kubernetes django app deployment¶

kubernetes django app deployment

Kubernetes has awesome deployment features. Deploying the django application with kubernetes even more awesome. We can use kubernetes cloud providers like Googles GKS, Amazons EKS and for local environment we can use microk8s.

Setting up microk8s in local environment¶

To install microk8s in local environment execute the below commands

sudo snap install microk8s --classic --channel=1.16/stable
sudo usermod -a -G microk8s $USER
su - $USER microk8s.status --wait-ready

After installation enable the DNS to allow services to communicate with one another using below command

microk8s.enable dns storage

Now, open ~/.bashrc and below line at the end of the file.

alias k8s="microk8s.kubectl"

Deploying the django app in micro kubernetes¶

I've written the deployment docs and django app and pushed the project to the github. We can find it at https://github.com/AnjaneyuluBatta505/kubernetes_django.

Clone kubernetes_django repo¶

git clone git@github.com:AnjaneyuluBatta505/kubernetes_django.git

Setting up postgres service¶

In the postgres we have 4 yaml configuration files.

secrets.yaml - Useful when using secret environment varibles in the apps
volume.yaml - Persistent Volume for Postgres database
volume_claim.yaml - This claim lets our deployment application store its data in an external location, so that if one of the application’s containers fails, it can be replaced with a new container and continue accessing its data stored externally, as though an outage never occurred.
deployment.yaml - Deployment to start the application containers
service.yaml - Allows other apps to comminicate with it's related deployment(i.e pods/containers)

Now, create above objects in kubernetes using below commands

k8s apply -f kubernetes/postgres/secrets.yaml
k8s apply -f kubernetes/postgres/volume.yaml
k8s apply -f kubernetes/postgres/volume_claim.yaml
k8s apply -f kubernetes/postgres/deployment.yaml
k8s apply -f kubernetes/postgres/service.yaml

We have successfully created the postgres service in k8s(i.e kubernetes cluster).

Setting up redis service¶

In the redis we have 2 yaml configuration files. To setup the redis service in k8s execute the below commands

k8s apply -f kubernetes/redis/deployment.yaml
k8s apply -f kubernetes/redis/service.yaml

Setting up django app¶

In django app we have 3 yaml configuration files. To setup the django app in k8s execute the below commands

k8s apply -f kubernetes/django/job-migration.yaml
k8s apply -f kubernetes/django/deployment.yaml
k8s apply -f kubernetes/django/service.yaml

Setting up celery app¶

In celery app we have 2 yaml configuration files. To setup the django app in k8s execute the below commands

k8s apply -f kubernetes/celery/beat-deployment.yaml
k8s apply -f kubernetes/celery/worker-deployment.yaml

Setting up celery-flower app¶

In flower app we have 2 yaml configuration files. To setup the django app in k8s execute the below commands

k8s apply -f kubernetes/flower/deployment.yaml
k8s apply -f kubernetes/flower/service.yaml

How to debug the pods for errors if any failures occurs ?¶

$ k8s get po
NAME                             READY   STATUS      RESTARTS   AGE
celery-beat-75b5f954-m2pzj       1/1     Running     0          64m
celery-worker-56fc7b88f5-shd87   1/1     Running     0          61m
django-688f76f576-kt4h6          1/1     Error       0          60m
django-migrations-mhng7          0/1     Completed   0          59m
flower-77bf99c799-9drnp          1/1     Running     87         20h
postgres-76dc76ffbb-hzmdc        1/1     Running     2          23h
redis-76f6f4857b-srxlw           1/1     Running     1          20h

We can debug the pod failure using command k8s logs <pod name>. Sometimes k8s describe <name of k8s object>

How to access the django app in our local browser¶

$ k8s get svc
NAME               TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
django-service     NodePort    10.152.183.146   <none>        8000:32693/TCP   21h
flower-service     NodePort    10.152.183.170   <none>        5555:30675/TCP   21h
kubernetes         ClusterIP   10.152.183.1     <none>        443/TCP          25h
postgres-service   ClusterIP   10.152.183.151   <none>        5432/TCP         24h
redis-service      ClusterIP   10.152.183.89    <none>        6379/TCP         21h

Visit url "10.152.183.146:8000" (i.e CLUSTER-IP:PORT). It's just for testing.

References: