VMware ESXI - 6.5.1 Resource Management
vSphere Resource Management
Update 1
VMware vSphere 6.5
VMware ESXi 6.5
vCenter Server 6.5
This document supports the version of each product listed and
supports all subsequent versions until the document is
replaced by a new edition. To check for more recent editions of
this document, see http://www.vmware.com/support/pubs.
EN-002644-00
vSphere Resource Management
You can find the most up-to-date technical documentation on the VMware Web site at:
hp://www.vmware.com/support/
The VMware Web site also provides the latest product updates.
If you have comments about this documentation, submit your feedback to:
docfeedback@vmware.com
Copyright © 2006–2017 VMware, Inc. All rights reserved. Copyright and trademark information.
VMware, Inc.
3401 Hillview Ave.
Palo Alto, CA 94304
www.vmware.com
2 VMware, Inc.
Contents
About vSphere Resource Management 7
Geing Started with Resource Management 9
1
Resource Types 9
Resource Providers 9
Resource Consumers 10
Goals of Resource Management 10
Conguring Resource Allocation
2
Seings 11
Resource Allocation Shares 11
Resource Allocation Reservation 12
Resource Allocation Limit 12
Resource Allocation Seings Suggestions 13
Edit Resource Seings 13
Changing Resource Allocation Seings—Example 14
Admission Control 15
CPU Virtualization Basics 17
3
Software-Based CPU Virtualization 17
Hardware-Assisted CPU Virtualization 18
Virtualization and Processor-Specic Behavior 18
Performance Implications of CPU Virtualization 18
VMware, Inc.
Administering CPU Resources 19
4
View Processor Information 19
Specifying CPU Conguration 19
Multicore Processors 20
Hyperthreading 20
Using CPU Anity 22
Host Power Management Policies 23
Memory Virtualization Basics 27
5
Virtual Machine Memory 27
Memory Overcommitment 28
Memory Sharing 28
Types of Memory Virtualization 29
Administering Memory Resources 33
6
Understanding Memory Overhead 33
How ESXi Hosts Allocate Memory 34
3
vSphere Resource Management
Memory Reclamation 35
Using Swap Files 36
Sharing Memory Across Virtual Machines 40
Memory Compression 41
Measuring and Dierentiating Types of Memory Usage 42
Memory Reliability 43
About System Swap 43
Conguring Virtual Graphics 45
7
View GPU Statistics 45
Add an NVIDIA GRID vGPU to a Virtual Machine 45
Conguring Host Graphics 46
Conguring Graphics Devices 47
Managing Storage I/O Resources 49
8
About Virtual Machine Storage Policies 50
About I/O Filters 50
Storage I/O Control Requirements 50
Storage I/O Control Resource Shares and Limits 51
Set Storage I/O Control Resource Shares and Limits 52
Enable Storage I/O Control 52
Set Storage I/O Control Threshold Value 53
Storage DRS Integration with Storage Proles 54
Managing Resource Pools 55
9
Why Use Resource Pools? 56
Create a Resource Pool 57
Edit a Resource Pool 58
Add a Virtual Machine to a Resource Pool 58
Remove a Virtual Machine from a Resource Pool 59
Remove a Resource Pool 60
Resource Pool Admission Control 60
Creating a DRS Cluster 63
10
Admission Control and Initial Placement 63
Virtual Machine Migration 65
DRS Cluster Requirements 67
Conguring DRS with Virtual Flash 68
Create a Cluster 68
Edit Cluster Seings 69
Set a Custom Automation Level for a Virtual Machine 71
Disable DRS 72
Restore a Resource Pool Tree 72
Using DRS Clusters to Manage Resources 73
11
Adding Hosts to a Cluster 73
Adding Virtual Machines to a Cluster 75
Removing Virtual Machines from a Cluster 75
4 VMware, Inc.
Removing a Host from a Cluster 76
DRS Cluster Validity 77
Managing Power Resources 82
Using DRS Anity Rules 86
Contents
Creating a Datastore Cluster 91
12
Initial Placement and Ongoing Balancing 92
Storage Migration Recommendations 92
Create a Datastore Cluster 92
Enable and Disable Storage DRS 93
Set the Automation Level for Datastore Clusters 93
Seing the Aggressiveness Level for Storage DRS 94
Datastore Cluster Requirements 95
Adding and Removing Datastores from a Datastore Cluster 96
Using Datastore Clusters to Manage Storage Resources 97
13
Using Storage DRS Maintenance Mode 97
Applying Storage DRS Recommendations 99
Change Storage DRS Automation Level for a Virtual Machine 100
Set Up O-Hours Scheduling for Storage DRS 100
Storage DRS Anti-Anity Rules 101
Clear Storage DRS Statistics 104
Storage vMotion Compatibility with Datastore Clusters 105
Using NUMA Systems with ESXi 107
14
What is NUMA? 107
How ESXi NUMA Scheduling Works 108
VMware NUMA Optimization Algorithms and Seings 109
Resource Management in NUMA Architectures 110
Using Virtual NUMA 110
Specifying NUMA Controls 111
Advanced
15
Aributes 115
Set Advanced Host Aributes 115
Set Advanced Virtual Machine Aributes 118
Latency Sensitivity 120
About Reliable Memory 120
Fault Denitions 123
16
Virtual Machine is Pinned 124
Virtual Machine not Compatible with any Host 124
VM/VM DRS Rule Violated when Moving to another Host 124
Host Incompatible with Virtual Machine 124
Host Has Virtual Machine That Violates VM/VM DRS Rules 124
Host has Insucient Capacity for Virtual Machine 124
Host in Incorrect State 124
Host Has Insucient Number of Physical CPUs for Virtual Machine 125
VMware, Inc. 5
vSphere Resource Management
Host has Insucient Capacity for Each Virtual Machine CPU 125
The Virtual Machine Is in vMotion 125
No Active Host in Cluster 125
Insucient Resources 125
Insucient Resources to Satisfy Congured Failover Level for HA 125
No Compatible Hard Anity Host 125
No Compatible Soft Anity Host 125
Soft Rule Violation Correction Disallowed 125
Soft Rule Violation Correction Impact 126
DRS Troubleshooting Information 127
17
Cluster Problems 127
Host Problems 130
Virtual Machine Problems 133
Index 137
6 VMware, Inc.
About vSphere Resource Management
vSphere Resource Management describes resource management for VMware® ESXi and vCenter® Server
environments.
This documentation focuses on the following topics.
Resource allocation and resource management concepts
n
Virtual machine aributes and admission control
n
Resource pools and how to manage them
n
Clusters, vSphere® Distributed Resource Scheduler (DRS), vSphere Distributed Power Management
n
(DPM), and how to work with them
Datastore clusters, Storage DRS, Storage I/O Control, and how to work with them
n
Advanced resource management options
n
Performance considerations
n
Intended Audience
This information is for system administrators who want to understand how the system manages resources
and how they can customize the default behavior. It’s also essential for anyone who wants to understand
and use resource pools, clusters, DRS, datastore clusters, Storage DRS, Storage I/O Control, or vSphere
DPM.
VMware, Inc.
This documentation assumes you have a working knowledge of VMware ESXi and of vCenter Server.
Task instructions in this guide are based on the vSphere Web Client. You can also perform most of the tasks
in this guide by using the new vSphere Client. The new vSphere Client user interface terminology, topology,
and workow are closely aligned with the same aspects and elements of the vSphere Web Client user
interface. You can apply the vSphere Web Client instructions to the new vSphere Client unless otherwise
instructed.
N Not all functionality in the vSphere Web Client has been implemented for the vSphere Client in the
vSphere 6.5 release. For an up-to-date list of unsupported functionality, see Functionality Updates for the
vSphere Client Guide at hp://www.vmware.com/info?id=1413.
7
vSphere Resource Management
8 VMware, Inc.
Getting Started with Resource
Management 1
To understand resource management, you must be aware of its components, its goals, and how best to
implement it in a cluster seing.
Resource allocation seings for a virtual machine (shares, reservation, and limit) are discussed, including
how to set them and how to view them. Also, admission control, the process whereby resource allocation
seings are validated against existing resources is explained.
Resource management is the allocation of resources from resource providers to resource consumers.
The need for resource management arises from the overcommitment of resources—that is, more demand
than capacity and from the fact that demand and capacity vary over time. Resource management allows you
to dynamically reallocate resources, so that you can more eciently use available capacity.
This chapter includes the following topics:
“Resource Types,” on page 9
n
“Resource Providers,” on page 9
n
“Resource Consumers,” on page 10
n
“Goals of Resource Management,” on page 10
n
Resource Types
Resources include CPU, memory, power, storage, and network resources.
N ESXi manages network bandwidth and disk resources on a per-host basis, using network trac
shaping and a proportional share mechanism, respectively.
Resource Providers
Hosts and clusters, including datastore clusters, are providers of physical resources.
For hosts, available resources are the host’s hardware specication, minus the resources used by the
virtualization software.
A cluster is a group of hosts. You can create a cluster using vSphere Web Client, and add multiple hosts to
the cluster. vCenter Server manages these hosts’ resources jointly: the cluster owns all of the CPU and
memory of all hosts. You can enable the cluster for joint load balancing or failover. See Chapter 10, “Creating
a DRS Cluster,” on page 63 for more information.
A datastore cluster is a group of datastores. Like DRS clusters, you can create a datastore cluster using the
vSphere Web Client, and add multiple datstores to the cluster. vCenter Server manages the datastore
resources jointly. You can enable Storage DRS to balance I/O load and space utilization. See Chapter 12,
“Creating a Datastore Cluster,” on page 91.
VMware, Inc.
9
vSphere Resource Management
Resource Consumers
Virtual machines are resource consumers.
The default resource seings assigned during creation work well for most machines. You can later edit the
virtual machine seings to allocate a share-based percentage of the total CPU, memory, and storage I/O of
the resource provider or a guaranteed reservation of CPU and memory. When you power on that virtual
machine, the server checks whether enough unreserved resources are available and allows power on only if
there are enough resources. This process is called admission control.
A resource pool is a logical abstraction for exible management of resources. Resource pools can be grouped
into hierarchies and used to hierarchically partition available CPU and memory resources. Accordingly,
resource pools can be considered both resource providers and consumers. They provide resources to child
resource pools and virtual machines, but are also resource consumers because they consume their parents’
resources. See Chapter 9, “Managing Resource Pools,” on page 55.
ESXi hosts allocate each virtual machine a portion of the underlying hardware resources based on a number
of factors:
Resource limits dened by the user.
n
Total available resources for the ESXi host (or the cluster).
n
Number of virtual machines powered on and resource usage by those virtual machines.
n
Overhead required to manage the virtualization.
n
Goals of Resource Management
When managing your resources, you must be aware of what your goals are.
In addition to resolving resource overcommitment, resource management can help you accomplish the
following:
Performance Isolation: Prevent virtual machines from monopolizing resources and guarantee
n
predictable service rates.
Ecient Usage: Exploit undercommied resources and overcommit with graceful degradation.
n
Easy Administration: Control the relative importance of virtual machines, provide exible dynamic
n
partitioning, and meet absolute service-level agreements.
10 VMware, Inc.
Configuring Resource Allocation
Settings 2
When available resource capacity does not meet the demands of the resource consumers (and virtualization
overhead), administrators might need to customize the amount of resources that are allocated to virtual
machines or to the resource pools in which they reside.
Use the resource allocation seings (shares, reservation, and limit) to determine the amount of CPU,
memory, and storage resources provided for a virtual machine. In particular, administrators have several
options for allocating resources.
Reserve the physical resources of the host or cluster.
n
Set an upper bound on the resources that can be allocated to a virtual machine.
n
Guarantee that a particular virtual machine is always allocated a higher percentage of the physical
n
resources than other virtual machines.
This chapter includes the following topics:
“Resource Allocation Shares,” on page 11
n
“Resource Allocation Reservation,” on page 12
n
“Resource Allocation Limit,” on page 12
n
“Resource Allocation Seings Suggestions,” on page 13
n
“Edit Resource Seings,” on page 13
n
“Changing Resource Allocation Seings—Example,” on page 14
n
“Admission Control,” on page 15
n
Resource Allocation Shares
Shares specify the relative importance of a virtual machine (or resource pool). If a virtual machine has twice
as many shares of a resource as another virtual machine, it is entitled to consume twice as much of that
resource when these two virtual machines are competing for resources.
Shares are typically specied as High, Normal, or Low and these values specify share values with a 4:2:1
ratio, respectively. You can also select Custom to assign a specic number of shares (which expresses a
proportional weight) to each virtual machine.
Specifying shares makes sense only with regard to sibling virtual machines or resource pools, that is, virtual
machines or resource pools with the same parent in the resource pool hierarchy. Siblings share resources
according to their relative share values, bounded by the reservation and limit. When you assign shares to a
virtual machine, you always specify the priority for that virtual machine relative to other powered-on
virtual machines.
VMware, Inc.
11
vSphere Resource Management
The following table shows the default CPU and memory share values for a virtual machine. For resource
pools, the default CPU and memory share values are the same, but must be multiplied as if the resource
pool were a virtual machine with four virtual CPUs and 16 GB of memory.
Table 2‑1. Share Values
Setting CPU share values Memory share values
High 2000 shares per virtual CPU 20 shares per megabyte of congured virtual
Normal 1000 shares per virtual CPU 10 shares per megabyte of congured virtual
Low 500 shares per virtual CPU 5 shares per megabyte of congured virtual machine
For example, an SMP virtual machine with two virtual CPUs and 1GB RAM with CPU and memory shares
set to Normal has 2x1000=2000 shares of CPU and 10x1024=10240 shares of memory.
N Virtual machines with more than one virtual CPU are called SMP (symmetric multiprocessing)
virtual machines. ESXi supports up to 128 virtual CPUs per virtual machine.
The relative priority represented by each share changes when a new virtual machine is powered on. This
aects all virtual machines in the same resource pool. All of the virtual machines have the same number of
virtual CPUs. Consider the following examples.
machine memory.
machine memory.
memory.
Two CPU-bound virtual machines run on a host with 8GHz of aggregate CPU capacity. Their CPU
n
shares are set to Normal and get 4GHz each.
A third CPU-bound virtual machine is powered on. Its CPU shares value is set to High, which means it
n
should have twice as many shares as the machines set to Normal. The new virtual machine receives
4GHz and the two other machines get only 2GHz each. The same result occurs if the user species a
custom share value of 2000 for the third virtual machine.
Resource Allocation Reservation
A reservation species the guaranteed minimum allocation for a virtual machine.
vCenter Server or ESXi allows you to power on a virtual machine only if there are enough unreserved
resources to satisfy the reservation of the virtual machine. The server guarantees that amount even when the
physical server is heavily loaded. The reservation is expressed in concrete units (megaher or megabytes).
For example, assume you have 2GHz available and specify a reservation of 1GHz for VM1 and 1GHz for
VM2. Now each virtual machine is guaranteed to get 1GHz if it needs it. However, if VM1 is using only
500MHz, VM2 can use 1.5GHz.
Reservation defaults to 0. You can specify a reservation if you need to guarantee that the minimum required
amounts of CPU or memory are always available for the virtual machine.
Resource Allocation Limit
Limit species an upper bound for CPU, memory, or storage I/O resources that can be allocated to a virtual
machine.
A server can allocate more than the reservation to a virtual machine, but never allocates more than the limit,
even if there are unused resources on the system. The limit is expressed in concrete units (megaher,
megabytes, or I/O operations per second).
CPU, memory, and storage I/O resource limits default to unlimited. When the memory limit is unlimited,
the amount of memory congured for the virtual machine when it was created becomes its eective limit.
12 VMware, Inc.
In most cases, it is not necessary to specify a limit. There are benets and drawbacks:
Benets — Assigning a limit is useful if you start with a small number of virtual machines and want to
n
manage user expectations. Performance deteriorates as you add more virtual machines. You can
simulate having fewer resources available by specifying a limit.
Drawbacks — You might waste idle resources if you specify a limit. The system does not allow virtual
n
machines to use more resources than the limit, even when the system is underutilized and idle
resources are available. Specify the limit only if you have good reasons for doing so.
Resource Allocation Settings Suggestions
Select resource allocation seings (reservation, limit and shares) that are appropriate for your ESXi
environment.
The following guidelines can help you achieve beer performance for your virtual machines.
Use Reservation to specify the minimum acceptable amount of CPU or memory, not the amount you
n
want to have available. The amount of concrete resources represented by a reservation does not change
when you change the environment, such as by adding or removing virtual machines. The host assigns
additional resources as available based on the limit for your virtual machine, the number of shares and
estimated demand.
When specifying the reservations for virtual machines, do not commit all resources (plan to leave at
n
least 10% unreserved). As you move closer to fully reserving all capacity in the system, it becomes
increasingly dicult to make changes to reservations and to the resource pool hierarchy without
violating admission control. In a DRS-enabled cluster, reservations that fully commit the capacity of the
cluster or of individual hosts in the cluster can prevent DRS from migrating virtual machines between
hosts.
Chapter 2 Configuring Resource Allocation Settings
If you expect frequent changes to the total available resources, use Shares to allocate resources fairly
n
across virtual machines. If you use Shares, and you upgrade the host, for example, each virtual machine
stays at the same priority (keeps the same number of shares) even though each share represents a larger
amount of memory, CPU, or storage I/O resources.
Edit Resource Settings
Use the Edit Resource Seings dialog box to change allocations for memory and CPU resources.
Procedure
1 Browse to the virtual machine in the vSphere Web Client navigator.
2 Right-click and select Edit Resource .
3 Edit the CPU Resources.
Option Description
Shares
Reservation
Limit
CPU shares for this resource pool with respect to the parent’s total. Sibling
resource pools share resources according to their relative share values
bounded by the reservation and limit. Select Low, Normal, or High, which
specify share values respectively in a 1:2:4 ratio. Select Custom to give each
virtual machine a specic number of shares, which expresses a
proportional weight.
Guaranteed CPU allocation for this resource pool.
Upper limit for this resource pool’s CPU allocation. Select Unlimited to
specify no upper limit.
VMware, Inc. 13
VM-QA
host
VM-Marketing
vSphere Resource Management
4 Edit the Memory Resources.
Option Description
Shares
Reservation
Limit
Memory shares for this resource pool with respect to the parent’s total.
Sibling resource pools share resources according to their relative share
values bounded by the reservation and limit. Select Low, Normal, or High,
which specify share values respectively in a 1:2:4 ratio. Select Custom to
give each virtual machine a specic number of shares, which expresses a
proportional weight.
Guaranteed memory allocation for this resource pool.
Upper limit for this resource pool’s memory allocation. Select Unlimited to
specify no upper limit.
5 Click OK.
Changing Resource Allocation Settings—Example
The following example illustrates how you can change resource allocation seings to improve virtual
machine performance.
Assume that on an ESXi host, you have created two new virtual machines—one each for your QA (VM-QA)
and Marketing (VM-Marketing) departments.
Figure 2‑1. Single Host with Two Virtual Machines
In the following example, assume that VM-QA is memory intensive and accordingly you want to change the
resource allocation seings for the two virtual machines to:
Specify that, when system memory is overcommied, VM-QA can use twice as much CPU and memory
n
resources as the Marketing virtual machine. Set the CPU shares and memory shares for VM-QA to
High and for VM-Marketing set them to Normal.
Ensure that the Marketing virtual machine has a certain amount of guaranteed CPU resources. You can
n
do so using a reservation seing.
Procedure
1 Browse to the virtual machines in the vSphere Web Client navigator.
2 Right-click VM-QA, the virtual machine for which you want to change shares, and select Edit .
3 Under Virtual Hardware, expand CPU and select High from the Shares drop-down menu.
4 Under Virtual Hardware, expand Memory and select High from the Shares drop-down menu.
5 Click OK.
6 Right-click the marketing virtual machine (VM-Marketing) and select Edit .
7 Under Virtual Hardware, expand CPU and change the Reservation value to the desired number.
8 Click OK.
14 VMware, Inc.
If you select the cluster’s Resource Reservation tab and click CPU, you should see that shares for VM-QA
are twice that of the other virtual machine. Also, because the virtual machines have not been powered on,
the Reservation Used elds have not changed.
Admission Control
When you power on a virtual machine, the system checks the amount of CPU and memory resources that
have not yet been reserved. Based on the available unreserved resources, the system determines whether it
can guarantee the reservation for which the virtual machine is congured (if any). This process is called
admission control.
If enough unreserved CPU and memory are available, or if there is no reservation, the virtual machine is
powered on. Otherwise, an Insufficient Resources warning appears.
N In addition to the user-specied memory reservation, for each virtual machine there is also an
amount of overhead memory. This extra memory commitment is included in the admission control
calculation.
When the vSphere DPM feature is enabled, hosts might be placed in standby mode (that is, powered o) to
reduce power consumption. The unreserved resources provided by these hosts are considered available for
admission control. If a virtual machine cannot be powered on without these resources, a recommendation to
power on sucient standby hosts is made.
Chapter 2 Configuring Resource Allocation Settings
VMware, Inc. 15
vSphere Resource Management
16 VMware, Inc.
CPU Virtualization Basics 3
CPU virtualization emphasizes performance and runs directly on the processor whenever possible. The
underlying physical resources are used whenever possible and the virtualization layer runs instructions
only as needed to make virtual machines operate as if they were running directly on a physical machine.
CPU virtualization is not the same thing as emulation. ESXi does not use emulation to run virtual CPUs.
With emulation, all operations are run in software by an emulator. A software emulator allows programs to
run on a computer system other than the one for which they were originally wrien. The emulator does this
by emulating, or reproducing, the original computer’s behavior by accepting the same data or inputs and
achieving the same results. Emulation provides portability and runs software designed for one platform
across several platforms.
When CPU resources are overcommied, the ESXi host time-slices the physical processors across all virtual
machines so each virtual machine runs as if it has its specied number of virtual processors. When an ESXi
host runs multiple virtual machines, it allocates to each virtual machine a share of the physical resources.
With the default resource allocation seings, all virtual machines associated with the same host receive an
equal share of CPU per virtual CPU. This means that a single-processor virtual machines is assigned only
half of the resources of a dual-processor virtual machine.
This chapter includes the following topics:
“Software-Based CPU Virtualization,” on page 17
n
“Hardware-Assisted CPU Virtualization,” on page 18
n
“Virtualization and Processor-Specic Behavior,” on page 18
n
“Performance Implications of CPU Virtualization,” on page 18
n
Software-Based CPU Virtualization
With software-based CPU virtualization, the guest application code runs directly on the processor, while the
guest privileged code is translated and the translated code runs on the processor.
The translated code is slightly larger and usually runs more slowly than the native version. As a result,
guest applications, which have a small privileged code component, run with speeds very close to native.
Applications with a signicant privileged code component, such as system calls, traps, or page table updates
can run slower in the virtualized environment.
VMware, Inc.
17
vSphere Resource Management
Hardware-Assisted CPU Virtualization
Certain processors provide hardware assistance for CPU virtualization.
When using this assistance, the guest can use a separate mode of execution called guest mode. The guest
code, whether application code or privileged code, runs in the guest mode. On certain events, the processor
exits out of guest mode and enters root mode. The hypervisor executes in the root mode, determines the
reason for the exit, takes any required actions, and restarts the guest in guest mode.
When you use hardware assistance for virtualization, there is no need to translate the code. As a result,
system calls or trap-intensive workloads run very close to native speed. Some workloads, such as those
involving updates to page tables, lead to a large number of exits from guest mode to root mode. Depending
on the number of such exits and total time spent in exits, hardware-assisted CPU virtualization can speed up
execution signicantly.
Virtualization and Processor-Specific Behavior
Although VMware software virtualizes the CPU, the virtual machine detects the specic model of the
processor on which it is running.
Processor models might dier in the CPU features they oer, and applications running in the virtual
machine can make use of these features. Therefore, it is not possible to use vMotion® to migrate virtual
machines between systems running on processors with dierent feature sets. You can avoid this restriction,
in some cases, by using Enhanced vMotion Compatibility (EVC) with processors that support this feature.
See the vCenter Server and Host Management documentation for more information.
Performance Implications of CPU Virtualization
CPU virtualization adds varying amounts of overhead depending on the workload and the type of
virtualization used.
An application is CPU-bound if it spends most of its time executing instructions rather than waiting for
external events such as user interaction, device input, or data retrieval. For such applications, the CPU
virtualization overhead includes the additional instructions that must be executed. This overhead takes CPU
processing time that the application itself can use. CPU virtualization overhead usually translates into a
reduction in overall performance.
For applications that are not CPU-bound, CPU virtualization likely translates into an increase in CPU use. If
spare CPU capacity is available to absorb the overhead, it can still deliver comparable performance in terms
of overall throughput.
ESXi supports up to 128 virtual processors (CPUs) for each virtual machine.
N Deploy single-threaded applications on uniprocessor virtual machines, instead of on SMP virtual
machines that have multiple CPUs, for the best performance and resource use.
Single-threaded applications can take advantage only of a single CPU. Deploying such applications in dualprocessor virtual machines does not speed up the application. Instead, it causes the second virtual CPU to
use physical resources that other virtual machines could otherwise use.
18 VMware, Inc.
Administering CPU Resources 4
You can congure virtual machines with one or more virtual processors, each with its own set of registers
and control structures.
When a virtual machine is scheduled, its virtual processors are scheduled to run on physical processors. The
VMkernel Resource Manager schedules the virtual CPUs on physical CPUs, thereby managing the virtual
machine’s access to physical CPU resources. ESXi supports virtual machines with up to 128 virtual CPUs.
This chapter includes the following topics:
“View Processor Information,” on page 19
n
“Specifying CPU Conguration,” on page 19
n
“Multicore Processors,” on page 20
n
“Hyperthreading,” on page 20
n
“Using CPU Anity,” on page 22
n
“Host Power Management Policies,” on page 23
n
View Processor Information
You can access information about current CPU conguration in the vSphere Web Client.
Procedure
1 Browse to the host in the vSphere Web Client navigator.
2 Click and expand Hardware.
3 Select Processors to view the information about the number and type of physical processors and the
number of logical processors.
N In hyperthreaded systems, each hardware thread is a logical processor. For example, a dual-core
processor with hyperthreading enabled has two cores and four logical processors.
Specifying CPU Configuration
You can specify CPU conguration to improve resource management. However, if you do not customize
CPU conguration, the ESXi host uses defaults that work well in most situations.
You can specify CPU conguration in the following ways:
Use the aributes and special features available through the vSphere Web Client. The
n
vSphere Web Client allows you to connect to the ESXi host or a vCenter Server system.
VMware, Inc.
19
vSphere Resource Management
Use advanced seings under certain circumstances.
n
Use the vSphere SDK for scripted CPU allocation.
n
Use hyperthreading.
n
Multicore Processors
Multicore processors provide many advantages for a host performing multitasking of virtual machines.
Intel and AMD have developed processors which combine two or more processor cores into a single
integrated circuit (often called a package or socket). VMware uses the term socket to describe a single
package which can have one or more processor cores with one or more logical processors in each core.
A dual-core processor, for example, provides almost double the performance of a single-core processor, by
allowing two virtual CPUs to run at the same time. Cores within the same processor are typically congured
with a shared last-level cache used by all cores, potentially reducing the need to access slower main memory.
A shared memory bus that connects a physical processor to main memory can limit performance of its
logical processors when the virtual machines running on them are running memory-intensive workloads
which compete for the same memory bus resources.
Each logical processor of each processor core is used independently by the ESXi CPU scheduler to run
virtual machines, providing capabilities similar to SMP systems. For example, a two-way virtual machine
can have its virtual processors running on logical processors that belong to the same core, or on logical
processors on dierent physical cores.
The ESXi CPU scheduler can detect the processor topology and the relationships between processor cores
and the logical processors on them. It uses this information to schedule virtual machines and optimize
performance.
The ESXi CPU scheduler can interpret processor topology, including the relationship between sockets, cores,
and logical processors. The scheduler uses topology information to optimize the placement of virtual CPUs
onto dierent sockets. This optimization can maximize overall cache usage, and to improve cache anity by
minimizing virtual CPU migrations.
Hyperthreading
Hyperthreading technology allows a single physical processor core to behave like two logical processors.
The processor can run two independent applications at the same time. To avoid confusion between logical
and physical processors, Intel refers to a physical processor as a socket, and the discussion in this chapter
uses that terminology as well.
Intel Corporation developed hyperthreading technology to enhance the performance of its Pentium IV and
Xeon processor lines. Hyperthreading technology allows a single processor core to execute two independent
threads simultaneously.
While hyperthreading does not double the performance of a system, it can increase performance by beer
utilizing idle resources leading to greater throughput for certain important workload types. An application
running on one logical processor of a busy core can expect slightly more than half of the throughput that it
obtains while running alone on a non-hyperthreaded processor. Hyperthreading performance
improvements are highly application-dependent, and some applications might see performance degradation
with hyperthreading because many processor resources (such as the cache) are shared between logical
processors.
N On processors with Intel Hyper-Threading technology, each core can have two logical processors
which share most of the core's resources, such as memory caches and functional units. Such logical
processors are usually called threads.
20 VMware, Inc.
Chapter 4 Administering CPU Resources
Many processors do not support hyperthreading and as a result have only one thread per core. For such
processors, the number of cores also matches the number of logical processors. The following processors
support hyperthreading and have two threads per core.
Processors based on the Intel Xeon 5500 processor microarchitecture.
n
Intel Pentium 4 (HT-enabled)
n
Intel Pentium EE 840 (HT-enabled)
n
Hyperthreading and ESXi Hosts
A host that is enabled for hyperthreading should behave similarly to a host without hyperthreading. You
might need to consider certain factors if you enable hyperthreading, however.
ESXi hosts manage processor time intelligently to guarantee that load is spread smoothly across processor
cores in the system. Logical processors on the same core have consecutive CPU numbers, so that CPUs 0 and
1 are on the rst core together, CPUs 2 and 3 are on the second core, and so on. Virtual machines are
preferentially scheduled on two dierent cores rather than on two logical processors on the same core.
If there is no work for a logical processor, it is put into a halted state, which frees its execution resources and
allows the virtual machine running on the other logical processor on the same core to use the full execution
resources of the core. The VMware scheduler properly accounts for this halt time, and charges a virtual
machine running with the full resources of a core more than a virtual machine running on a half core. This
approach to processor management ensures that the server does not violate any of the standard ESXi
resource allocation rules.
Consider your resource management needs before you enable CPU anity on hosts using hyperthreading.
For example, if you bind a high priority virtual machine to CPU 0 and another high priority virtual machine
to CPU 1, the two virtual machines have to share the same physical core. In this case, it can be impossible to
meet the resource demands of these virtual machines. Ensure that any custom anity seings make sense
for a hyperthreaded system.
Enable Hyperthreading
To enable hyperthreading, you must rst enable it in your system's BIOS seings and then turn it on in the
vSphere Web Client. Hyperthreading is enabled by default.
Consult your system documentation to determine whether your CPU supports hyperthreading.
Procedure
1 Ensure that your system supports hyperthreading technology.
2 Enable hyperthreading in the system BIOS.
Some manufacturers label this option Logical Processor, while others call it Enable Hyperthreading.
3 Ensure that hyperthreading is enabled for the ESXi host.
a Browse to the host in the vSphere Web Client navigator.
b Click .
c Under System, click Advanced System and select VMkernel.Boot.hyperthreading.
You must restart the host for the seing to take eect. Hyperthreading is enabled if the value is
true.
4 Under Hardware, click Processors to view the number of Logical processors.
Hyperthreading is enabled.
VMware, Inc. 21
vSphere Resource Management
Using CPU Affinity
By specifying a CPU anity seing for each virtual machine, you can restrict the assignment of virtual
machines to a subset of the available processors in multiprocessor systems. By using this feature, you can
assign each virtual machine to processors in the specied anity set.
CPU anity species virtual machine-to-processor placement constraints and is dierent from the
relationship created by a VM-VM or VM-Host anity rule, which species virtual machine-to-virtual
machine host placement constraints.
In this context, the term CPU refers to a logical processor on a hyperthreaded system and refers to a core on
a non-hyperthreaded system.
The CPU anity seing for a virtual machine applies to all of the virtual CPUs associated with the virtual
machine and to all other threads (also known as worlds) associated with the virtual machine. Such virtual
machine threads perform processing required for emulating mouse, keyboard, screen, CD-ROM, and
miscellaneous legacy devices.
In some cases, such as display-intensive workloads, signicant communication might occur between the
virtual CPUs and these other virtual machine threads. Performance might degrade if the virtual machine's
anity seing prevents these additional threads from being scheduled concurrently with the virtual
machine's virtual CPUs. Examples of this include a uniprocessor virtual machine with anity to a single
CPU or a two-way SMP virtual machine with anity to only two CPUs.
For the best performance, when you use manual anity seings, VMware recommends that you include at
least one additional physical CPU in the anity seing to allow at least one of the virtual machine's threads
to be scheduled at the same time as its virtual CPUs. Examples of this include a uniprocessor virtual
machine with anity to at least two CPUs or a two-way SMP virtual machine with anity to at least three
CPUs.
Assign a Virtual Machine to a Specific Processor
Using CPU anity, you can assign a virtual machine to a specic processor. This allows you to restrict the
assignment of virtual machines to a specic available processor in multiprocessor systems.
Procedure
1 Find the virtual machine in the vSphere Web Client inventory.
a To nd a virtual machine, select a data center, folder, cluster, resource pool, or host.
b Click the Related Objects tab and click Virtual Machines.
2 Right-click the virtual machine and click Edit .
3 Under Virtual Hardware, expand CPU.
4 Under Scheduling Anity, select physical processor anity for the virtual machine.
Use '-' for ranges and ',' to separate values.
For example, "0, 2, 4-7" would indicate processors 0, 2, 4, 5, 6 and 7.
5 Select the processors where you want the virtual machine to run and click OK.
22 VMware, Inc.
Potential Issues with CPU Affinity
Before you use CPU anity, you might need to consider certain issues.
Potential issues with CPU anity include:
For multiprocessor systems, ESXi systems perform automatic load balancing. Avoid manual
n
specication of virtual machine anity to improve the scheduler’s ability to balance load across
processors.
Anity can interfere with the ESXi host’s ability to meet the reservation and shares specied for a
n
virtual machine.
Because CPU admission control does not consider anity, a virtual machine with manual anity
n
seings might not always receive its full reservation.
Virtual machines that do not have manual anity seings are not adversely aected by virtual
machines with manual anity seings.
When you move a virtual machine from one host to another, anity might no longer apply because the
n
new host might have a dierent number of processors.
The NUMA scheduler might not be able to manage a virtual machine that is already assigned to certain
n
processors using anity.
Chapter 4 Administering CPU Resources
Anity can aect the host's ability to schedule virtual machines on multicore or hyperthreaded
n
processors to take full advantage of resources shared on such processors.
Host Power Management Policies
You can apply several power management features in ESXi that the host hardware provides to adjust the
balance between performance and power. You can control how ESXi uses these features by selecting a power
management policy.
Selecting a high-performance policy provides more absolute performance, but at lower eciency and
performance per wa. Low-power policies provide less absolute performance, but at higher eciency.
You can select a policy for the host that you manage by using the VMware Host Client. If you do not select a
policy, ESXi uses Balanced by default.
Table 4‑1. CPU Power Management Policies
Power Management Policy Description
High Performance Do not use any power management features.
Balanced (Default) Reduce energy consumption with minimal performance
Low Power Reduce energy consumption at the risk of lower
Custom User-dened power management policy. Advanced
compromise
performance
conguration becomes available.
When a CPU runs at lower frequency, it can also run at lower voltage, which saves power. This type of
power management is typically called Dynamic Voltage and Frequency Scaling (DVFS). ESXi aempts to
adjust CPU frequencies so that virtual machine performance is not aected.
When a CPU is idle, ESXi can apply deep halt states, also known as C-states. The deeper the C-state, the less
power the CPU uses, but it also takes longer for the CPU to start running again. When a CPU becomes idle,
ESXi applies an algorithm to predict the idle state duration and chooses an appropriate C-state to enter. In
power management policies that do not use deep C-states, ESXi uses only the shallowest halt state for idle
CPUs, C1.
VMware, Inc. 23
vSphere Resource Management
Select a CPU Power Management Policy
You set the CPU power management policy for a host using the vSphere Web Client.
Prerequisites
Verify that the BIOS seings on the host system allow the operating system to control power management
(for example, OS Controlled).
N Some systems have Processor Clocking Control (PCC) technology, which allows ESXi to manage
power on the host system even if the host BIOS seings do not specify OS Controlled mode. With this
technology, ESXi does not manage P-states directly. Instead, the host cooperates with the BIOS to determine
the processor clock rate. HP systems that support this technology have a BIOS seing called Cooperative
Power Management that is enabled by default.
If the host hardware does not allow the operating system to manage power, only the Not Supported policy
is available. (On some systems, only the High Performance policy is available.)
Procedure
1 Browse to the host in the vSphere Web Client navigator.
2 Click .
3 Under Hardware, select Power Management and click the Edit buon.
4 Select a power management policy for the host and click OK.
The policy selection is saved in the host conguration and can be used again at boot time. You can
change it at any time, and it does not require a server reboot.
Configure Custom Policy Parameters for Host Power Management
When you use the Custom policy for host power management, ESXi bases its power management policy on
the values of several advanced conguration parameters.
Prerequisites
Select Custom for the power management policy, as described in “Select a CPU Power Management Policy,”
on page 24.
Procedure
1 Browse to the host in the vSphere Web Client navigator.
2 Click .
3 Under System, select Advanced System .
4 In the right pane, you can edit the power management parameters that aect the Custom policy.
Power management parameters that aect the Custom policy have descriptions that begin with In
Custom policy. All other power parameters aect all power management policies.
5 Select the parameter and click the Edit buon.
N The default values of power management parameters match the Balanced policy.
Parameter Description
Power.UsePStates
Power.MaxCpuLoad
24 VMware, Inc.
Use ACPI P-states to save power when the processor is busy.
Use P-states to save power on a CPU only when the CPU is busy for less
than the given percentage of real time.
Parameter Description
Power.MinFreqPct
Power.UseStallCtr
Power.TimerHz
Power.UseCStates
Power.CStateMaxLatency
Power.CStateResidencyCoef
Power.CStatePredictionCoef
Power.PerfBias
Do not use any P-states slower than the given percentage of full CPU
speed.
Use a deeper P-state when the processor is frequently stalled waiting for
events such as cache misses.
Controls how many times per second ESXi reevaluates which P-state each
CPU should be in.
Use deep ACPI C-states (C2 or below) when the processor is idle.
Do not use C-states whose latency is greater than this value.
When a CPU becomes idle, choose the deepest C-state whose latency
multiplied by this value is less than the host's prediction of how long the
CPU will remain idle. Larger values make ESXi more conservative about
using deep C-states, while smaller values are more aggressive.
A parameter in the ESXi algorithm for predicting how long a CPU that
becomes idle will remain idle. Changing this value is not recommended.
Performance Energy Bias Hint (Intel-only). Sets an MSR on Intel processors
to an Intel-recommended value. Intel recommends 0 for high performance,
6 for balanced, and 15 for low power. Other values are undened.
6 Click OK.
Chapter 4 Administering CPU Resources
VMware, Inc. 25
vSphere Resource Management
26 VMware, Inc.
Memory Virtualization Basics 5
Before you manage memory resources, you should understand how they are being virtualized and used by
ESXi.
The VMkernel manages all physical RAM on the host. The VMkernel dedicates part of this managed
physical RAM for its own use. The rest is available for use by virtual machines.
The virtual and physical memory space is divided into blocks called pages. When physical memory is full,
the data for virtual pages that are not present in physical memory are stored on disk. Depending on
processor architecture, pages are typically 4 KB or 2 MB. See “Advanced Memory Aributes,” on page 116.
This chapter includes the following topics:
“Virtual Machine Memory,” on page 27
n
“Memory Overcommitment,” on page 28
n
“Memory Sharing,” on page 28
n
“Types of Memory Virtualization,” on page 29
n
Virtual Machine Memory
Each virtual machine consumes memory based on its congured size, plus additional overhead memory for
virtualization.
The congured size is the amount of memory that is presented to the guest operating system. This is
dierent from the amount of physical RAM that is allocated to the virtual machine. The laer depends on
the resource seings (shares, reservation, limit) and the level of memory pressure on the host.
For example, consider a virtual machine with a congured size of 1GB. When the guest operating system
boots, it detects that it is running on a dedicated machine with 1GB of physical memory. In some cases, the
virtual machine might be allocated the full 1GB. In other cases, it might receive a smaller allocation.
Regardless of the actual allocation, the guest operating system continues to behave as though it is running
on a dedicated machine with 1GB of physical memory.
Shares
Reservation
VMware, Inc. 27
Specify the relative priority for a virtual machine if more than the reservation
is available.
Is a guaranteed lower bound on the amount of physical RAM that the host
reserves for the virtual machine, even when memory is overcommied. Set
the reservation to a level that ensures the virtual machine has sucient
memory to run eciently, without excessive paging.
vSphere Resource Management
After a virtual machine consumes all of the memory within its reservation, it
is allowed to retain that amount of memory and this memory is not
reclaimed, even if the virtual machine becomes idle. Some guest operating
systems (for example, Linux) might not access all of the congured memory
immediately after booting. Until the virtual machines consumes all of the
memory within its reservation, VMkernel can allocate any unused portion of
its reservation to other virtual machines. However, after the guest’s workload
increases and the virtual machine consumes its full reservation, it is allowed
to keep this memory.
Limit
Is an upper bound on the amount of physical RAM that the host can allocate
to the virtual machine. The virtual machine’s memory allocation is also
implicitly limited by its congured size.
Memory Overcommitment
For each running virtual machine, the system reserves physical RAM for the virtual machine’s reservation (if
any) and for its virtualization overhead.
The total congured memory sizes of all virtual machines may exceed the amount of available physical
memory on the host. However, it doesn't necessarily mean memory is overcommied. Memory is
overcommied when the combined working memory footprint of all virtual machines exceed that of the
host memory sizes.
Because of the memory management techniques the ESXi host uses, your virtual machines can use more
virtual RAM than there is physical RAM available on the host. For example, you can have a host with 2GB
memory and run four virtual machines with 1GB memory each. In that case, the memory is overcommied.
For instance, if all four virtual machines are idle, the combined consumed memory may be well below 2GB.
However, if all 4GB virtual machines are actively consuming memory, then their memory footprint may
exceed 2GB and the ESXi host will become overcommied.
Overcommitment makes sense because, typically, some virtual machines are lightly loaded while others are
more heavily loaded, and relative activity levels vary over time.
To improve memory utilization, the ESXi host transfers memory from idle virtual machines to virtual
machines that need more memory. Use the Reservation or Shares parameter to preferentially allocate
memory to important virtual machines. This memory remains available to other virtual machines if it is not
in use. ESXi implements various mechanisms such as ballooning, memory sharing, memory compression
and swapping to provide reasonable performance even if the host is not heavily memory overcommied.
An ESXi host can run out of memory if virtual machines consume all reservable memory in a memory
overcommied environment. Although the powered on virtual machines are not aected, a new virtual
machine might fail to power on due to lack of memory.
N All virtual machine memory overhead is also considered reserved.
In addition, memory compression is enabled by default on ESXi hosts to improve virtual machine
performance when memory is overcommied as described in “Memory Compression,” on page 41.
Memory Sharing
Memory sharing is a proprietary ESXi technique that can help achieve greater memory density on a host.
Memory sharing relies on the observation that several virtual machines might be running instances of the
same guest operating system. These virtual machines might have the same applications or components
loaded, or contain common data. In such cases, a host uses a proprietary Transparent Page Sharing (TPS)
technique to eliminate redundant copies of memory pages. With memory sharing, a workload running on a
28 VMware, Inc.
virtual machine often consumes less memory than it might when running on physical machines. As a result,
virtual machine
1
guest virtual memory
guest physical memory
machine memory
a b
a
a b b c
b
c b
b c
virtual machine
2
higher levels of overcommitment can be supported eciently. The amount of memory saved by memory
sharing depends on whether the workload consists of nearly identical machines which might free up more
memory. A more diverse workload might result in a lower percentage of memory savings.
N Due to security concerns, inter-virtual machine transparent page sharing is disabled by default and
page sharing is being restricted to intra-virtual machine memory sharing. Page sharing does not occur
across virtual machines and only occurs inside a virtual machine. See “Sharing Memory Across Virtual
Machines,” on page 40 for more information.
Types of Memory Virtualization
There are two types of memory virtualization: Software-based and hardware-assisted memory
virtualization.
Because of the extra level of memory mapping introduced by virtualization, ESXi can eectively manage
memory across all virtual machines. Some of the physical memory of a virtual machine might be mapped to
shared pages or to pages that are unmapped, or swapped out.
A host performs virtual memory management without the knowledge of the guest operating system and
without interfering with the guest operating system’s own memory management subsystem.
The VMM for each virtual machine maintains a mapping from the guest operating system's physical
memory pages to the physical memory pages on the underlying machine. (VMware refers to the underlying
host physical pages as “machine” pages and the guest operating system’s physical pages as “physical”
pages.)
Chapter 5 Memory Virtualization Basics
Each virtual machine sees a contiguous, zero-based, addressable physical memory space. The underlying
machine memory on the server used by each virtual machine is not necessarily contiguous.
For both software-based and hardware-assisted memory virtualization, the guest virtual to guest physical
addresses are managed by the guest operating system. The hypervisor is only responsible for translating the
guest physical addresses to machine addresses. Software-based memory virtualization combines the guest's
virtual to machine addresses in software and saves them in the shadow page tables managed by the
hypervisor. Hardware-assisted memory virtualization utilizes the hardware facility to generate the
combined mappings with the guest's page tables and the nested page tables maintained by the hypervisor.
The diagram illustrates the ESXi implementation of memory virtualization.
Figure 5‑1. ESXi Memory Mapping
The boxes represent pages, and the arrows show the dierent memory mappings.
n
The arrows from guest virtual memory to guest physical memory show the mapping maintained by the
n
page tables in the guest operating system. (The mapping from virtual memory to linear memory for
x86-architecture processors is not shown.)
VMware, Inc. 29
The arrows from guest physical memory to machine memory show the mapping maintained by the
n
VMM.
vSphere Resource Management
The dashed arrows show the mapping from guest virtual memory to machine memory in the shadow
n
page tables also maintained by the VMM. The underlying processor running the virtual machine uses
the shadow page table mappings.
Software-Based Memory Virtualization
ESXi virtualizes guest physical memory by adding an extra level of address translation.
The VMM maintains the combined virtual-to-machine page mappings in the shadow page tables. The
n
shadow page tables are kept up to date with the guest operating system's virtual-to-physical mappings
and physical-to-machine mappings maintained by the VMM.
The VMM intercepts virtual machine instructions that manipulate guest operating system memory
n
management structures so that the actual memory management unit (MMU) on the processor is not
updated directly by the virtual machine.
The shadow page tables are used directly by the processor's paging hardware.
n
There is non-trivial computation overhead for maintaining the coherency of the shadow page tables.
n
The overhead is more pronounced when the number of virtual CPUs increases.
This approach to address translation allows normal memory accesses in the virtual machine to execute
without adding address translation overhead, after the shadow page tables are set up. Because the
translation look-aside buer (TLB) on the processor caches direct virtual-to-machine mappings read from
the shadow page tables, no additional overhead is added by the VMM to access the memory. Note that
software MMU has a higher overhead memory requirement than hardware MMU. Hence, in order to
support software MMU, the maximum overhead supported for virtual machines in the VMkernel needs to
be increased. In some cases, software memory virtualization may have some performance benet over
hardware-assisted approach if the workload induces a huge amount of TLB misses.
Performance Considerations
The use of two sets of page tables has these performance implications.
No overhead is incurred for regular guest memory accesses.
n
Additional time is required to map memory within a virtual machine, which happens when:
n
The virtual machine operating system is seing up or updating virtual address to physical address
n
mappings.
The virtual machine operating system is switching from one address space to another (context
n
switch).
Like CPU virtualization, memory virtualization overhead depends on workload.
n
Hardware-Assisted Memory Virtualization
Some CPUs, such as AMD SVM-V and the Intel Xeon 5500 series, provide hardware support for memory
virtualization by using two layers of page tables.
The rst layer of page tables stores guest virtual-to-physical translations, while the second layer of page
tables stores guest physical-to-machine translation. The TLB (translation look-aside buer) is a cache of
translations maintained by the processor's memory management unit (MMU) hardware. A TLB miss is a
miss in this cache and the hardware needs to go to memory (possibly many times) to nd the required
translation. For a TLB miss to a certain guest virtual address, the hardware looks at both page tables to
translate guest virtual address to machine address. The rst layer of page tables is maintained by the guest
operating system. The VMM only maintains the second layer of page tables.
30 VMware, Inc.
Loading...