Transcript
your OpenNebula hybrid cloud using Scaleway. This integration represents a significant step forward in hybrid and edge cloud computing deployments, offering unmatched flexibility, scalability, and performance. Let's dive and see what makes this combination so powerful. Let's discuss the key elements of this integration. The first is OneProvision, an OpenNebula's tool that makes it possible to expand your hybrid cloud with minimal effort. OneProvision, by using Terraform for automatic resource deployment and Ansible for automatic host configuration, can automate the provisioning of hosts and other resources in a single click, ensuring that your cloud is ready to meet the demand at any moment. OpenNebula provides a unified platform that enables rapid deployment of resources across cloud and edge environments. Whether you need to scale up in a central data center or expand to remote edge locations, this integration delivers instant results. With these tools, you gain the ability to automate both host configuration and resource deployment, significantly reducing setup time and operational complexity. Let's now look at the real-world example on how this integration works. Imagine a hybrid cloud environment spanning across the multiple locations. The bare-metal instances in Frankfurt hosted by AWS, in Amsterdam hosted by Equinix, and two locations, Paris and Warsaw, hosted by Scaleway, to showcase the OneProvision capabilities and the new Scaleway provision driver, we are going to simultaneously deploy these four bare-metal clusters from scratch, followed by the real-life application deployment, the Distributed Object Storage Platform, MiniIO, to utilize the full capability of this infrastructure. This is the main interface of OpenNebula. Currently, we are running the default cluster with the default set of resources. Three datastores. The host view under the Infrastructure section is empty, as we currently don't have any hosts running. Same is applicable to the virtual network view. In order to provision the bare-metal instances at the cloud provider, you must have the credentials configured under the Infrastructure, Providers in the first place. You must supply the authentication credentials for each new provider, and for the purpose of this demonstration, we've pre-created all necessary providers up front. Let's start with deploying clusters. To create and configure the cluster, you should navigate to the provisions under the Infrastructure. First cluster to deploy is going to be Scaleway cluster. We are going to stick to defaults to avoid any complexity during the demonstration. By default, Scaleway is deploying with local storage and VXLAN networking. The Warsaw region is going to be the first one. Let's have two servers in the cluster, and two public IPs. Once all is set, start a deployment. It's going to take a while. Let's repeat the same steps and deploy a similar cluster in Paris region. Next deployment is going to take place in AWS. When deploying AWS, you are supposed to pick the cluster type. The Hyper-Converged Infrastructure cluster is equipped with Ceph Datastore, while Edge cluster is going to have the same configuration as AWS. Let's have two servers in the cluster, and two public IPs. The Hyper-Converged Infrastructure cluster is equipped with Ceph Datastore, while Edge cluster is going to have the same configuration as the ones we have in Scaleway. Let's proceed with Edge cluster. Pick the pre-configured provider in the Frankfurt region. We are going to proceed with two instances and two public IPs, leaving the machine image, instance type and virtualization technology on default. The Terraform kicked in and is going to perform all necessary tasks in the background. Let's proceed with Equinix, keeping configuration alike the all previous deployments, two IPs and two servers. The time to deploy the cluster depends on the provider and some other factors like networks, resource limitations and etc. Overall, it should be in between 6 to 12 minutes. If we will return back to the clusters page, you can see that four new clusters have been registered. By opening the cluster, you can see the cluster resources. Please note that some of these are still deploying, thus the status light is red. In the provisioning state, we can see that the clusters were deployed successfully and are ready to be used. Each host provides the information of its public IP, the state and numerous CPU assignments as well as overall CPU and memory statistics. Now, when we have all the resources online, let's deploy an actual application. In the scope of this demonstration, we are going to deploy the distributed object storage. Please have a look at the deployment schema. You can see that we are going to deploy the MiniIO virtual machines to all of the nodes in all four clusters. To achieve the goal of deploying a synchronized, orchestrated and scalable application, we are going to use OpenNebulas OneFlow service that takes care of the multi-VM deployments. In OpenNebula terminology, the multi-VM deployments are being called services. Service template depends on the VM template. Inside the VMs there is a contextualization package installed which allows to pass the configuration and perform certain tasks with the virtual machine. For example, automate the startup of the MiniIO service and register them in the correct MiniIO cluster. We already had downloaded the MiniIO appliance to the OpenNebula, thus we can find it under the templates, service templates page. If it's not there, you can always download it from the official OpenNebula marketplace if you navigate to storage, apps and search for MiniIO service template. As you can see, during the deployment we have four separate roles that represent four different locations we have. Each location is supposed to host two virtual machines. Once all virtual machines are deployed, we can access the MiniIO service using the assigned public IP address. Once logged in, you can see the list containing all MiniIO nodes in the cluster and their respective FQDNs. Let's create a new bucket. As the final test, let's upload a few files to the bucket and then let's access them from the object storage. Going back to the OpenNebula's web UI, you can see some of the network traffic data that spiked during our upload and download actions. This demonstration is recorded as a part of the One Next Gen project in the scope of EPSA Cloud. Thank you for watching and see you in the next video.
