Place: reserve workload resources

From the Workload Placement Page, you can set up reservations to save the resources you will need to deploy VMs at a future date. Turbonomic calculates optimal placement for these VMs and then reserves the host and storage resources that they need.

To reserve VMs, you will need to choose a VM template, specify any placement constraints, set how many instances to reserve, and then indicate whether to reserve now or in the future. Because reserved VMs do not yet exist, they do not participate in the real-time market.

About VM templates for reservations

VM templates specify the resource requirements for each reserved VM, including:

  • Compute and storage resources that are allocated to each VM

  • Consumed factor. This is the percentage of allocated CPU, memory, or storage that the reserved VM will utilize.

For more information about these templates, see VM Template Settings.

About placement of reserved VMs

To determine the best placement for the VMs you want to reserve, Turbonomic runs a plan that uses the last-generated data in nightly-run headroom plans.

Note:

If you change your environment by adding targets or changing policies, wait until the next run of headroom plans for the affected scope before you create reservations.

When making placement decisions, Turbonomic considers the following:

  • Placement constraints set in the reservation

  • Demand capacity

    Turbonomic calculates demand based on the resource allocation and consumed factor set in VM templates. For example, to create a reserved VM from a template that assigns 3 GB of virtual memory and a consumed factor of 50%, Turbonomic calculates 1.5 GB of demand capacity for the reservation.

  • Overprovisioned capacity

    For reserved VMs, this corresponds to the resource allocation set in VM templates. Continuing from the previous example, Turbonomic assumes 3 GB of overprovisioned capacity for a reserved VM created from a template that assigns 3 GB of virtual memory.

    For providers (hosts and storage), Turbonomic calculates overprovisioned capacity. The default overprovisioned capacity is 1000% for host Mem and CPU, and 200% for storage. A host with 512 GB of memory has an overprovisioned capacity of 5 TB (5120 GB).

Providers must have sufficient demand and overprovisioned capacity to place a reservation. Turbonomic analyzes the current and historical utilization of cluster, host, and storage resources to identify viable providers for the VMs when they are deployed to your on-prem environment. In this way, Turbonomic can prevent congestion issues after you deploy the VMs.

Note:

Turbonomic persists historical utilization data in its database so it can continue to calculate placements accurately when market analysis restarts.

The initial placement attempt either succeeds or fails.

  • Successful Initial Placements

    If the initial placement attempt is successful, Turbonomic adds the reserved VM to your inventory.

    In the previous example, a reserved VM that requires 1.5 GB of demand capacity and 3 GB of overprovisioned capacity can be placed on a host with 512 GB of memory (5 TB of overprovisioned capacity), assuming no constraints prevent the placement.

    Note that actual and reserved VMs share the same resources on providers. This means that provider capacity changes as demand from the actual VMs changes. Turbonomic polls your environment once per day to identify changes in provider capacity. It then evaluates if it can continue to place the reserved VMs within the same cluster, and then shows the latest placement status.

    For example, if the host for a reserved VM is congested at the time of polling, Turbonomic might decide to move the VM to another host in the cluster that has sufficient capacity. In this case, the placement status stays the same (Reserved). Should you decide to deploy the VM at that point, you need to deploy it to the new host. If, on the other hand, there is no longer a suitable host in the cluster, the placement fails and the status changes to Placement Failed. Deploying the VM at that point results in congestion. Turbonomic does not retry fulfilling the reservation.

    Reserved VMs are listed on the Workload Placement page. You can also get a list of reserved VMs, with information about each reservation, by making the following API request to the /reservations endpoint: 

    GET https://10.10.10.10/api/v3/reservations?status=RESERVED

    However, reserved VMs are not visible in your application map in the Turbonomic user interface.

  • Failed Initial Placements

    If the initial placement attempt is unsuccessful (for example, if all providers have seen historical congestion), the Workload Placement page shows that the placement has failed and Turbonomic does not retry fulfilling the reservation. You can get a list of VMs for which the placement failed by making the following API request to the /reservations endpoint: 

    GET https://10.10.10.10/api/v3/reservations?status=PLACEMENT_FAILED

Current and future reservations

You can create a current or future reservation from the Workload Placement Page.

  • Current Reservation

    Turbonomic calculates placement immediately and then adds the reserved VMs to your inventory if placement is successful.

    This reservation stays in effect for 24 hours, or until you delete it.

  • Future Reservation

    Set the reservation for some time in the future.

    Turbonomic does not calculate placement at this time — the future reservation saves the definition, and Turbonomic will calculate placement at the time of the reservation start date.

    This reservation stays in effect for the duration that you set, or until you delete it.

Displaying the Workload Placement page

Place button

To see the reservations that are in effect and to create new reservations, click the PLACE button in the Navigation Menu.

Workload Placement page