DocsKnowledge BaseAPI Explorer
Documentation›HCI›Virtualization Edition›User GuideHCI

Cluster DRS

Dynamic Resource Scheduling (DRS) monitors CPU or memory load on hosts at the cluster level and dynamically adjusts VM services running on hosts according to the configured scheduling policy.

ZStack Cube Virtualization Edition supports both manual and automatic scheduling policies. A manual scheduling policy provides scheduling suggestions, and administrators can manually migrate VMs according to the suggestions. An automatic scheduling policy allows the system to execute resource scheduling based on the scheduling algorithm. Both scheduling policies can balance cluster loads and improve platform stability.

Configure DRS

Before you begin

  • Only ZCE distributed storage or SAN storage data stores are attached to the cluster.
  • All hosts in the cluster use the same CPU model.

Procedure

  1. Go to the DRS configuration entry.
    • On the DRS tab of the target cluster, click Enable.
    • In the navigation bar on the left side of the platform page, choose Reliability > DRS Policy. Then, for the target cluster, choose Actions > Modify Policy.
  2. Configure the DRS policy.
    Option Description
    DRS Mode
    • Manual Scheduling: If the CPU or memory usage of the hosts in the cluster reaches the specified threshold, administrators manually schedule the resources based on the scheduling suggestions.
    • Automatic Scheduling: If the CPU or memory usage of the hosts in the cluster reaches the specified threshold, the system auto-schedules the resources based on the scheduling algorithm.
    Monitoring Item
    • CPU Utilization: Defines the trigger condition for the CPU utilization monitoring item.
    • Memory Utilization: Defines the trigger condition for the memory utilization monitoring item.
    • CPU or Memory Utilization: Triggers scheduling evaluation when either resource utilization reaches the trigger condition.
    Duration Defines how long the threshold must be met. The unit can be seconds, minutes, or hours. The maximum duration is 6 hours.
    VM Migration Concurrency Specifies the maximum number of VMs that can be migrated from their current hosts to recommended target hosts at the same time during DRS. The default value is 1, and the maximum value is 100.
    Cluster Scanning Interval Specifies the interval for scanning the cluster balance status during DRS. The default value is 10 minutes. The value cannot be less than 5 minutes or greater than 168 hours.
    Note: If any host in the cluster reaches the threshold of the monitoring item trigger condition and the configured duration is met, the cluster is considered imbalanced and scheduling suggestions are generated.
  3. Confirm the configuration and click OK.

Perform DRS Operations

About this task

After DRS is configured, you can scan the cluster balance status, execute scheduling suggestions, view scheduling tasks, reconfigure the policy, or disable DRS as needed.

  • Manually scan the cluster balance status.
  • When the scheduling policy is manual, migrate VMs to recommended hosts according to the scheduling suggestions to balance the cluster load.
  • View scheduling tasks.

    Click O&M Management > Task > Scheduling Task to view scheduling execution history, execution results, execution time, and other information.

  • Reconfigure the DRS policy.
  • Disable Cluster DRS.
    Note: After DRS is disabled, host loads might not be balanced in high-load scenarios, which can affect service performance. Proceed with caution.
PreviousResource Allocation and OvercommitNextVM Scheduling Policy