DRS

As discussed in Chapter 11, Configuring and Managing vSphere 6.7, a vSphere cluster is a collection of ESXi hosts that share resources and a management interface. Some of the vSphere's features are available only on the cluster level and DRS is one of them. Once the DRS is enabled on the cluster, the capability to automatically balance loads across the ESXi hosts will be available. vSphere DRS provides two main functions:

  • Executing the placement of the just-powered-on VM on a specific host in the cluster
  • Periodically (every 5 minutes by default), DRS checks the load on the cluster, providing recommendations for migration or automatically migrate the VM (using vMotion) to get a balanced cluster
If you have a DRS-enabled cluster and one of the hosts is heavily loaded compared to other host members, you might notice DRS doesn't vMotion any running VM off the host, leaving the workload unchanged. Until the ESXi host can satisfy resource demand from the VM, DRS doesn't perform any action. DRS ensures that the cluster is balanced, regardless of the workloads distributed on individual host members. To get balanced clusters and host members, there are third-party applications that provide real-time automation to allocate resources efficiently.

You can find a basic overview of your cluster's balance in the Summary tab of your vSphere cluster, as shown in the following screenshot:

When a VM in a DRS-enabled cluster is powered on, the vCenter Server checks whether the cluster has enough resources to support the VM, that is, it performs admission control. If the available resources in the cluster are not sufficient to power on the VM, a warning message appears. If the resources are sufficient to support the VM, a recommendation on which host the VM should run is generated by the DRS and, based on the automation level configured in the cluster, one of the following actions is taken:

  • The placement recommendation is executed automatically
  • The placement recommendation is displayed, leaving the user with the option to accept or override

When DRS is disabled, no recommendations are provided, and VMs are not moved among the cluster's hosts.
To enable DRS in a cluster, proceed as follows:

  1. From the vSphere client, log into vCenter Server and right-click the cluster in which you want to enable DRS and select Settings.
  2. Under Services, select vSphere DRS and click the Edit button.
  1. Enable the vSphere DRS option, and from the Automation drop-down menu, select the level of automation you want to apply to the cluster:
    • Manual: Placement and migration recommendations are displayed, but must be applied manually
    • Partially Automated: The initial placement is performed automatically, but migration recommendations are only displayed without running
    • Fully Automated: Placement and migration recommendations run automatically:

  1. Select your Migration Threshold:
    • Priority 1: Those recommendations are not based on cluster imbalance but rather on system requirements such as anti-affinity and affinity rules, maintenance mode, or DPM.
    • Priority 2 to 5: Based on the cost/benefit of the VM migration, DRS rules are generated for priority 2 to 5. This configuration option basically adjusts the sensitivity of the DRS.

For more information about DRS migration thresholds, feel free to visit the following website: https://blogs.vmware.com/vsphere/2016/10/drs-migration-thresholds.html.

  1. If you have vRealize Operations in place, you can also enable Predictive DRS. With Predictive DRS, migration can occur even before your cluster is unbalanced based on historical data.
  2. Virtual Machine Automation is enabled by default, and it gives you the option to override the automation level on a per-VM basis.

If you need to, you can also check the advanced parameters of DRS:

  • VM Distribution: This setting overrules this logic and incurs the cost of migration to achieve a more even distribution of VMs. Please note that this setting will still keep VM happiness in mind, so even the distribution of VMs is done on a best-effort basis.
  • Memory Metric for Load Balancing: This setting can be helpful for environments that attempt to minimize the impact of host failures or attempt to balance the load on network IP connections across the ESXi hosts in the cluster. Please note that this setting can increase the number of VM migrations without specifically benefitting the application's performance.
  • CPU Over-Commitment: By default, DRS uses a default CPU over-commit (vCPU to pCPU) ratio that is approximately 80 to 1. A latency-sensitive workload can benefit from a lower CPU over-commit ratio by reducing the number of vCPUs waiting to be scheduled. This setting limits the number of vCPUs that can be powered on in the vSphere cluster.
If DRS is disabled, resource pools configured in the cluster are removed.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset