Part II vsphere DRS (distributed resource Scheduling)
Chapter One introduction of Vsphere DRS
VMware vsphere Distributed Resource Scheduling (DRS) is an infrastructure service that runs on a VMware vcenter server (vcenter), and DRS aggregates ESXi host resources, which are automatically allocated to virtual machines by monitoring utilization, and can continuously carry on the virtual machine resource optimization across the ESXi host.
DRS calculates resources for virtual machines based on static resource configuration, dynamic resource utilization, and dynamic resource contention levels.
DRS leverages Vmotion to meet the available resource quotas for virtual machines in a cluster, vmotion migrate virtual machines to ESXi hosts with more available resources to balance utilization, and DRS can use vmotion for decentralized resource consolidation, making virtual machines more resource-intensive.
Basic design Principles
We recommend that you start DRS to improve the consolidation rate
Cluster resource management level
The cluster group ESXi the various resources of the host and treats them as a pool of resources. DRS aggregates these resources, which are equivalent to a large host for a virtual machine, which allows all hosts in the DRS cluster to create resource pools and apply resource-allocation policies. It may not be necessary to point out that, although resources are pooled by DRS but virtual machine resource usage cannot exceed its host resources, DRS relies on local host resource scheduling to allocate resources, and in addition to resource pooling and resource allocation policies, DRS provides the following resource management capabilities.
Initial drop location: When the virtual machine in the cluster is powered on, DRS places the virtual machine on the appropriate host, or suggests it according to the automation level.
Load balancing: DRS allocates the workload of the virtual machines on the ESXi host within the cluster, the workloads and available resources of the DRS continuous health activity, DRS compares the monitoring results to complete the desired resource execution and allocation, or recommends that the virtual machine be migrated to ensure that the workload gets the resources The goal is to maximize performance.
Power Management: When distributed power Management (DPM) is enabled, DRS compares cluster-level and host-level capacity and virtual machine requirements, including recent historical requirements, historical locations, or if excess capacity is detected, recommend ESXi hosts in standby mode or require additional resources, and recommend opening ESXi hosts.
Cluster Maintenance Mode: DRS evaluates a set of hosts that can be placed in maintenance mode while also speeding up the repair process for VMware Update Manager. When determining which hosts are in the maintenance mode, DRS will also consider factors such as ha,dpm,ft (Fault tolerance), vmotion, and so on.
Rule revision: When the host enters maintenance mode or Standby mode, DRS will evacuate the user request to reassign the virtual machine to ESXi, host, and migrate the virtual machines to adhere to user-defined association or association rules.
Support Agent Virtual machine: The proxy virtual machine is a virtual machine that is configured and activated on a per-host basis, and is a solution for the ESX Agent Management tool, and DRS and DPM fully support the ESX agent and compliance requirements of the ESX Agent virtual machine, DRS and DPM understand as follows:
If the virtual machine does not power up, the proxy virtual machine reservation rule should still comply
Host into maintenance mode or Standby mode, Agent virtual machine will not be withdrawn
The agent virtual machine must be available before the virtual machine completes migration or the host is turned on
Demand
For DRS to work properly, the virtual infrastructure needs to meet the following requirements:
ESXi in a cluster
VMware vcenter Server
VMware vsphere Enterprise or Enterprise Plus License
Meet Vmotion requirements (not mandatory, but strongly recommended)
Shared storage is accessible to ESXi hosts within a cluster
Private Migration Network
Gigabit Network
CPU compatibility
DRS allows automatic load balancing, vmotion is required, and if the initial position is set, vmotion is not required
Basic design Principles
Configuration vmotion will benefit fully from Drs.