We are pleased to announce the General Availability (GA) of VM Cluster Node Subsetting capability on Gen 2 Exadata Cloud@Customer. Previously, Exadata Cloud@Customer customers provisioned VM clusters across all DB servers in the Exadata Infrastructure and allocated resources from each DB server. With VM Cluster Node Subsetting, you have the flexibility to choose specific DB Servers to provision the VMs in your cluster and grow or shrink your VM clusters on-demand to meet your changing business needs.
With the VM Cluster Node Subsetting capability, you can now
You can easily shrink existing provisioned VM clusters running on all DB servers in the infrastructure by terminating VMs running on specific DB Servers without affecting existing running workloads. Once the VM cluster is running on DB Servers of your choice, you can seamlessly scale resources (OCPU, memory, local storage) allocated per VM to ensure optimal resource utilization for the cluster. Newly provisioned VM clusters will only host VMs on selected DB servers. You can dynamically scale any VM cluster to add more VMs to the cluster or add more resources per VM based on changing workload requirements without disrupting your RAC-aware databases.
Let's go over the following core user journey highlights related to VM cluster node subsetting using the OCI console.
On the Exadata Infrastructure details page, you can navigate to the VM Cluster section and initiate the create VM Cluster workflow to provision a new cluster on this infrastructure.
The create VM Cluster workflow now includes selecting the DB servers on which you want to host the VMs for your new cluster.
You can specify the placement of each VM in your cluster by selecting the DB server to host the VM for this cluster. All DB Servers part of your Exadata Infrastructure are listed and available for selection to place your VM. You can see the available OCPU, memory, and local storage resources for each DB server, along with the list of existing VM clusters already hosting VMs on the respective DB Servers. Based on your isolation and co-location preferences and planned resource allocation limits, you can choose the DB servers best suited for your specific use case to be part of this VM cluster.
Once the DB servers to host VMs for the cluster are selected, you can specify the allocation for OCPU, memory, and local file system storage resources per VM using the presented controls. The maximum resources available for assignment per VM depends on the selected DB servers that will host these VMs for the cluster. The DB server with the least resources will determine the maximum limit available for allocation per VM, given the symmetric resource allocation across all VMs in the cluster.
The VM cluster details page shows the total number of VMs and the total resources allocated across all VMs after the provisioning completes successfully. You can also view the list of VMs in the cluster and their respective resource allocation, client IP address, and a hyperlink to view the DB Server hosting the VM. The DB server details page will list all the VMs from various clusters hosted on that DB Server.
You can expand a provisioned cluster on-demand by adding VMs from the cluster details page.
While adding new VMs to your provisioned cluster, you can choose the specific DB servers on which you want to add new VMs and extend your VM Cluster. DB servers already hosting a VM from a particular cluster are not available to host another VM from the same cluster. For every DB Server, you can see the available OCPU, memory, and local storage along with the list of VM clusters hosting VMs on that DB server. You can choose the DB server(s) best suited to host your newly added VM(s) based on the information presented.
Note: Newly added VMs have the same resource allocation for OCPU, memory, and local storage as existing VMs part of the cluster.
Total resources allocated across the cluster are updated to reflect the newly added VM resources. Each new VM added to the cluster is listed along with existing VMs and displays the allocated resources, assigned client IP address, and the DB server hosting the VM.
Additionally, you can navigate to a specific VM listed as part of the cluster and use the action menu dropdown for the list row to terminate the VM.
Deleting a virtual machine will terminate any database instances running on the VM and requires additional confirmation to proceed.
For your provisioned VM cluster, you can always navigate to the cluster details page and initiate a scale action to change the resource allocation for the VMs in your cluster.
The scale VM cluster resources workflow shows the number of VMs that are part of the cluster and presents controls to change the allocation for OCPU, memory, and local file system storage resources per VM. You can view the total resources allocated across all VMs in the cluster as a read-only summary similar to the view shown during cluster creation.
VM Cluster Node Subsetting capability is now available on Gen2 Exadata Cloud@Customer in all OCI commercial regions.