To deploy Apache Mesos and the pertinent components, we need servers where we can install these components. So, in this module, we will work on creating EC2 instances on AWS.

To deploy Mesos on AWS, we should have a Virtual Private Cloud (VPC), a subnet, in order to launch our instances inside different subnets as per our requirements, a route table where our instances can communicate with each other, and an internet gateway for public access, and then we will create the EC2 instances. After the creation of EC2 instances, we will deploy our Mesos software and their components.

Before we start with EC2 instances, let's understand the architecture of our system. The following diagram shows the architecture in which we are working on AWS and creating a VPC. Now we will create the EC2 instance, which will have two Marathon servers and three Mesos master...

In order to add the mesosphere repository, run and install a command, as shown:

After this, install ZooKeeper with yum by running the sudo yum install mesosphere-zookeeper command. This will install ZooKeeper on CentOS:

We can see the mesosphere-zookeeper package with version 3.4.6 from the mesosphere repository, and the size is 2.8 MB.

This screenshot shows that the package has been installed:

Now, install the Mesos master and Mesos slave, as shown:

We can see the mesos package version 1.5 from the mesosphere repository, and the size is 76 MB.

Hence, this will download and install Mesos, which will comprise the Mesos master and slave service, as seen here:

The sudo yum-config-manager command will add a Docker repository in our CentOS, as seen in the following screenshot:

Run the sudo yum install docker-ce command:

Type yes, which will install Docker Community Edition:

Perform the following steps to configure ZooKeeper:

1. Type cd /etc /mesos.
2. Open the vi zk file. The following screenshot shows the default Mesos definition. This is the default ZooKeeper information. The line starts with zk and ends with /mesos, and in between, we will have the IP address and the ZooKeeper port:

3. Replace this localhost with the host IP address by typing ifconfig in order to get the IP. As we have a standalone server, we will use the IP address that is highlighted here:

4. Go to cd /etc /mesos and run sudo vi zk to replace the IP.
5. Save the file. This completes the ZooKeeper connection information for Mesos.
6. Map the unique ID by typing var/lib/zookeeper and cat myid.
7. It has defined 1. As we have a standalone system, we will keep ID 1 for this server only.

To start ZooKeeper, type sudo service zookeeper start. Check the status:

You can see status -l has more information:

Run sudo service mesos-master start, and see the status:

You can also add -l to get more information:

In order to access the console, here is the Mesos-master server and the public IP is 34.237.145.106 port is 5050. But we won't be able to access this as we need to enable the traffic and security group.

To do so, check the Security groups, which is mesos. We need to add a rule and input 5050 as Port Range, in Inbound | Edit. Where, we will consider the source as Anywhere but you should define your IP or your custom IP range and save it.

Now let's go and refresh our page(34.237.145.106:5050). We can see that we are able to access Mesos:

In this way, you can enable the inbound rule for your Mesos master server.

The following are the steps to install the Marathon package on the EC2 instance:

1. The following screenshot shows the Marathon server in which we will install the package and add the Mesosphere repository to install the Marathon package:

2. Install the Marathon by run the following command:

3. Perform the sudo rpm -qa | grep marathon command, which will show the Marathon package installed:

We need to perform two configurations where we will define the Mesos master-related information and ZooKeeper information.

The following are the steps to work on the configuration of Marathon:

1. Edit the marathon.service file by opening the file that is inside /etc/systemd...

To check the Mesos web page, use the name on the local desktop. There will be no agent activity, so we will start our Mesos slave by typing sudo service mesos-slave start:

Check the status with sudo service mesos-slave status command. It will be in running state. Next, check the Mesos console. Click the Agents tab, and you will see that we have activated the agent that is running on the Mesos host. Hence, the host, CPUs, memory, and disk space are visible, which are registered:

Click on the Frameworks tab, and then you will see which frameworks are available via the Mesos console:

To deploy the sample nginx application, carry out the following steps:

1. Click on Create Application...

In this chapter, you learned how to create a VPC as well as multiple subnets in different availability zones. We also created a route table so traffic can flow between them, then we created an internet gateway to give internet access to our instances. We learned how to create an EC2 instance inside our VPC and subnet. We created multiple instances and different subnets to achieve high availability and reliability.

You also saw that you can connect to an EC2 instance in two ways: by using PuTTY, or by converting them into a PPK file and using PuTTY to connect to your instances and next using SSH. If you have Apache, you can use SSH to connect to your instances.

In the next chapter, we will see how to install Mesos and our required components on all these servers.