment is a k8s object that acts as a wrapper around ReplicaSet and makes it easier to use

● The major reason for Deployment is

○ it maintains the history of revisions (i.e. history of your application Deployments)
○ it makes it easy to roll back to previous revisions
○ it gives a way to do zero downtime deployments
● Deployment Strategy:
○ Rolling Update:
■ This give an option to deploy your application without having downtime
■ This gives maxUnavailable, maxSurge
○ Recreate:
■ This option kills all the running pods and recreates new pods
● Refer Here for deployment commands
● Now lets write a sample deployment Refer Here
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: gameoflife-deploy
labels:
app: gameoflife
spec:
replicas: 3
strategy:
type: RollingUpdate
rollingUpdate:
maxSurge: 1
maxUnavailable: 1
selector:
matchLabels:
app: gameoflife
version: "1.0"
template:
metadata:
labels:
app: gameoflife
version: "1.0"
spec:
containers:
- image: qualitythought/gameoflife:07112020
name: gol-dev
ports:
- containerPort: 8080
protocol: TCP

---
apiVersion: v1
kind: Service
metadata:
name: gol-dev
spec:
type: LoadBalancer
selector:
app: gameoflife
ports:
- targetPort: 8080
port: 80
● Follow the complete application creation in k8s Refer Here
Topic:- K8s Jobs and CronJobs

● Let's assume you want to run a Pod which runs a task and then terminates gracefully, To
solve these kind of Problems k8s has Jobs & CronJobs
● The difference b/w k8s job & CronJob is CronJobs can be scheduled to run on a
particular time frame
● Run the job as mentioned in the changeset Refer Here

---
apiVersion: batch/v1
kind: Job
metadata:
name: one-time-job
spec:
template:
spec:
restartPolicy: OnFailure
containers:
- name: ubuntu-container
image: ubuntu
args:
- /bin/bash
- -c
- sleep 10m
Topic: K8s Ingress
● Ingress is an object in k8s that is used manage external access to services in
the k8s cluster. Ingress provides layer 7 load balancing, so it will be aware
of url’s
● Refer Here {https://docs.microsoft.com/en-us/azure/aks/ingress-basic}

Scenario:
● I have a application component which is running in a pod
● I have a database pod
● Application Pod is configured to scale automatically
 ● Now application pod needs to access db
● Approach 1:
○ Create a service for db

● Whenever you add a service an environmental variables will be added in all pods
containing svc ip and port

Approach 2:
● Use a Config Map. Config Map is a k8s object that allows us to define the
application-related data.
● Config maps are key value pairs
● You can create config maps from
○ literal value
 ○ file
○ directory
● Lets create a sample config map

kubectl create configmap demovalue-map --from-literal=db-ip=10.0.83.222

---
apiVersion: v1
kind: Pod
metadata:
name: configmapdemo
spec:
containers:
- name: configmapdemo
image: alpine
command: ["/bin/bash", "-c", "sleep 1d"]
envFrom:
- configMapRef:
name: demovalue-map
NOTE:- In K8s we also have secret which are much like config maps but they can store secret
of size 1kb in base 64 encoded form

Topic: Helm
● Helm is a Package manager for k8s
● Helm charts represent a k8s package
● Installing helm Refer Here
● Lets execute an example using helm in AKS Refer Here
● In k8s we write yaml manifests, helm simplifies and makes manifests reusable
● Refer Here
● Journey of App from docker to k8s Refer Here