DevOps Mini Concept | Git Branching Strategy and Helm Charts | Sagar Kakkala's World

Highly Recommend using repo here - Day8 to practice this project instead of code present in snippets, In case of confusion, Please do watch video that is explained in English, the code here in the blog is not changed to keep screenshots intact

What is Git Branching Strategy

Git Branching Strategy may differ from every company, The Production branch will be the main branch and direct changes to this code may impact the environment and ongoing business

We divide this into 3 replicas here

DEV , UAT and PROD

All these have same set of servers and same setup, same code in different AWS IAM accounts with minimal variable changes to have different Deployments

Any change that needs to be done should be done first in feature branch, Be it a Developer, tester, devops guy , All would create feature branches for their respective changes and all will be integrated to Dev branch code with Manager Approval.

Once Dev Branch is updated, QA will proceed with approvals and Pull Request(Merge Request) will be raised to get Dev Branch and UAT Branch in sync

once UAT Branch code is deployed, QA will again proceed with tests and approve, Only then Merge Request is raised from UAT to PROD branch

UAT Branch and PROD Branch code will be deployed only on particular days of week

In case you see any application v1.2.4, here 1 denotes major version, 2- denotes minor version, 4- denotes patch version.

In case, you see application with v1.2.4.3 , here 4 denotes hotfix version, Hotfix is generally done when there is a production issue and it needs to be fixed irrespective of release date. something serious happening in production and that needs to be addressed

What are Helm Charts?

In Simple Words..Helm Charts are Package Managers of Kubernetes,

In Kubernetes Deployment, We deployed "pods" with labels and then "service" using same labels to use same connection.Then "Ingress" that connects to service

INGRESS > SERVICE > PODS

But in Helm Chart , We don't need to worry about connecting these kubernetes services together, Helm Chart templates care of this

Take an Example of Docker Hub, We do get Image from Docker hub which we can use and deploy in Kubernetes

But in Helm, instead of Image, you get a complete application

For Suppose, take an example of Grafana, application would be same across any environment.

So you can get complete Grafana image from official Helm Hub

You will get a list of commands once you click on Grafana, just installing Helm and then running those commands, you can get Grafana complete application into your setup

Now let us understand, Helm Charts with our Ongoing Project

Configuration to be used for this project:

This can be done with existing setup of your resources from Previous Projects, especially kubernetes setup

#Note: AWS Free Tier is only for 750 Hours per year, so make sure you stop instance once you are done with the project

As an initial step, Install helm from official documentation or simply run these commands

curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/main/scripts/get-helm-3
chmod 700 get_helm.sh
./get_helm.sh

Now once helm is installed, check helm version

helm version

Once installed, let's create helm chart by simple command

helm create react-chart

here react-chart is the name given for chart directory, you can give any name of your choice, once you created, you will be able to see the directory name with the chart

Now navigate into this directory and you will be able to see files created by helm chart for us

cd react-chart

ls

Chart.yaml contains metadata about our helm chart, Charts contains files if our helm chart is linked with any other helm chart, templates is where we have our kubernetes manifest file, and values.yaml file is the place where we update our values.

if you navigate to templates file, this is what you will see

cd templates

Now if you see any file, here they will have values as some values.xyz, which means they are referring to values.yaml file

for example, let us see service.yaml file

cat service.yaml

As per above screenshot, you can see, we have type in spec is .Values.service.type, To understand what value will be given as input, .check values.yaml file

Now if you open values.yaml file

cat ../values.yaml

As you can see, it has service.type: ClusterIP, so this value will be updated in our templates/service.yaml file

Now let us understand better by adding namespace variable in all templates file

sudo vi deployment.yaml

add line namespace: {{ .Values.namespace }}

since we want to have our deployment, service and ingress files to be in same namespace, update all the files to have same values

sudo vi service.yaml

add line namespace: {{ .Values.namespace }}

sudo vi ingress.yaml

add line namespace: {{ .Values.namespace }}

Now once you added these, lets get back to our values.yaml file as to update variable, now our file searches for variable called namespace and if it is not there, helm does not allow us to deploy our application

sudo vi ../values.yaml

By default you will have values.yaml filled with some date, we need to change it

first change to do is add this variable namespace in the file 

Now, let us make other changes in values.yaml according to our application.We need to be aware of kubernetes manifest files, so as to make changes to values.yaml files, You can refer our project Kubernetes manifest files here - Kubernetes Manifest files

let us understand how we are making changes step by step, as soon as you open values.yaml file, you will be shown this

Now let us check our deployment.yaml file from Github repo to compare

As said earlier, we don't need to worry about labels, according to manifest file, image: sagarkakkala385/react-nginx and replicas are 2, and also in helm chart values.yaml file, you can see there is an option called tag: "" , if you don't specify anything on tag, it will default take Chart Api Version as you can see in the comment on file itself

Now if you want to know what version it will update, check cat Chart.yaml 

cat Chart.yaml

You can see at the end of the file, it says 

so this will try to pull image sagarkakkala385/react-nginx:1.16.0 , which actually is not present in our docker hub. 

As you can see, we only have tag latest, if we leave tag as empty, it will try to fetch sagarkakkala385/react-nginx:1.16.0, which will result only in Imagepullerro, 

and also since we don't use service account, we will change those values to false, so update the initial values like this as shown in screenshot. make changes according to your Docker hub account if necessary

leave the other values as default

scroll down to update other values in values.yaml

we don't need to worry about service here as we are using clusterIp, and as per above screenshot, you can see ingress enabled: false, it means even if we install helm ingress wont be installed, let us make changes , let us compare with our existing ingress file

Now the new updated values.yaml file would look like this,

since i advised, not to take SSL Certificate, we can ignore tls part, In case you have taken SSL, the the values.yaml file would look like this

You can leave the rest values as default, By default Helm is configured to have kubernetes run on http://localhost:8080/ and if kubernetes is running on different port, we need to set the configuration to helm 

kubectl config view

As you can see in my server, it runs at 16443, run this command before installing helm chart else you will get error like 

Why is this error? Cause Helm tries to contact kubernetes at Port 8080 but our Kubernetes is running at Port 16443

so run this command to fix this issue

kubectl config view --raw > ~/.kube/config

and this config file must have only read access, so run

chmod 600 ~/.kube/config

Now we are ready to install our first helm chart, run the command at place where you can see values.yaml file

helm install react-chart .

here react-chart is the name of the chart, you can give any name of your choice

only for the first time, We will use Helm install command, even if you try to run the command again, helm won't allow you 

Once helm is installed, confirm if the pods are running.

kubectl get pods -n react-nginx

from the next time, You need to use helm upgrade command.

if there are any changes that are done in the code, now you want to deploy application, we will use helm upgrade

helm upgrade react-chart .

As you can see, here it displayed REVISION:2, if we run upgrade again , it will display revision as 3

Relate it like Version 1, Version 2 and Version 3 of application, to check all the versions or Revisions, use helm history command

helm history react-chart

As you can see from above screenshot, our application is deployed 3 times, as STATUS: deployed is current running version

for Suppose Version3 has some issue, In order to fix things and business to keep smoothly running, we can jump to Version2 with just single command

helm rollback react-chart 2 .

note: this Rollback also will be deployed like newer Revision

Now if you want to check all the helm charts that are running inside your server

helm list

also to uninstall helm chart, use command, this removes application and its related dependencies

helm uninstall react-chart

You can run uninstallation from any part of server

Helm Charts will not redeploy application unless there is a change in values.yaml file, for suppose, we used tag:latest in file, even if our code is updated and image is pushed again. Helm chart would not fetch latest file.

To overcome this, we will discuss on Creating tag for each commit and pushing every code change with tag attached and attaching this tag value to our docker file.

🔹 Important NoteAlso, before proceeding to the next session, please do the homework to understand the session better - DevOps Homework

I Post most of my content in Telugu related to contrafactums(changing lyrics to original songs),fun vlogs, Travel stories and much more to explore, You can use this link as single point of link to access - Sagar Kakkala One Stop

🖊feedback,queries and suggestions about blog are welcome in the comments.