part-3.md

Part 3

3.1

Current frontend and backend image from 1.12 are 260MB and 164MB respectively.

Working directory: ./part-3/3-01.

To optimize the frontend image, we will use Docker multi-stage build to build a production version and use nginx to serve static files and remove the rest of the source code: Dockerfile

docker build --file Dockerfile_front --rm -t mighty_thor ../..

The new image is only 24.1MB now. Since we are using nginx now, the frontend will now run on port 80.

docker run -d -p 5000:80 --name furious_thor --rm mighty_thor
docker run -d -p 8000:8000 --name tricky_loki --rm mighty_loki

3.2

Working directory: ./part-3/3-02

docker build --rm -t moomin .

The image size is 234MB.

Download an audio file usign moomin:

docker run -v $(pwd)/audios:/app moomin "https://areena.yle.fi/1-50325682"

3.3

Working directory: ./part-3/3-03

docker build --file Dockerfile_front --rm -t mighty_thor ../..
docker build --file Dockerfile_back --rm -t mighty_loki ../..

Run the containers:

docker run -d -p 5000:5000 --name furious_thor --rm mighty_thor
docker run -d -p 8000:8000 --name tricky_loki --rm mighty_loki

3.4

Current Dockerfiles in ./part-3/3-03 are already using Alpine. The frontend image is 24.1MB and the backend image is 164MB.

3.5

Current Dockerfile of frontend is using Multi-stage builds. Dockerfile.

3.6

Working directory: ./part-3/3-07. Optimize the calculator image in part-1.

• Before optimization: Dockerfile. The image size is 115MB.

  docker build --file Dockerfile_before --rm -t vision_before ../..
  

• After optimization: Dockerfile. The image size is 23.7MB.

  docker build --file Dockerfile_after --rm -t vision_after ../..
  

3.8

Create an example setup that uses Kubernetes, Docker Swarm or equivalent using any project and deploy it online. You can use any applications available to you (such as the frontend - backend example). Make sure that the application consists of atleast two different containers that are deployed on two different hosts. Explain what you created in your submission.

I will use Kubernetes to set up a monitor system using Grafana and Prometheus to track MySQL and JVM.

1. Install kubectl and aws-cli.
2. Install kops. kops helps you create, destroy, upgrade and maintain production-grade, highly available, Kubernetes clusters from the command line.
3. Follow this instruction to create a IAM user on AWS.
4. Configure DNS for the domain by following the instruction. In this case, I will use my subdomain k8s.jebetech.online.

ID=$(uuidgen) && aws route53 create-hosted-zone --name k8s.jebetech.online --caller-reference $ID | jq .DelegationSet.NameServers

Output:

[
  "ns-1269.awsdns-30.org",
  "ns-1644.awsdns-13.co.uk",
  "ns-111.awsdns-13.com",
  "ns-819.awsdns-38.net"
]

aws route53 list-hosted-zones | jq '.HostedZones[] | select(.Name=="jebetech.online.") | .Id'

Output:

/hostedzone/Z3KWWM38J8231S

Create a new JSON file (k8s.json) with these data:

{
  "Comment": "Demo k8s",
  "Changes": [
    {
      "Action": "CREATE",
      "ResourceRecordSet": {
        "Name": "k8s.jebetech.online",
        "Type": "NS",
        "TTL": 300,
        "ResourceRecords": [
          {
            "Value": "ns-1269.awsdns-30.org"
          },
          {
            "Value": "ns-1644.awsdns-13.co.uk"
          },
          {
            "Value": "ns-111.awsdns-13.com"
          },
          {
            "Value": "ns-819.awsdns-38.net"
          }
        ]
      }
    }
  ]
}

Apply the sub-domain NS record to the parent hosted zone:

aws route53 change-resource-record-sets \
 --hosted-zone-id Z3KWWM38J8231S \
 --change-batch k8s.json

Output:

{
    "ChangeInfo": {
        "Id": "/change/C34YUVWC1RVQH2",
        "Status": "PENDING",
        "SubmittedAt": "2019-11-10T12:26:18.261000+00:00",
        "Comment": "Demo k8s"
    }
}

5. Create a dedicated S3 bucket to store k8s assets: Cluster State Storage

aws s3api create-bucket \
    --bucket kops-uh-s3 \
    --region us-east-1

6. Create a configuration file kops.env for the clusters

export NAME=k8s.jebetech.online
export KOPS_STATE_STORE=s3://kops-uh-s3

Then source the env file:

source kops.env

7. Create a cluster:

kops create cluster \
    --zones=ap-southeast-1a \
    --cloud=aws
    --node-count=4
    --name=${NAME}
    --yes

Now run kops validate cluster and all 4 nodes are up:

Using cluster from kubectl context: k8s.jebetech.online

Validating cluster k8s.jebetech.online

INSTANCE GROUPS
NAME                    ROLE    MACHINETYPE     MIN     MAX     SUBNETS
master-ap-southeast-1a  Master  m3.medium       1       1       ap-southeast-1a
nodes                   Node    t2.medium       4       4       ap-southeast-1a

NODE STATUS
NAME                                                    ROLE    READY
ip-172-20-35-60.ap-southeast-1.compute.internal         master  True
ip-172-20-43-139.ap-southeast-1.compute.internal        node    True
ip-172-20-45-85.ap-southeast-1.compute.internal         node    True
ip-172-20-49-41.ap-southeast-1.compute.internal         node    True
ip-172-20-62-100.ap-southeast-1.compute.internal        node    True