DEV Community

AWS Community Builders profile image Shakir
Shakir for AWS Community Builders

Posted on • Edited on • Originally published at networkandcode.github.io

Launch AWS EC2 instances the hard way with CLI

#aws #ec2 #linux #kubernetes

Hey All 👋, in this post we shall launch 3 AWS EC2 instances and test SSH connectivity to those...

We are not going to use the GUI / Web console ☺️ for this purpose, we would be using the CLI 😅, and also create individual components along the way, that are required for the instances to function properly, instead of relying on default ones, and thus touch bits of networking and security areas. Hope this approach gives someone a better understanding(like it gave me) of the different components that glue together underhood / behind the scenes, which we don't usually notice when we quickly setup instances with all the default options.

Hence, please ensure you have the following installed and configured: aws cli and jq. Note that jq is used to parse JSON content at different places in this post, though you can use the AWS CLI's built in filters.

You would see the naming conventions such as kubeadm or k8s, as these nodes are created for the purpose of creating a kubernetes cluster, and I thought I would keep the instance launching setup as a separate post.

Alright, let's get started...

Configuration

I have configured us-west-2 as the default region. You can change this as required.

$ cat ~/.aws/config 
[default]
region = us-west-2
Enter fullscreen mode Exit fullscreen mode

And you should also see a credentials file like this.

$cat ~/.aws/credentials 
[default]
aws_access_key_id=<aws_access_key_id>
aws_secret_access_key=<aws_secret_access_key>
Enter fullscreen mode Exit fullscreen mode

Key Pair

Create an ssh key pair, as we need it to login via SSH, to the instances we are about to launch.

$ aws ec2 create-key-pair --key-name kubeadmKeyPair --query 'KeyMaterial' --output text > kubeadmKeyPair.pem

$ ls
kubeadmKeyPair.pem
Enter fullscreen mode Exit fullscreen mode

You can set permission 400 so that only you can read it, beneficial if its a shared system.

$ chmod 400 kubeadmKeyPair.pem
Enter fullscreen mode Exit fullscreen mode

And move it to the .ssh directory, where ssh keys are usually stored.

$ mv kubeadmKeyPair.pem ~/.ssh/.
Enter fullscreen mode Exit fullscreen mode

VPC

Let's create a VPC from which we can reserve IPs for the EC2 instances.

$ aws ec2 create-vpc --cidr-block 10.0.0.0/28
Enter fullscreen mode Exit fullscreen mode

This is a /28 subnet which should give a total of 2 ^ (32-28) = 2 ^ 4 = 16 IPs, including the gateway and broadcast IP. We can not create VPCs with lesser IPs, 16 IPs are minimum, no matter we need it or not.

Let's retreive the VPC ID and save it in a variable.

$ export KUBEADM_VPC_ID=$(aws ec2 describe-vpcs | jq -r '.Vpcs[] | select(.CidrBlock == "10.0.0.0/28") | .VpcId')
$ echo $KUBEADM_VPC_ID
<vpc-id>
Enter fullscreen mode Exit fullscreen mode

Internet Gateway

We shall create an Internet Gateway, as we need internet access to our instances, to access them via SSH from the local machine.

$ aws ec2 create-internet-gateway --tag-specifications 'ResourceType=internet-gateway,Tags=[{Key=Name,Value=kubeadm-k8s-igw}]'
Enter fullscreen mode Exit fullscreen mode

Retrieve the gateway id.

$ export KUBEADM_IGW_ID=$(aws ec2 describe-internet-gateways --filters Name=tag:Name,Values=kubeadm-k8s-igw --query "InternetGateways[*].InternetGatewayId" --output text)

$ echo $KUBEADM_IGW_ID
<igw-id>
Enter fullscreen mode Exit fullscreen mode

Now, let's attach this to the VPC.

$ aws ec2 attach-internet-gateway --internet-gateway-id $KUBEADM_IGW_ID --vpc-id $KUBEADM_VPC_ID
Enter fullscreen mode Exit fullscreen mode

Route Table

A route table will be associated with the VPC we created. Let' find it.

$ KUBEADM_RTB_ID=$(aws ec2 describe-route-tables | jq -r '.RouteTables[] | select(.VpcId == env.KUBEADM_VPC_ID) | .RouteTableId')

$ echo $KUBEADM_RTB_ID
<rtb-id>
Enter fullscreen mode Exit fullscreen mode

We can add a default route in this table, that points to our Internet Gateway.

$ aws ec2 create-route --route-table-id $KUBEADM_RTB_ID --destination-cidr-block 0.0.0.0/0 --gateway-id $KUBEADM_IGW_ID
{
    "Return": true
}
Enter fullscreen mode Exit fullscreen mode

Subnet

A VPC can have more than one subnet, however the minimum number of IPs in the EC2 subnet is also 16, hence we can only have one subnet in this case, which suits our requirement though. Sixteen IPs in the subnet is more than enough, as we plan to launch only 3 instances in this subnet.

$ aws ec2 create-subnet --cidr-block 10.0.0.0/28 --vpc-id $KUBEADM_VPC_ID
Enter fullscreen mode Exit fullscreen mode

The subnet is created, let's save the subnet id too.

$ export KUBEADM_SUBNET_ID=$(aws ec2 describe-subnets | jq -r '.Subnets[] | select(.CidrBlock == "10.0.0.0/28") | .SubnetId')

$ echo $KUBEADM_SUBNET_ID
<subnet-id>
Enter fullscreen mode Exit fullscreen mode

Note that if you have duplicate subnets, then you may also have to add a condition above to check if the subnet belongs to the correct VPC.

We can check the availabilty zone of the subnet.

$ export KUBEADM_AVAILABILITY_ZONE=$(aws ec2 describe-subnets | jq -r '.Subnets[] | select(.CidrBlock == "10.0.0.0/28") | .AvailabilityZone')

$ echo $KUBEADM_AVAILABILITY_ZONE
us-west-2b
Enter fullscreen mode Exit fullscreen mode

Security Group

We would need a security group, that we can attach to our VPC, so that the instances launched in the subnet of the VPC, would leverage the rules defined in the security group.

$ aws ec2 create-security-group --group-name kubeadm-sg --description "kubeadm security group" --vpc-id $KUBEADM_VPC_ID
Enter fullscreen mode Exit fullscreen mode

Let's save the security group's id in a variable.

$ export KUBEADM_SG_ID=$(aws ec2 describe-security-groups | jq -r '.SecurityGroups[] | select(.GroupName == "kubeadm-sg") | .GroupId')

$ echo $KUBEADM_SG_ID
<sg-id>
Enter fullscreen mode Exit fullscreen mode

Image ID

We need to choose an Image for our instances. I am going to choose Ubuntu 20.04 LTS on amd64 architecture, you can get the ami for Ubuntu from this link, I am going to pick ami-036d46416a34a611c.

Instance Type

I am creating these instances to make a kubernetes cluser out of those. Hence, we need to choose an instance type that suffices kubernetes resource requirements, we need to have a minimum of 2 GB Ram and 2 CPUs. Hence I am going to filter for instance types that have 4 GB Ram, which is 4/1.074, i.e. approximately 3.75 GiB or 3840 MiB.

$ aws ec2 describe-instance-types | jq '.InstanceTypes[] | select(.MemoryInfo.SizeInMiB == 3840) | (.InstanceType, .VCpuInfo.DefaultVCpus)'
"c3.large"
2
"c4.large"
2
"m3.medium"
1
Enter fullscreen mode Exit fullscreen mode

Let's check for instance types with 4GiB RAM this time, which also has 2 Default CPUs

$ aws ec2 describe-instance-types | jq '.InstanceTypes[] | select(.MemoryInfo.SizeInMiB == 4096) | select (.VCpuInfo.DefaultVCpus == 2) | .InstanceType' | sort
"a1.large"
"c5ad.large"
"c5a.large"
"c5d.large"
"c5.large"
"c6gd.large"
"c6g.large"
"c6gn.large"
"c6i.large"
"t2.medium"
"t3a.medium"
"t3.medium"
"t4g.medium"
Enter fullscreen mode Exit fullscreen mode

Let's check if t2.medium is available in our subnet's availability zone. Please refer to this link for more details.

$ aws ec2 describe-instance-type-offerings --location-type availability-zone | jq '.InstanceTypeOfferings[] | select(.Location == env.KUBEADM_AVAILABILITY_ZONE) | .InstanceType' | grep t2.medium
"t2.medium"
Enter fullscreen mode Exit fullscreen mode

So, let's fix t2.medium.

Instances

We can now go ahead and create the EC2 instances.

$ aws ec2 run-instances --image-id ami-036d46416a34a611c --count 3 --instance-type t2.medium --key-name kubeadmKeyPair --security-group-ids $KUBEADM_SG_ID --subnet-id $KUBEADM_SUBNET_ID --associate-public-ip-address
Enter fullscreen mode Exit fullscreen mode

We have also enabled public IP address as we need to SSH in to the instances from our machine.

Great, so our instances are created finally.

$ aws ec2 describe-instances | jq -r '.Reservations[0] | .Instances[] | select(.SubnetId==env.KUBEADM_SUBNET_ID) | .InstanceId' 
<i-id1>
<i-id2>
<i-id3>
Enter fullscreen mode Exit fullscreen mode

Let's give these instances, some names.

$ aws ec2 create-tags --resources <i-id1> --tags Key=Name,Value=k8s-master
$ aws ec2 create-tags --resources <i-id2> --tags Key=Name,Value=k8s-node1
$ aws ec2 create-tags --resources <i-id3> --tags Key=Name,Value=k8s-node2
Enter fullscreen mode Exit fullscreen mode

We can get the public IP as follows.

$ export K8S_MASTER_IP=$(aws ec2 describe-instances --filter Name=tag:Name,Values=k8s-master --query "Reservations[*].Instances[*].PublicIpAddress" --output text)

$ echo $K8S_MASTER_IP
<master-public-ip>

$ export K8S_NODE1_IP=$(aws ec2 describe-instances --filter Name=tag:Name,Values=k8s-node1 --query "Reservations[*].Instances[*].PublicIpAddress" --output text)

$ echo $K8S_NODE1_IP
<node1-public-ip>

$ export K8S_NODE2_IP=$(aws ec2 describe-instances --filter Name=tag:Name,Values=k8s-node2 --query "Reservations[*].Instances[*].PublicIpAddress" --output text)

$ echo $K8S_NODE2_IP
<node2-public-ip>
Enter fullscreen mode Exit fullscreen mode

SSH

We need to access the instances via SSH from the local machine, for which we need allow port 22 on the security group.

Get the local machine's public IP.

$ export MY_PUBLIC_IP=$(curl ifconfig.me --silent)
$ echo $MY_PUBLIC_IP
<public-ip>
Enter fullscreen mode Exit fullscreen mode

We can now 📝 add the rule.

$ aws ec2 authorize-security-group-ingress --group-id $KUBEADM_SG_ID --protocol tcp --port 22 --cidr $MY_PUBLIC_IP/32
Enter fullscreen mode Exit fullscreen mode

Let's add the node IPs to a file.

$ cat > k8s-node-ips.txt <<EOF
$K8S_MASTER_IP
$K8S_NODE1_IP
$K8S_NODE2_IP
EOF
Enter fullscreen mode Exit fullscreen mode

And then test the connection.

$ ips=$(<k8s-node-ips.txt)

$ for ip in $ips; do ssh ubuntu@$ip -i ~/.ssh/kubeadmKeyPair.pem 'echo -n "Hello World!, my AWS EC2 hostname is "; hostname'; done
Hello World!, my AWS EC2 hostname is ip-10-0-0-5
Hello World!, my AWS EC2 hostname is ip-10-0-0-11
Hello World!, my AWS EC2 hostname is ip-10-0-0-12
Enter fullscreen mode Exit fullscreen mode

Summary

So we have successfully launched 3 EC2 instances using the AWS CLI, you can customize the options in the commands used, to create instances with different configuration as required. Thank you !!! 👏.

--end-of-post--

Top comments (0)

Read next

nandinirajaram profile image

Quickly Set Up a GraphQL API for DynamoDB CRUD operations with Just a Few Clicks Using AWS AppSync

Nandini Rajaram -

bailon profile image

How to Set Up Selenium as a Linux Daemon with systemd

Juan Bailon -

jutionck profile image

Kubernetes: Install Tools

Jution Candra Kirana -

smkhub profile image

AWS CDK

Sharad Khambe -