- Virtual machine application performance is poor when compared to another host.
- Applications run slower than expected. When they are moved to another host, performance improves.
- Virtual machines encounter higher than expected CPU %Ready times in ESXTOP.
- Processor Power Management (Power Saving, Power Scaling) is impacting application performance on virtual machines.
Poor performance may be related to processor power management implemented either by ESXi/ESX or by the server hardware. Certain applications that are very sensitive to processing speed latencies may show less than expected performance when processor power management features are enabled. It may be necessary to to turn off ESXi/ESX and server hardware power management features to achieve the best performance for such applications.
Disabling power management usually results in more power being consumed by the system, especially when it is lightly loaded. The majority of applications benefit from the power savings offered by power management, with little or no performance impact. Therefore, if disabling power management does not realize any increased performance, VMware recommends that power management be re-enabled to reduce power consumption
Disabling ESXi/ESX Power Management
The ESXi/ESX kernel can manage processor power usage on supported hardware whose BIOS settings are configured to allow OS-controlled power management. In this case, you can select from various ESXi/ESX power management policies using the vSphere Client Host Configuration panel.
ESXi 5.0, 5.1, 5.5 and ESXi/ESX 4.1
There are four Power Management policies available in ESXi 5.0, 5.1, 5.5 and ESXi/ESX 4.1:
- High Performance
- Balanced (Default)
- Low Power
Choose High Performance to disable power management.
If you see only one policy (not supported), then the hardware either does not support power management, or its BIOS settings are configured to disallow OS-controlled power management.
Note: When using vSphere 5.x Host Profiles with ESXi 5.x hosts that support power management, the CPU power system policy is retained within the host configuration policy and can be applied to other hosts within the environment. For more information, see vSphere Host Profiles.
There are two Power Management policies available in ESXi/ESX 4.0:
Choose Static to disable power management.
Disabling Hardware Power Management
Disabling processor power management in the server hardware is vendor specific. Here are some common server platforms:
- HP ProLiant servers with the HP Power Profile option (ProLiant G6 or greater)
Review both the HP Power Profile and HP Power Regulator Mode in the BIOS settings. For HP Power Profile, you see these options:
- Balanced Power and Performance
- Minimum Power Usage
- Maximum Performance
Choose Maximum Performance to disable power management. If Custom is already selected, refer to the Power Regulator Mode options below. If you choose Custom, set the HP Power Regulator Mode to OS Control Mode.
- HP ProLiant servers without the HP Power Profile option
Review the Power Regulator Modes in the BIOS settings. You see these options:
- HP Static High Performance mode
- HP Static Low Power mode
- HP Dynamic Power Savings mode
- OS Control mode
- Ultra Low Power Mode
Choose HP Static High Performance mode to disable power management. If OS Control mode is selected, see the section above for disabling power management within the OS.
Note: Some HP BIOS settings may include an option called Ultra Low Power Mode. This only applies in HP Static Low Power Mode or HP Dynamic Power Savings Mode, and allows the CPU to transition to its lowest power state.
- Dell PowerEdge 11th Generation servers with BIOS
Review the Power Management in the BIOS settings. You see these options:
- OS Control
- Active power control
- Maximum Performance
Choose Maximum Performance to disable power management. If Custom is selected, this allows for modification of the CPU Power and Performance, Fan Power and Performance, and Memory Power and Performance Management. If using Custom, set the various subsystems mentioned previously to Maximum Performance If OS Control mode is selected, set the various subsystems to OS DBPM and see the section above for disabling power management within the OS.
- Dell PowerEdge 12th Generation servers with UEFI
Review the System Profile modes in the System Setup > System BIOS settings. You see these options:
- Performance Per Watt (DAPC-System)
- Performance Per Watt (OS)
- Dense Configuration (DAPC-System)
Choose Performance to disable power management. If Custom is selected, this allows for modification of the CPU Power Management, Memory Frequency and Turbo Boost (when applicable CPUs are used). If using Custom, set the various subsystems mentioned previously to Maximum Performance If you choose Performance Per Watt (OS), see the section above for disabling power management within the OS.
Note: Some Dell UEFI settings may include an option called Dense Configuration (DAPC-System). This power management setting is a branch of Performance Per Watt (DAPC-System) that disables the Intel Turbo Boost technology on the CPUs.
For more information, see:
- HP document – Power Regulator for ProLiant Servers.
- Dell documentation – PowerEdge 12g BIOS Turning for Performance Power.
- Dell documentation – PowerEdge 11g BIOS Settings for High Performance Computing