Vamshi Krishna Santhapuri

Kubernetes is a Orchestration mechanism for running your container infrastructure on linux based machines.
In this tutorial we will be looking at the server based kubernetes installation on centos7 linux server OS.

Installing the kubernetes minimum requirements

Have 2 CPU cores with 2 GB or more RAM.

Have the swap memory disabled.

The swap memory can be disabled using the swapoff -a command.

Now, Lets take a look at the prerequisites to perform a kubernetes installation:

The Docker as the runtime container engine.
We make sure that the docker is already installed on the system.

[root@node01 ~]# docker --version
Docker version 1.13.1, build b2f74b2/1.13.1

Ensure you are loggedin as the root user to the machine to perform the remaining procedure.
We now start of the Kubernetes installation by adding the yum repo as demonstrated below:

STEP 1:

[root@node01 ~]# cat <<EOF > /etc/yum.repos.d/kubernetes.repo
> [kubernetes]
> name=Kubernetes
> baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
> enabled=1
> gpgcheck=1
> repo_gpgcheck=1
> gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
> EOF

Now update the repositories with yum update command:

# yum update
-- OUTPUT TRUNCATED --
kubernetes/signature                                                                                                                             |  454 B  00:00:00     
Retrieving key from https://packages.cloud.google.com/yum/doc/yum-key.gpg
Importing GPG key 0xA7317B0F:
 Userid     : "Google Cloud Packages Automatic Signing Key <[email protected]>"
 Fingerprint: d0bc 747f d8ca f711 7500 d6fa 3746 c208 a731 7b0f
 From       : https://packages.cloud.google.com/yum/doc/yum-key.gpg
Is this ok [y/N]: y
Retrieving key from https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
-- OUTPUT TRUNCATED --

Till this step the repository addition is complete.

STEP2:

We now Hop onto the proposed Kubernetes master server, proceed with setup of the kubenetes master and Container cluster management components..
Downloading the kubernetes master and the kubernetes network interface binaries to configure the kubernetes master.
The yum package manager offer the following components which have to installed as dependencies to configure the kubernetes-master.

We should do some configuration before hand to enable the bridging net.bridge.bridge-nf-call-iptables

Enabling the bridging on the master node by adding the following to /etc/sysctl.d/kubernetes.conf. Create this file under /etc/sysctl.d

[root@node01 ~]# cat /etc/sysctl.d/kubernetes.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1

Or else we might run into errors like the one as follows:

[ERROR FileContent--proc-sys-net-bridge-bridge-nf-call-iptables]: /proc/sys/net/bridge/bridge-nf-call-iptables contents are not set to 1

Now run the below command to read the new bridging rules.

# sysctl --system

Disabling SELINUX on the kubernetes master

We need to ensure the selinux is disabled for the purpose of simplifying the installation, You may encounter many cases where the selinux context obstructing the kublet to send the information to the kube-controller and kube-scheuler

[vamshi@node01 ~]$ sudo setenforce 0

Setting it to 0 using setenforce will set the selinux to permissive mode, and Verify it with the getenforce will display the results.

[vamshi@node01 ~]$ sudo getenforce 
Permissive

To make the SELINUX rules persistent across the reboot you need to modify its configuration file

[root@node01 ~]# sed -i 's/SELINUX=enabled/SELINUX=disabled/' /etc/selinux/config 
[root@node01 ~]# cat /etc/selinux/config 

# This file controls the state of SELinux on the system.
# SELINUX= can take one of these three values:
#     enforcing - SELinux security policy is enforced.
#     permissive - SELinux prints warnings instead of enforcing.
#     disabled - No SELinux policy is loaded.
SELINUX=disabled
# SELINUXTYPE= can take one of three values:
#     targeted - Targeted processes are protected,
#     minimum - Modification of targeted policy. Only selected processes are protected. 
#     mls - Multi Level Security protection.
SELINUXTYPE=targeted

STEP3

Now lets shift our focus onto the Kubernetes and see the following core components of Kubernetes:

• kube-apiserver
• kube-controller-manager
• kube-scheduler
• kubelet
• kube-proxy

We shall now beign installing kubeadm and kubernetes-cni

# yum install kubeadm kubernetes-cni

Here we have marked the kubernetes-cni because of the network components which goes along well with the kubernetes network scope management.

The important component is kubeadm which presides over the kubernetes cluster initialization.
To access the kubernetes we need the we need to install the kubectl, Although It will installed along with kubernetes-client package and if required can be install with the following command:

# yum install kubectl

STEP 4: Your Kubernetes worker Node

This is exclusive for the worker nodes which will be connected to the working kubernetes master.
STEP 1 is required to setup on the worker node so we can install and configure the kubernetes-node binary.
We will download the kubernetes-node Binaries from the yum package manager.

# yum install kubernetes-node kubernetes-client

STEP 5: Enabling the Full potential on the control-plane

The important step to enable and start the core kubernetes master services.
Here are the core important kubernetes services in the control-plane.

 kube-apiserver
 kube-controller-manager
 kube-scheduler

The Below services contributes on the data-plane or the worker-nodes and are also important on the contol-plane

 kubelet
 kube-proxy

The important configuration files on the kubernetes master:

• /etc/kubernetes/manifests

• /etc/kubernetes/pki

The important config files are:

• /etc/kubernetes/admin.conf

• /etc/kubernetes/kubelet.conf

• /etc/kubernetes/bootstrap-kubelet.conf

• /etc/kubernetes/controller-manager.conf

• /etc/kubernetes/scheduler.conf

The stateful data directories in Kubernetes are as below:

• /var/lib/etcd

• /var/lib/kubelet

• /var/lib/dockershim

• /var/run/kubernetes

• /var/lib/cni

Now we initialize the kubernetes with kubeadm as we see as follows:

kubeadm init --apiserver-advertise-address [preferred-master-node-ip-address|FQDN]

With the kubernetes successfully configured as follows you can begin digging deep onto the setup.

[addons] Applied essential addon: CoreDNS
[addons] Applied essential addon: kube-proxy

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

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 10.100.0.10:6443 --token 123jei.123456783n6o8bq \
    --discovery-token-ca-cert-hash sha256:12345678906bff25a6d132a539e87321833181

Upon the successful installation you should see the following information with the client and the server version information:

Copy the kubeconfig file from the path /etc/kubernetes/config to the desired home directory under .kube/config

[root@node01 ~]# kubectl version
Client Version: version.Info{Major:"1", Minor:"15", GitVersion:"v1.15.1", GitCommit:"4485c6f18cee9a5d3c3b4e523bd27972b1b53892", GitTreeState:"clean", BuildDate:"2019-07-18T09:18:22Z", GoVersion:"go1.12.5", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"15", GitVersion:"v1.15.0", GitCommit:"e8462b5b5dc2584fdcd18e6bcfe9f1e4d970a529", GitTreeState:"clean", BuildDate:"2019-06-19T16:32:14Z", GoVersion:"go1.12.5", Compiler:"gc", Platform:"linux/amd64"}

STEP 6: Initialize the Networking in Kubernetes

Here we enable the kubernetes networking with the preferred network provider:

kubectl apply -f https://docs.projectcalico.org/v3.11/manifests/calico.yaml

We should be able to get the nodes

[root@node01 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
master.linuxcent.com Ready master 3h7m v1.18.2

Common Errors during the setup:

There can be some common errors during the installation I have faced and able to reproduce them in-order to find a quick resolution.

The kubelet is unhealthy due to a misconfiguration of the node in some way (required cgroups disabled)
[E0509 9645 kubelet_node_status.go:92] Unable to register node

If you encounter the above error, then please ensure the following things:
Ensure that you have the kubelet service running,
The selinux is in disabled state. and then reinitialize, kubeadm reset and then kubeadm init command.

There may be errors related to the DNS not functioning:

Warning  FailedScheduling    default-scheduler  0/1 nodes are available: 1 node(s) had taint {node.kubernetes.io/not-ready: }, that the pod didn't tolerate.

runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: cni config uninitialized

Then it definetly needs to apply the kubernetes networking plugin, please choose the calico or Weavenet or your preferred network plugin and apply those components.

The Dockerfile is a very simple to understand format containing of Statements often referred to as Docker DSL(Domain Specific Language), It tends to become quiet complex and difficult to understand over the time.

[vamshi@docker01 ~]$ cat Dockerfile

# Dockerfile which runs a Latest Ubuntu image and sleeps for 100 seconds
FROM ubuntu:18.04
LABEL maintainer="vamshi" version="1.0.0" description="My First Docker Image"
RUN apt-get update && apt-get dist-upgrade -y && apt-get autoremove -y && apt-get install -y tomcat
RUN apt-get remove --purge -y $(apt-mark showauto) && rm -rf /var/lib/apt/lists/*
WORKDIR /data
ENTRYPOINT ["sleep", "100"]

Let’s examine the Dockerfile and the Statements as follows:

The First line starting with # is a comment.
The Second line with FROM tag determines the image and the latest tag; .
Eg:

FROM ubuntu:latest FROM ubuntu:18.04 or FROM centos:7

The LABEL is a Descriptive tag and contains the information about the original author credits
RUN command simulates a shell command and the subsequent statements are executed as a shell command inside the container.
WORKDIR determines the  directory context inside the running container.
ENTRYPOINT is the invocation of the container main process which runs when the docker container runs and its failure to run means the termination or the end of the particular container.
From the above Dockerfile you would have noticed 2 formats of Instruction, Now lets discuss them in details:
Shell Form
The instructions are written as shell commands

RUN apt-get update , this in turn is formatted as bin sh -c “apt-get update” and enables for command expansion, inclusion of the special characters and it enables combining of multiple commands.

Exec Form
It is JSON array style
These instructions are also shell commands but they are represented in the form of elements in a list.

This format has the following drawbacks
Here the shell is not provided
No scope for variable expansion and
also this format doesn’t allow the special characters like (&&, ||, >….) to be included into the command expression statements.

While running the docker container, the CMD takes the run time arguments and the JSON list format works as a preventive measure.

Advantage of CMD and ENTRYPOINT using square bracket JSON array notation

The most advantageous point with CMD or ENTRYPOINT being written in JSON list format is the during the container runtime, the CMD can take certain arguments which can alter the main container process.. and Thus we can shield against variable expansion, Injecting special characters and not providing a shell as a counter measure security practice.

How to build a Dockerfile?

From the current working directory navigate to the location where Dockerfile is present and run the below command.

# docker build --tag first-docker-image -f Dockerfile .

How to extract Build description from docker image?

We are able to extract the docker LABEL description and MAINTAINER information from the docker command which will help in identifying its purpose when have some hundreds of docker images.

[vamshi@node01 ~]$ docker image inspect first-image --format='{{.Config.Labels}}'
map[description:My First Docker Image maintainer:vamshi version:1.0.0]

Best Practices while building Docker images.

• While writing the Dockerfile, Its a good practice to include the version information.
• Include the description of the image which can be easily understood while inspecting the docker image from the inspect command.

• Grouping RUN shell commands together with && which are inter-dependent and relevant. The Docker by design stores a single STATEMENT as One Layer Image. This will enable better compressing and storing of docker image layers. This technique is called cache-busting.

  RUN apt-get update && apt-get dist-upgrade -y

• Its a good practice to cleanup the installed package sources and build files which lightens up the size of docker image layer.

  RUN apt-get autoremove -y && rm -rf /var/lib/apt/lists/*

• The Docker statements can be grouped together that have require less to no modification to save the network bandwidth and increase the docker build time.

The endpoints in kubernetes are the mechanism that directly interact and implement the Kubernetes Service

The Endpoints are underlying mechanism which are created in the background and enable us to talk to the kubernetes Services.
As we know that by creating a Kubernetes service we automatically generate the FQDN with the help of core-dns services

There was a requirement for me to setup the specific endpoint and create a service to convert a outside IP into the kubernetes FQDN [svc-name.namespace.cluster.local]
I managed to work around it by creating an endpoint of my external IP which was running mySQL.

[root@master01 ~]# cat Mysql-ep.yaml 
---
kind: Endpoints
apiVersion: v1
metadata:
 name: mysql-svc
 namespace: actoneye
subsets:
 - addresses:
     - ip: 172.22.110.130
   ports:
     - port: 3306

Lets take a look at Kubernetes Service yaml file.

[root@master01 ~]# cat Mysql-svc.yaml

---
kind: Service
apiVersion: v1
metadata:
 name: mysql-svc
spec:
  ports:
    - protocol: TCP
      port: 3306
      targetPort: 3306

The mysql FQDN mysq-svc.dev.cluster.local was used in my application code for my cluster which the mysql resource was outside the kubernetes cluster.

Please try this out and let me know if you had any similar experiences to share.

How to Inject Variables into Docker image in build time without modifying the Dockerfile

We have a requirement wherein we need to modify the specific Dockerfile with the build information and we make use of string replace operations like SED to modify the data in Dockerfile, But we can make use of the docker build time arguments to achieve the results efficiently.

Here’s a snippet of Dockerfile

FROM ubuntu:17.04
LABEL maintainer="vamshi" version="1.0.0" description="JRUBY Docker image"
WORKDIR /app
ARG JRUBY_VER 9.2.11.1
ENV JRUBY_VER ${JRUBY_VER}
ADD https://repo1.maven.org/maven2/org/jruby/jruby-dist/${JRUBY_VER}/jruby-dist-${JRUBY_VER}-bin.tar.gz .

We now build this Dockerfile as follows passing the –build-arg:

# docker build --build-arg JRUBY_VER=9.2.8.0 -t jruby:v2  -f ./Dockerfile .

Here we have passed in the build arguments of –build-arg the JRUBY_VER=9.2.8.0 and this then assigns the Argument to the ENV and passed it to ADD command which downloads tar.gz.

When run the command without any build-args it readsup the predefined ARG JRUBY_VER 9.2.11.1 from the Dockerfile.

Conclusion:
Another best usecase is while you have a continuous Docker build pipeline system and want to pass the build time arguments on user input, Its best to use ARG statements inside while writing your Dockerfile

The Docker Environment variables can be declared during the (1) Docker image creation inside the Dockerfile and (2) During the docker container run time.

The Dockerfile ENV syntax is shown as follows:

ENV VAR_NAME value

We will look at the Dockerfile syntax in the following snippet:

# Exporting the Environment Variables
ENV MAVEN_HOME /usr/share/maven/
ENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64/
# Added as per Java team
ENV JAVA_OPTS -Xms256m -Xmx768m

We can even reference the ENV variable after the declaration at the next statement.

RUN ${JAVA_HOME}/bin/java -version

Conclusion:

Setting the ENV is a lot better than Hard coding the values into the Dockerfile, as it makes maintaining the Dockerfile a simpler task.

Its a good practice for storing the path names, version of packages which are set to be modified over a longer course of time.

The docker container is simply a run time copy of a docker image resources, The docker container utilizes the filesystem structure originally packed into it via the union filesystem packaged from various image layers during the docker image creation.

The docker provides a standard diff command which compares the filesystem data in docker image with the container.

Syntax:

# docker diff [CONTAINER ID | CONTAINER NAME]

Before jumping in lets examine a docker container below and take a look at filesystem by logging into it.

root@node03:~# docker ps
CONTAINER ID        IMAGE               COMMAND             CREATED             STATUS              PORTS               NAMES
e3de85aaf61c        9a0b6e4f0956        "sh"                2 months ago        Up 3 minutes                            jovial_hertz

We have a running container with a random name jovial_hertz, and we login to the container as follows

root@node03:~# docker exec -it jovial_hertz bash

We are now inside the container and now will create a directory linuxcent and the also create a ASCII text file test and then exit out from the container.

root@e3de85aaf61c:~# mkdir linuxcent
root@e3de85aaf61c:~# cd linuxcent
root@e3de85aaf61c:~# touch test
root@e3de85aaf61c:~# exit

Created a directory called linuxcent and a touched a test file and now logged out from the container with the exit command.
The docker diff command will run against the container should result in the modified data and we contemplate the results

root@node03:~# docker diff jovial_hertz 
C /root
A /root/linuxcent
A /root/linuxcent/test
C /root/.bash_history

The Flags A in front of /root/linuxcent and /root/linuxcent/test indicate that these are directory and file that were the new additions to the container and Flag C indicates that the other 2 files were changed.
Thus it helps us to compare and contrast the new changes to a container filesystem for better auditing.

STEP 1: Docker login to private registry

Lets see the syntax of docker login command followed by the authorized username and the repository URL.
Syntax:

[root@docker03:~]# docker login [DOCKER-REGISTRY-SERVER] -u <username> [-p][your password will be seen here]

The -p is the option for password which can be given along with the docker command or you can type it in the password prompt after hitting enter on the docker login command.

Example given:

[root@docker03:~]# docker login nexusreg.linuxcent.com:5000 -u vamshi
Password:
Login Succeeded

Once the docker login is succeeded a json file will be generated under your home directory at the following path which contains the auth metadata information.

[root@docker03:~]# cat $HOME/.docker/config.json
{
	"auths": {
		"nexusreg.linuxcent.com:5000": {
			"auth": "1234W46TmV0ZW5yaWNoMjAxOQ=="
		}
	},
	"HttpHeaders": {
		"User-Agent": "Docker-Client/18.09.1 (linux)"
	}
}

The docker login repository URL can be found out from your docker client machine using docker info command if you had previously logged in, as we see below:

[root@docker03:~]# docker info | grep Registry
Registry: https://index.docker.io/v1/

How to logout from the specific docker registry use the docker logout command.
The syntax is shown as below:
docker logout [DOCKER-REGISTRY-SERVER]

Example given:

[root@docker03:~]# docker logout 
Removing login credentials for https://index.docker.io/v1/

We would need java runtime environment to install and operate the kafka and zookeeper programs on our linux environment as a dependency which uses the java runtime environment.
So lets quickly check the current java version on the system with java -version.
If its not present Lets now begin with the setup of Java and quickly download the java stable version from the epel repository.

# yum install java-1.8.0-openjdk

Once the java version is installed we verify it with the java -version command.
Creating the kafka user account.

# useradd kafka -m -d /opt/kafka

Adding the kafka password

# echo "kafka" | passwd kafka --stdin

usermod -aG wheel kafka

# su -l kafka

Now login to the system as kafka user.

# cd /opt

Navigate to the url : https://downloads.apache.org/kafka/ and download the latest kafka version

# curl https://downloads.apache.org/kafka/2.5.0/kafka_2.13-2.5.0.tgz -o /opt/kafka.tgz
# tar -xvzf ~/opt/kafka.tgz --strip 1
# cd /opt/kafka
# cp -a /opt/kafka/config/server.properties ~/opt/kafka/config/server.properties.bkp
# echo -e "\ndelete.topic.enable = true\nauto.create.topics.enable = true" >> /opt/kafka/config/server.properties

Adding the Kafka start/stop scripts to systemctl controller daemon services

# sudo vim /etc/systemd/system/zookeeper.service
Add the following lines to /etc/systemd/system/zookeeper.service
[Unit]
Requires=network.target remote-fs.target
After=network.target remote-fs.target

[Service]
Type=simple
User=kafka
ExecStart=/opt/kafka/bin/zookeeper-server-start.sh /opt/kafka/config/zookeeper.properties
ExecStop=/opt/kafka/bin/zookeeper-server-stop.sh
Restart=on-abnormal

[Install]
WantedBy=multi-user.target

Now adding the Kafka control scripts to systemd service

# sudo vim /etc/systemd/system/kafka.service
Add the following lines to /etc/systemd/system/kafka.service
[Unit]
Requires=zookeeper.service
After=zookeeper.service

[Service]
Type=simple
User=kafka
ExecStart=/bin/sh -c '/opt/kafka/bin/kafka-server-start.sh /opt/kafka/config/server.properties > /opt/kafka/kafka.log 2>&1'
ExecStop=/opt/kafka/bin/kafka-server-stop.sh
Restart=on-abnormal

[Install]
WantedBy=multi-user.target

# sudo chown kafka.kafka -R /opt/kafka/logs

# sudo mkdir /var/log/kafka-logs

# sudo chown kafka.kafka -R /var/log/kafka-logs

Now start the zookeeper and kafka services as follows:

# sudo systemctl enable zookeeper --now

# sudo systemctl enable kafka --now