best practices for vmware horizon view vdi based on huawei ... pr… · best practices for vmware...

Best Practices for VMware Horizon View VDI Based on HUAWEI OceanStor 5500 V3 Converged Storage System

Yuan Chuanhu

Storage Solution Dept, IT, EBG

2015-05-11 V1.1

This document is aimed at the scenario where HUAWEI OceanStor V3 converged storage systems are used to serve

VMware Horizon View 6.1. This document focuses on how to efficiently deploy VMware Horizon View application

based on HUAWEI OceanStor V3 converged storage systems (the OceanStor V3 for short), and verifies VMware

Horizon View application in typical enterprise user scenarios. The best practices described in this document help you

obtain higher deployment efficiency and better service quality, thereby ensuring VMware VDI's performance and

availability.

Huawei Technologies Co., Ltd.

mailto:[email protected]

Huawei Proprietary and Confidential Copyright © Huawei Technologies Co., Ltd. 2

Best Practices for VMware Horizon View VDI Based on HUAWEI OceanStor 5500 V3

Converged Storage System

Contents

1 About This Document .................................................................................................................. 4

1.1 Overview ...................................................................................................................................................................... 4

1.2 About the Best Practices ............................................................................................................................................... 4

1.3 Intended Audience ........................................................................................................................................................ 5

1.4 Workload Models .......................................................................................................................................................... 5

2 Products and Technologies ......................................................................................................... 6

2.1 OceanStor V3 Converged Storage Systems .................................................................................................................. 6

2.1.2 OceanStor OS ............................................................................................................................................................ 7

2.1.3 Next-Generation Hardware ........................................................................................................................................ 7

2.1.4 Convergence Design .................................................................................................................................................. 7

2.1.5 Virtualization, Intelligence, and Efficiency................................................................................................................ 8

2.2 VMware Horizon View ................................................................................................................................................. 8

2.2.1 Overview ................................................................................................................................................................... 8

2.2.2 Linked Clone ............................................................................................................................................................. 9

2.2.3 User File Management ............................................................................................................................................... 9

2.2.4 View Storage Accelerator ........................................................................................................................................ 10

2.2.5 SE Sparse Virtual Disk............................................................................................................................................. 10

3 Best Practices for VDI Planning ............................................................................................... 11

3.1 Capacity Planning ....................................................................................................................................................... 11

3.1.1 Capacity Planning for System Disks ........................................................................................................................ 11

3.1.2 Capacity Planning for Data Disks ............................................................................................................................ 12

3.1.3 Exemplification ........................................................................................................................................................ 12

3.2 Performance Planning ................................................................................................................................................. 12

3.2.1 Disk Performance .................................................................................................................................................... 13

3.2.2 RAID Performance .................................................................................................................................................. 13

3.2.3 Exemplification ........................................................................................................................................................ 14

4 Best Practices for VDI Storage Configuration ....................................................................... 15

4.1 Overview .................................................................................................................................................................... 15

4.2 Storage Configuration ................................................................................................................................................. 15

4.2.1 Creating Disk Domains ............................................................................................................................................ 15

4.2.2 Creating Storage Pools............................................................................................................................................. 16

4.2.3 Creating LUNs ......................................................................................................................................................... 16




4.2.4 Configuring Hosts and Host Groups ........................................................................................................................ 17

4.2.5 Configuring Port Groups ......................................................................................................................................... 17

4.2.6 Creating Mapping Views ......................................................................................................................................... 17

4.2.7 Configuring VAAI ................................................................................................................................................... 18

4.3 Best Practices on Applications .................................................................................................................................... 18

4.3.1 Configuring Multipathing ........................................................................................................................................ 18

4.3.2 Creating Clusters ..................................................................................................................................................... 19

4.3.3 Creating VM Templates ........................................................................................................................................... 19

4.3.4 Configuring View Storage Accelerator and Space Reclamation .............................................................................. 19

5 Example of VDI Planning and Configurations ..................................................................... 22

5.1 Introduction to the Verification Test ........................................................................................................................... 22

5.1.1 Network Diagram .................................................................................................................................................... 22

5.1.2 Hardware and Software Configurations ................................................................................................................... 23

5.2 Test Results ................................................................................................................................................................. 23

5.2.1 Medium Workload Test in a Scenario with 500 VDI Users ..................................................................................... 23

5.2.2 Medium Workload Test in a Scenario with 1000 VDI Users ................................................................................... 24




1 About This Document

1.1 Overview

As virtual desktop technologies mature, an increasing number of enterprises are deploying

virtual desktop infrastructure (VDI) to their IT architecture and transforming their office

environment from traditional physical PCs to virtual VDI cloud desktops. Performance and

storage capacity are major factors to be considered in planning and deploying VDI virtual

desktop environments. With virtualization, hybrid, thin IT, and low carbon footprint,

HUAWEI OceanStor V3 converged storage systems (the OceanStor V3 for short) are the best

storage platform for the new-generation data center. It is also the optimum storage platform

for constructing secure, efficient, and scalable VDI.

Based on VMware Horizon View and the OceanStor V3, the best practices aim to help

enterprises construct the best infrastructure for virtual desktops by fully utilizing advantages

brought by the OceanStor 5500 V3. This document contains the following parts:

OceanStor 5500 V3

VMware Horizon View 6.1

Best practices for VDI storage planning

Best practices for VDI storage configuration

1.2 About the Best Practices

The best practices involve the basic performance of the OceanStor 5500 V3. The suggestions

and best practices proposed in this document are based on the following conditions:

Fibre Channel storage protocol is used for its high performance.

2.5-inch 10k rpm SAS disks are used as storage media.

VDI users are office users. (For details about user types, see section 1.4 "Workload

Models.")

VDI desktops are linked clone desktops. The capacity for storing the desktops is

negligible.

All the suggestions in this document apply to only one scenario: VMware Horizon View is

deployed on the OceanStor 5500 V3. In addition to this document, Huawei and its partners

provide professional services to help customers complete the VDI design covered in this document. These services enable customers to build the best storage architecture and solution

for their VDI environments.




1.3 Intended Audience

The target audiences include personnel responsible for the construction, design, management,

and support of VMware VDI solutions. It is assumed that the readers are familiar with the

following products and technologies:

VMware vSphere and VMware Horizon View

OceanStor V3

1.4 Workload Models

Login Virtual Session Indexer (VSI) is a standard test tool used to test the performance and

scalability of VMware Horizon View, Citrix XenDesktop and XenApp, Microsoft

Remote Desktop Services (Terminal Services), or other Windows-based virtual desktop

solutions in virtual desktop environments.

The best practices discuss the workloads of office desktops only.

Table 1-1 distinguishes between different workloads and associate user types with IOPS

values, based on empirical values collected in Huawei lab tests.

Table 1-1 Mapping between user type and IOPS

User Type Description IOPS of Each User

Task

(light workload)

Such type of users is typically engaged in

data recording and paperwork, using

Outlook, Excel, Word, and Web browsers

(Internet Explorer or Firefox) to complete

routine tasks.

3 to 7

Knowledge

(heavy workload)

Such type of users has comprehensive

knowledge. In addition to all tools used by

task-type users, knowledge-type users need

to process large-scale PowerPoint

documents and perform other large file

operations. Such users include business

managers, management officers, and

marketing engineers.

8 to 16

Workloads of different user types vary depending on the IOPS, read/write I/Os, and I/O size. The best

practices use the medium workload type of Login VSI to simulate the workloads of knowledge-type

users.




2 Products and Technologies

2.1 OceanStor V3 Converged Storage Systems

HUAWEI OceanStor V3 converged storage systems are next-generation converged storage

products designed for enterprise-class applications. Leveraging a storage operating

system with a cloud-oriented architecture, a powerful next-generation hardware platform, and

a full range of intelligent management software, OceanStor V3 converged storage systems

deliver industry-leading functionality, performance, efficiency, reliability, and ease-of-use.

They provide data storage for applications such as large-scale database OLTP/OLAP, file

sharing, and cloud computing, and can be used in industries ranging from government,

finance, telecommunications, energy, to media and entertainment (M&E). Meanwhile,

OceanStor V3 converged storage systems can provide a wide range of efficient and flexible

backup and disaster recovery solutions to ensure business continuity and data security,

delivering excellent storage services.

For details about the OceanStor V3, click the following link:

http://e.huawei.com/en/products/cloud-computing-dc/storage/unified-storage/mid-range

Figure 2-1 OceanStor V3 converged storage systems

http://e.huawei.com/en/products/cloud-computing-dc/storage/unified-storage/mid-range




2.1.2 OceanStor OS

OceanStor OS cloud-oriented architecture is the core of OceanStor V3 converged storage

systems.

Figure 2-2 Storage operating system employing a cloud-oriented architecture

Convergence of SAN and

NAS storage

Primary storage

convergence

Backup convergence

Third-party convergence

Convergence of SSD and

HDD storage

Convergence of entry-level,

mid-range, and high-end

storage systems

Convergence of primary and

backup storage

Convergence of

heterogeneous storage

systems

2.1.3 Next-Generation Hardware

The OceanStor V3 employs next-generation Intel multi-core processors, PCIe 3.0 buses, 12

Gbit/s SAS 3.0 disk ports, and a variety of host ports such as 16 Gbit/s Fibre Channel, 10

Gbit/s FCoE, and 56 Gbit/s InfiniBand. The storage system provides up to 28 GB/s of system

bandwidth to meet the requirements of bandwidth-intensive application scenarios. It also

offers million-level IOPS performance, outshining products from other vendors.

The OceanStor V3 is equipped with exclusive SmartIO cards. A SmartIO card supports 8

Gbit/s Fibre Channel, 16 Gbit/s Fibre Channel, 10 Gbit/s iSCSI, and 10 Gbit/s FCoE. Users

can specify the protocols that a SmartIO card is required to support.

The deduplication/compression cards used by the OceanStor V3 support lossless data

deduplication and compression, efficiently reducing data storage costs. In addition, the

storage system can leverage data encryption to secure data.

2.1.4 Convergence Design

Convergence of SAN and NAS storage: SAN and NAS services are converged to

provide elastic storage, simplify service deployment, improve storage resource

utilization, and reduce total cost of ownership (TCO). Underlying storage resource pools

directly provide both block and file services, thereby shortening storage resource access

paths to ensure that the two services are equally efficient.

Convergence of heterogeneous storage systems: Thanks to the built-in heterogeneous

virtualization function, the OceanStor V3 can efficiently manage storage systems from

other mainstream vendors and consolidate these storage systems into unified resource

pools for central and flexible resource allocation.




Convergence of entry-level, mid-range, and high-end storage systems: The

OceanStor V3 is the only storage system in the industry that enables entry-level,

mid-range, and high-end storage systems to interwork seamlessly with one another. Data

can freely flow among storage products of different models without the assistance of

third-party systems.

Convergence of SSD and HDD storage: The advantages of traditional and solid-state

storage media are combined, bringing the performance of different types of storage

media into full play and striking an optimal balance between performance and cost.

Convergence of primary and backup storage: The built-in backup function enables

data to be efficiently backed up without additional backup software, simplifying backup

solution management.

2.1.5 Virtualization, Intelligence, and Efficiency

RAID 2.0+ underlying virtualization: RAID 2.0+ used by the OceanStor V3 employs

two-layer virtualized management, namely, underlying disk management plus

upper-layer resource management. In an OceanStor V3 converged storage system, the

space of each disk is divided into data blocks with a small granularity and RAID groups

are created based on data blocks so that data is evenly distributed onto all disks in a

storage pool. Besides, using data blocks as the smallest units greatly improves the

efficiency of resource management.

SmartTier (intelligent storage tiering): SmartTier automatically analyzes data access

frequencies per unit time and migrates data to disks of different performance levels

based on the analysis result. (High-performance disks store most frequently accessed

data, performance disks store less frequently accessed data, and large-capacity disks

store seldom accessed data.) In this way, the optimal overall performance is achieved,

and the IOPS cost is reduced.

SmartQoS (intelligent service quality control): SmartQoS categorizes service data based

on data characteristics (each category represents a type of application) and sets a priority

and performance objective for each category. In this way, resources are allocated to

services properly, fully utilizing system resources.

SmartThin (thin provisioning): SmartThin allocates storage space on demand rather than

pre-allocating all storage space at the initial stage. It is more cost-effective because

customers can start business with a few disks and add disks based on site requirements.

In this way, the initial purchase cost and TCO are reduced.

SmartCache (Intelligent storage cache): SmartCache enables storage systems to use

SSDs as cache resources to improve system read performance in scenarios where read

operations are more than write operations and hotspot data exists.

2.2 VMware Horizon View

2.2.1 Overview

VMware Horizon View, formerly known as VMware View, is a VDI solution that simplifies

desktop management and provides on-demand services. It expands VMware server

deployments to bring VDI advantages to virtual desktops.




Figure 2-3 Architecture of VMware Horizon View

Based on the vSphere 5.5 virtualization architecture, Horizon View 6.1 offers the following:

Desktop management

End user experience

Flexible access

Cloud environment compatibility

The best practices introduce configurations based on VMware Horizon View 6.1.

This chapter covers storage cost– or performance-related technologies used in the best

practices.

2.2.2 Linked Clone

The linked clone function, provided by Horizon View Composer, can create multiple desktop

images based on a golden master image. The latest View Composer version has removed the

restriction of allowing a maximum of only eight hosts in a cluster. It supports large host pools,

accelerates deployment and reconstruction, and ensures the instant and uniform update of any

number of virtual desktops. When used in conjunction with ThinApp, Horizon View

Composer reduces the total amount of images, required storage capacity, and operating costs.

The best practices use the latest version of Horizon View Composer to link clone desktops,

improving deployment and maintenance efficiency and reducing the required number of

disks.

2.2.3 User File Management

Horizon View Persona Management provides consistent personalized experience for end users

and offers cost-effective stateless desktops. Such desktops bring down IT costs while

providing end users with "familiar" desktop appearances (last desktop state retained) and

faster login.




2.2.4 View Storage Accelerator

vSphere 5.0 and later versions allow VM disk data to be cached to ESXi hosts. This function

is called View Storage Accelerator (VSA), which leverages content-based read cache (CBRC)

of ESXi hosts. With VSA, hosts no longer need to read the entire storage operating system.

Instead, it reads common data blocks from the CBRC to improve access performance.

The best practices use VSA to ease the pressure of partial storms in VDI.

2.2.5 SE Sparse Virtual Disk

Linked clone desktops can use SE sparse virtual disks as system disks to reclaim storage

space, further reducing the storage space required by permanent desktops.

The best practices use SE Sparse virtual disks to greatly save storage space and mitigate the

risk of system disk capacity shortage.




3 Best Practices for VDI Planning

3.1 Capacity Planning

This section describes user data capacity planning. NL-SAS disks are the recommended

media for storing user data. The following information must be considered for capacity

planning:

Actual available capacity of each disk

Capacity overhead of RAID groups during storage pool creation

The necessity of thin LUN provisioning to address future data growth

3.1.1 Capacity Planning for System Disks

In the best practices, system disks of VDI office desktops use linked clone technology of

VMware Horizon View. LUN capacity required by system disks can be calculated using the

following formula:

Capacity = System Disk Capacity + Number of VMs * (Swap file size + Headspace of

every VM)

Capacity indicates the capacity of a LUN.

Number of VMs indicates the number of VMs on a LUN.

Swap file size indicates the memory size of a VDI virtual desktop.

Headspace of every VM includes VM files and differential data between each VM and template VM.

Table 3-1 Disk utilization of different RAID levels

RAID Level Space Utilization

RAID 5 (4D+1P) 80%

RAID 6 (4D+2P) 67%

RAID 10 50%

Different RAID levels have different percentages of storage space consumption to maximum

allocated space in a storage pool. Therefore, based on the information in Table 3-1, the required disk storage space can be estimated using the following formula:

app:ds:capacity

app:ds:capacity

app:ds:%20%20virtual%20machine

app:ds:headspace




Total capacity = Number of LUNs * Capacity/Space utilization/0.8

The percentage of storage space consumption to maximum allocated space in a storage pool is the alarm

threshold of space utilization in the storage pool. By default, the threshold is 80%. If used capacity of a

storage pool in the OceanStor 5500 V3 exceeds the threshold, an alarm will be reported.

The required disk quantity can be calculated using the following formula:

Number of disks = Total capacity/Disk capacity

3.1.2 Capacity Planning for Data Disks

The required data disk capacity can be calculated using the following formula:

Total capacity = Number of VMs * Data disk capacity/Space utilization/0.8

The required disk quantity can be calculated using the following formula:

Number of disks = Total capacity/Disk capacity

For details about the preceding parameters, see "Capacity Planning for System Disks."

3.1.3 Exemplification

1000 VDI office desktops are deployed, each of which is configured with 20 GB of system

disks, 20 GB of data disks, one vCPU, and 1 GB RAM. The storage capacity of these

desktops can be planned as follows:

RAID 10 is used as a RAID policy in storage pools and 10 LUNs are created to store VM

system disks. The capacity of a LUN can be calculated using the following formula:

20 GB + 100 * (1 GB + 2 GB) = 320 GB

The total capacity for storing system disks can be calculated using the following formula:

10 * 320 GB/50%/0.8 = 8000 GB

If 600 GB SAS disks are used and each disk has 550 GB of available capacity, the number of

system disks required can be calculated using the following formula:

8000 GB/550 GB = 15

RAID 5 is adopted as a RAID policy in storage pool to store VM data disks. The total

capacity required can be calculated using the following formula:

1000 * 20 GB/0.8/0.8 = 31,250 GB

If 600 GB SAS disks are used and each disk has 550 GB of available capacity, the number of

data disks required can be calculated using the following formula:

31,250 GB/550 GB = 57

3.2 Performance Planning

Performance is a key factor in the best practices. Workloads of VDI in a steady state are

mostly random small I/Os, and the storage performance is mainly determined by IOPS.

Therefore, this section chooses a RAID type based on an understanding of estimated disk

app:ds:total%20capacity

app:ds:capacity

app:ds:utility%20rate%20of%20space



app:ds:capacity

app:ds:utility%20rate%20of%20space





performance, RAID performance overhead, and workload features of VDI in a steady state.

Finally, the section calculates the number of disks enough to meet the performance

requirements of storage pools with different RAID levels.

3.2.1 Disk Performance

Table 3-2 Estimated IOPS of s single disk

Disk Type Estimated IOPS

10k rpm SAS 150

SSD (eMLC) 2500

SSD (SLC) 3500

The preceding table shows the estimated IOPS values of different disks. The IOPS value

indicates the performance of a disk assuming that it meets the latency requirement.

3.2.2 RAID Performance

After the type of needed disks is determined, it is necessary to understand the IOPS

consumption of different RAID groups. The IOPS consumption of a RAID group is caused by

its write operations, as RAID working principles suggest. The following table describes

the write overhead of RAID groups.

Table 3-3 Write consumption of RAID groups

RAID Level RAID Penalty

RAID 10 2

RAID 5 4

RAID 6 6

Based on a Login VSI benchmark test, the following tables describe the percentages of

I/O writes and reads of VDI in a steady state, as well as the percentages of I/O writes and

reads delivered to disks.

Table 3-4 Percentages of I/O writes and reads in a steady state

Read I/O Percentage (%) Write I/O Percentage (%)

20 80




Table 3-5 Percentages of I/O writes and reads delivered to disks in a steady state

Percentage of Read I/Os Delivered Percentage of Write I/Os Delivered

60 30

The required IOPS can be easily calculated using the following formulas based on the

information provided in Table 3-2, Table 3-3, and Table 3-4.

IOPS = (Target IOPS * Read I/O% * Disk read I/O%) + (Target IOPS * Write I/O%

* Disk write I/O%) * RAID penalty

-Formula 1

Based on information in Table 3-1, the required number of disks can be estimated using the

following formula:

Number of disks = IOPS/Estimated IOPS

-Formula 2

Target IOPS is obtained by the average IOPS (shown in Table 1-1) of each user multiplied by the

number of users.

The values of read I/O%, write I/O%, RAID penalty, disk read I/O%, and disk write I/O% are provided

in Table 3-2, Table 3-3, and Table 3-4.

The value of estimated IOPS is provided in Table 3-1.

The calculated IOPS indicates IOPS provided by back-end disks after the performance of read

hit and write cache is improved and enters a steady state.

3.2.3 Exemplification

1000 knowledge-type users (10 IOPS per user on average generated by Login VSI

benchmark workloads) use RAID 10. The required performance can be calculated using

formula 1.

IOPS = (1000 * 10 * 20% * 60%) + (1000 * 10 * 90% * 30%) * 2 = 400 + 4800 = 6000

The number of required disks for the 1000 users can be calculated by using the calculated

IOPS value and formula 2.

Number of 10k rpm SAS disks = 6000/150 = 40

Based on the exemplification, Huawei recommends that customers use high-reliability and

high-performance RAID 10 as the RAID level of storage pools on the OceanStor 5500

V3 when deploying a large number of VMware View linked clone desktops, if the

preconditions described in this document are met.

In the best practices, system disks of VDI virtual desktops occupy small storage space because they use

the linked clone function. For this reason, system disk capacity will not become a bottleneck.




4 Best Practices for VDI Storage Configuration

4.1 Overview

This chapter describes recommended configurations for VMware View VDI on the OceanStor

V3. It includes storage planning, storage configuration, cluster configuration, VM template

configuration.

4.2 Storage Configuration

Storage is the basis for VMware Horizon View implementation. This section describes the

best practices of deploying a desktop pool based on Horizon View linked clone for

knowledge-type users. The OceanStor 5500 V3 uses 2.5-inch 10k rpm SAS disks and the

Fibre Channel protocol.

The configuration of storage arrays mainly involves the allocation of storage resources to

hosts. Log in to OceanStor DeviceManager. For details, refer to the OceanStor 5500 V3

manual. Then complete the following configuration:

Figure 4-1 Resource allocation flowchart

4.2.1 Creating Disk Domains

In this white paper, 2.5-inch 10k rpm SAS disks are used to provide storage space for virtual

desktops. As user data storage requires modest performance and large capacity, NL-SAS disks

are recommended.




For details about determining the number of disks for linked clone desktop pools provided by

VMware Horizon View, see section 3.1 "Capacity Planning" and section 3.2 "Performance

Planning." Based on VMware View, it is recommended that:

Different disk domains be used to provide storage space for system disks and data disks

to facilitate management and maintenance.

Disk enclosure amounts be averaged to storage engines for load balancing.

No more than 100 disks be allocated for each tier of a disk domain.

4.2.2 Creating Storage Pools

A storage pool, a container that stores storage space resources, is created in a disk domain. A

storage pool can dynamically allocate resources from a disk domain and define the RAID

level of each storage tier.

A storage tier is a collection of storage media providing the same performance level in a

storage pool. Different storage tiers manage storage media with different performance levels

and provide different storage spaces for applications whose performance requirements vary.

There are three types of storage tiers: high-performance tier that consists of SSDs (including

SLC and eMLC), performance tier that consists of SAS disks (including 15k rpm and 10k

rpm), and capacity tier that consists of NL-SAS disks.

The OceanStor V3 supports six RAID levels: RAID 6, RAID 10, RAID 5, RAID 3, RAID 50,

and RAID 1. The most commonly used RAID levels are RAID 6, RAID 10, and RAID 5.

From the perspective of stripe width, RAID 6 and RAID 5 are classified into RAID 6-4

(2D+2P), RAID 6-6 (4D+2P), RAID 6-10 (8D+2P), RAID 5-3 (2D+1P), RAID 5-5 (4D+1P),

and RAID 5-9 (8D+1P).

In VDI scenarios, the best practices for creating storage pools for storing system disks of

virtual desktops are as follows:

RAID 10 is strongly recommended because this RAID level can provide high

performance using minimum disks.

If reliability takes precedence over performance, RAID 6 is recommended.

If capacity takes precedence over reliability, RAID 5 is recommended.

In VDI scenarios, the best practices for creating storage pools for storing data disks of virtual

desktops are as follows:

RAID 5 is strongly recommended because this RAID level can provide largest available

capacity using minimum disks.

If reliability takes precedence over performance, RAID 6 is recommended.

If performance takes precedence over reliability, RAID 10 is recommended.

During the creation of a storage pool, you can set an alarm threshold for the capacity

allocation ratio. The default threshold is 80%. Capacity alarming is particularly important in

scenarios where value-added features such as thin LUN, snapshot, remote replication, and

clone are used. You can set a proper alarm threshold based on the speed of application data

growth to prevent insufficient capacity of the storage pool from casing application

interruption.

4.2.3 Creating LUNs

A LUN is a storage unit that can be directly mapped to a host for data reads and writes.




In VDI scenarios, the best practices for creating LUNs for system disks of virtual desktops are

as follows:

Create an even number of LUNs so that they can be evenly distributed to controllers for

load balancing.

When creating LUNs, do not enable SmartThin for performance consideration.

Assign not more than 128 VMs for each LUN.

Configure the advanced properties for the LUNs and set the prefetch policy to No

Prefetch.

Do not enable SmartThin for the LUNs used by system disks because the system disks occupy only a

small amount of space but have demanding performance requirements. It is recommended that

SmartThin be enabled for user data disks to improve storage flexibility.

4.2.4 Configuring Hosts and Host Groups

You can create and manage hosts so that the hosts can obtain and use the storage resources

allocated by the storage system. You can also create host groups for easily managing multiple

hosts. This management mode adapts to the VMware cluster management mode.

In VDI scenarios, the best practices for creating hosts and host groups are as follows:

When creating a host, use the automatic scan function to scan for hosts.

After creating a host, modify its initiator and disable ALUA.

When creating a host group, match it with a VMware cluster and name the host group

after the cluster name for easy management and maintenance.

To use the automatic scan function, you need to install UltraPath on the ESX host.

4.2.5 Configuring Port Groups

If a storage device is exclusively used by the VDI, it is recommended not to configure port

groups. If a storage device is shared by the VDI and other services, it is recommended that

port groups be configured to isolate these services from affecting each other's performance.

In VDI scenarios, the best practices for configuring port groups are as follows:

Add all ports that are connected to VDI hosts to a host group for simplified deployment

and management.

8 Gbit/s Fibre Channel ports are preferred for front-end connection. It is recommended

that at least two front-end ports be configured on each controller for redundancy.

4.2.6 Creating Mapping Views

A mapping view enables you to flexibly allocate storage resources to hosts and complete

configuration on the storage array. You can determine whether to use port groups in the

mapping view based on site requirements.




Figure 4-2 Creating a mapping view

4.2.7 Configuring VAAI

As a vSphere API of VMware, VAAI offloads storage operations to supported storage arrays

for high performance and efficiency. The OceanStor 5500 V3 in the best practices supports

VAAI to enable vSphere to quickly execute key tasks and reduce usage of CPU, memory, and

storage bandwidth for higher performance and lower costs.

The efficiency can be improved in the following scenarios: VM cloning, VM migration, and

VM power-on and power-off.

By default, VAAI is enabled on the OceanStor 5500 V3.

4.3 Best Practices on Applications

This chapter describes the best practices for configuring VMware Horizon View based on the

planning in this document and Huawei's experience.

The configuration involves multipathing, VMware clusters, data stores, and VM templates.

4.3.1 Configuring Multipathing

In this white paper, Fibre Channel SAN is used to provide storage resources. The redundant

Fibre Channel switches provide multiple paths between the host and storage system.

Therefore, the multipathing software must be used to choose the optimum path.

It is recommended that UltraPath for VMware that can automatically discover and select the

best path be installed.




4.3.2 Creating Clusters

Note the following when configuring a cluster:

If HA is enabled, ensure that the cluster has sufficient VM resources and provide at least

one more host for switchover.

If DRS is enabled, provide a dedicated channel for vMotion.

The latest Horizon View Composer supports up to 32 clustered hosts, whereas earlier

versions support 8.

4.3.3 Creating VM Templates

Because the best practices assume that the desktop users are knowledge-type users, ensure the

following when creating and configuring a VM template:

Set VM hardware version to 10. You can use SE Sparse virtual disks to improve disk

utilization.

If memory resources are sufficient, reserve at least 1 GB memory for each VM to

prevent a waste of storage space.

4.3.4 Configuring View Storage Accelerator and Space Reclamation

On the View management platform, configure the host cache size for VMware vCenter Server.

Log in to VMware View Administrator, and choose View Configuration > Server > vCenter

Server. Set the host cache size to 2048 MB, as shown in Figure 4-3.




Figure 4-3 Configuring the host cache size and space reclamation for vCenter Server

When adding a linked clone desktop pool, select Use View Storage Accelerator and use it for

operating system disks, as shown in Figure 4-4:




Figure 4-4 Enabling host caching

As shown in Figure 4-4, you can select a time range for host caching to avoid host resource contention

during peak usage. Select the time range based on service requirements.




5 Example of VDI Planning and Configurations

5.1 Introduction to the Verification Test

5.1.1 Network Diagram

Figure 5-1 Network diagram for testing

ES

T

ES

T

......

RH2288R720R720

5500T 5500 V3

Management network

Fibre Channel network

Cluster A

(test management)

Cluster B

(service cluster)

vCenter

server

Connection

serverAD/DHCP/DNS Login VSI VM VM...

Two switches are used in the network.




Three R720 servers are used in cluster A (test cluster) to carry workloads of testing machines and

management nodes.

Eight RH2288 servers are used in cluster B (service cluster) to carry workloads of service VMs.

5.1.2 Hardware and Software Configurations

Hardware Configurations

Device Type Quantity Function Configuration

HUAWEI

Tecal RH2288

V2

8

Provides a

computing cluster

to carry virtual

desktops and to

connect VMs.

2 x 8 Gbit/s Fibre Channel ports

2 x Intel(R) Xeon(R) CPU E5-2660

@ 2.20 GHz

256 GB RAM

GE network adapter

600 GB local SAS disk

HUAWEI

5500 V3

1

(2

controllers)

Provides

high-reliability and

high-performance

shared storage

resources for

virtual desktops.

1 x Intel(R) Xeon(R) CPU E5-2620

v2 @ 2.10 GHz

Storage array cache size: 24 GB

(per controller)

50 x 2.5-inch 10k rpm SAS disks

Software Configurations

Component Software Version

Virtualization

infrastructure VMware vSphere (VMware ESXi) 5.5.0

Virtualization

management software VMware vCenter 5.5.0

Virtualization desktop

software VMware Horizon View 6.0.1 build-2088845

Workload test tool Login VSI 3.7

5.2 Test Results

5.2.1 Medium Workload Test in a Scenario with 500 VDI Users

Figure 5-2 shows the test results after the medium workload type of Login VSI is used to

simulate the workloads of 500 VDI users.




Figure 5-2 Test results

This test uses twenty 2.5-inch 10k rpm SAS disks to carry VM system disks (not VM data disks).

The RAID level in the storage pool is RAID 10.

Five 300 GB LUNs are created, each of which carries 100 VMs.

5.2.2 Medium Workload Test in a Scenario with 1000 VDI Users

Figure 5-3 shows the test results after the medium workload type of Login VSI is used to

simulate the workloads of 1000 VDI users.

Figure 5-3 Test results

This test uses forty 2.5-inch 10k rpm SAS disks to carry VM system disks (not VM data disks).

The RAID level in the storage pool is RAID 10.

Ten 300 GB LUNs are created, each of which carries 100 VMs.

Copyright © Huawei Technologies Co., Ltd. 2015. All rights reserved.

No part of this document may be reproduced or transmitted in any form or by any means without prior written

consent of Huawei Technologies Co., Ltd.

Trademarks and Permissions

and other Huawei trademarks are trademarks of Huawei Technologies Co., Ltd.

All other trademarks and trade names mentioned in this document are the property of their respective holders.

Notice

The purchased products, services and features are stipulated by the contract made between Huawei and the

customer. All or part of the products, services and features described in this document may not be within the

purchase scope or the usage scope. Unless otherwise specified in the contract, all statements, information,

and recommendations in this document are provided "AS IS" without warranties, guarantees or

representations of any kind, either express or implied.

The information in this document is subject to change without notice. Every effort has been made in the

preparation of this document to ensure accuracy of the contents, but all statements, information, and

recommendations in this document do not constitute a warranty of any kind, express or implied.

Huawei Technologies Co., Ltd.

Address: Huawei Industrial Base

Bantian, Longgang

Shenzhen 518129

People's Republic of China

Website: http://www.huawei.com

Email: [email protected]

http://www.huawei.com/

mailto:[email protected]

best practices for vmware horizon view vdi based on huawei ... pr… · best practices for vmware...

Documents