Docker

Learning docker from this udemy course.

command to run code

1. docker build -t <container_name> .
2. docker run -p <local_port>:<docker_exposed_port> <container_name> example: docker run -p 8080:8080 abir/simpleweb

3. note: host.docker.internal works if you install docker desktop. I mean development version.

• But if production version installed using sudo apt-get install .... it does not support it.
• but we can use it by passing an argument in docker run --add-host=host.docker.internal:host-gateway ...

Docker Commands

• FROM image-name is used to specify the base image from which you are building.

• WORKDIR is used to set the working directory for any RUN, CMD, ENTRYPOINT, COPY and ADD instructions that follow it in the Dockerfile.

  • If the WORKDIR doesn't exist, it will be created

• COPY is used to copy files or directories from a source to a destination.

• RUN is used to execute a command when the image is being built.

• EXPOSE is used to specify the port on which the container will listen for incoming connections.

• CMD is used not to run anything when the image is being built, but to run something when a container is being started.

• ATTACH is used to attach to a running container.

  • docker attach <container_id>

• logs is used to fetch the logs of a container.

  • docker logs <container_id>
  • docker logs -f <container_id> to follow the logs

• ENV is used to set the environment variable.

  • ENV <key> <value>
  • docker run --env <key>=<value> <container_name>

• --env-file is used to set the environment variable from a file.

  • docker run --env-file <file_name e.g .env> <container_name>

• ARG is used to pass the argument to the dockerfile.

  • ARG <key>
  • docker build --build-arg <key>=<value> .
  • so it's bind in image building and locked not in container running.

• host.docker.internal is used to access the host machine from the container.

  • {mongo: 'host.docker.internal'} rather than {mongo: 'localhost'} to connect to the mongoDB running on the host machine.

• docker network create <network_name> is used to create a network.

  • This is how we can connect multiple containers together.
  • docker run --network <network_name> <container_name>
  • docker network connect <network_name> <container_name> to connect a container to a network.
  • in this case we put the container name in the place of the host name in the connection string.
    • mongodb://<mongo_container_name>:27017/<db_name>
  • note: we can connect to other container using the ip address of the container as well.
  • but it's not recommended.

• -v is used to mount a volume.

  • docker run -v /app/node_modules is anonymous volume.
  • docker run -v data:/app is named volume.
  • docker run -v $(pwd):/app is bind mount.

Docker Compose notes

Though when we run docker compose it creates container name with some suffix and prefix.

• it still refers to the container name we provided in the docker-compose.yml file. So we can use the container name in this case service name specified by us in docker-compose in the connection string and for other purposes.
• But in command line we have to use the container name with the suffix and prefix.

Docker Compose Commands

• docker-compose up is used to start the containers.

  • docker-compose up -d to run in the background.
  • docker-compose up --build to rebuild the image.
  • it will create a network and connect the containers to the network.
  • it will create a volume and mount the volume to the container.
  • all automatic name will be prefixed with the parent folder name.

• docker-compose down is used to stop the containers.

  • docker-compose down -v/--volumes to remove the volumes as well.

Kubernetes

kube control commands:

• kubectl create deployment <deployment_name> --image=<image_name> to create a deployment.

  • local image cannot be used. It has to be pushed to the docker hub. Because the kubernetes cluster is a separate environment. and will through error ErrImagePull.
  • kubectl get deployments to get the deployments.
  • kubectl delete deployment <deployment_name> to delete the deployment.
  • kubectl get pods to get the pods.
  • kubectl describe pod <pod_name> to get the details of the pod.

• kubectl expose deployment <deployment_name> --type=TypeName --port=portId to expose the deployment.

  • type ClusterIP is used to expose the deployment to the inside the cluster.
  • type NodePort is used to expose the deployment to the outside world but it will expose the deployment to the port range of the worker node.
  • type LoadBalancer is used to expose the deployment to the outside world. It will create a load balancer.
  • kubectl get services to get the services.
  • kubectl delete service <service_name> to delete the service.

• kubectl scale deployment/<deployment_name> --replicas=<no_of_replica> to scale up or down the deployment.

  • fact is if we scale down those pods will be deleted which are not in use or has more restarts.

• kubectl set image deployment/<deployment_name> <container_name>=<new_image_name> to update the image of the deployment.

  • kubectl rollout status deployment/<deployment_name> to check the status of the rollout.
  • kubectl rollout history deployment/<deployment_name> to check the history of the rollout.
  • kubectl rollout undo deployment/<deployment_name> to undo the latest deployment.
  • kubectl rollout undo deployment/<deployment_name> --to-revision=<revision_number> to undo the rollout to a specific revision.

• kubectl apply -f <file_name> to apply the configuration file.

  • kubectl apply -f <file_name> -f <file_name> or -f <file_name1>,<file_name2> to apply multiple configuration files.
  • kubectl apply -f <directory_name> to apply all the configuration files in the directory.

• kubectl delete -f <file_name> to delete the configuration defined in that file.

  • kubectl delete -f <file_name> -f <file_name> or -f <file_name1>,<file_name2> to delete multiple configuration files.
  • kubectl delete -f <directory_name> to delete all the configuration files in the directory.

Minikube commands:

• minikube start --driver=driverName to start the kubernetes cluster.
  • drivers for linux can be docker, QEMU, kvm2, virtualbox, Podman, none
• minikube status to check the status of the cluster.
• minikube dashboard to open the dashboard.
• minikube service <service_name> to open the service url in the browser.
• minikube stop to stop the cluster.
• minikube delete to delete the cluster.

foot note about kubernetes:

There are two types of volumes in kubernetes.

1. "normal" volumes: Which are tied to the pod lifecycle or worker lifecycle.
2. "persistent" volumes: Which are not tied to anything. Rather it is a standalone cluster resource which we claim in our pod.

Kubernetes files [YAML]:

look at master-deployment.yaml file. which i'll be referencing

• --- is used to separate the different objects in the same file.
• apiVersion is used to specify the version of the kubernetes api.
  • apiVersion: apps/v1 is used to specify the version of the apps. for deployment kind
  • apiVersion: v1 is used to specify the version of the api. for service kind
• kind is used to specify the kind of the object.
• metadata is used to specify the metadata of the kind object.
  • name is used to specify the name of the kind object.
  • labels is used to specify the labels of the kind object.
    • exmaple: labels: group: example
    • Now to delete by name, kubectl delete [deployment,service,...] -l group=example
• spec is used to specify the specification of the object.
  • replicas is used to specify the number of replicas.
  • selector is used to specify the selector of the object.
    • matchLabels is used to specify the labels of the object.
    • matchExpressions is used to specify the expressions of the object.
      • example: matchExpressions: [{key: environment, operator: In, values: [first-app, meta-app]}]
  • template is used to specify the template of the kind object which is pod.

    • metadata is used to specify the metadata of the pod object.

      • labels is used to specify the labels of the pod object.

    • spec is used to specify the specification of the pods object. (container specification)

      • containers is used to specify the containers of the pods object.
        
        • name1 is used to specify the name of the container.
          image1 is used to specify the image of the container.
          volumeMounts is used to specify the volume mounts of the container.
          - mountPath is used to specify the mount path of the volume.
          name is used to specify the name of the volume.
          imagePullPolicy is used to specify the image pull policy of the container.
          • [Always, IfNotPresent, Never] are some policy
        • livenessProbe is used to specify how we check if the image container is live.
          • httpGet is used to specify the http get of the liveness probe.
            • path is used to specify the path of the http get.
            • port is used to specify the port of the http get.
          • initialDelaySeconds is used to specify the initial delay seconds of the liveness probe.
          • periodSeconds is used to specify the period seconds of the liveness probe.
      • volumes is used to specify the volumes of the pods object.
        • name is used to specify the name of the volume. like story-volume emptyDir is used to specify the empty directory for each pod.
          or
          hostPath is used to specify the host path of the volume into a single worker node.
          path is used to specify the path of the host path. like /app
type is used to specify the type of the host path. like DirectoryOrCreate

    • spec for service specification.

      • type is used to specify the type of the service.
        • types can be ['ClusterIP', 'NodePort', 'LoadBalancer']
      • ports is used to specify the ports of the service.
        • port is used to specify the port of the service.
        • targetPort is used to specify the target port of the service.

reverse proxy is used in nginx when put to deployment

• we can use nginx as a reverse proxy to route the request to the appropriate service. for example in the nginx.conf file we can specify the configuration like this

    listen 80;
  
    location /api/ {
      <!-- proxy_pass http://<service_name>:<port>/; -->
      proxy_pass http://tasks-service.default:8000/;
    }
  }```
  

• after that we change all url in the client side from http://url:port to http://url/api and it will route the request to the appropriate service. note: the trailing / is important in the proxy_pass directive.

Always check if port matches your call

Always check if aws is configured properly

if some pod has issue, see

• kubectl describe pod <pod_name> to see the error