Flashstack With Cisco Ucs and Pure Storage Flasharray//M For 5000 Vmware Horizon View 6.2 Users
Flashstack With Cisco Ucs and Pure Storage Flasharray//M For 5000 Vmware Horizon View 6.2 Users
Flashstack With Cisco Ucs and Pure Storage Flasharray//M For 5000 Vmware Horizon View 6.2 Users
http://www.cisco.com/go/designzone.
THE DESIGNS ARE SUBJECT TO CHANGE WITHOUT NOTICE. USERS ARE SOLELY RESPONSIBLE FOR THEIR
APPLICATION OF THE DESIGNS. THE DESIGNS DO NOT CONSTITUTE THE TECHNICAL OR OTHER
PROFESSIONAL ADVICE OF CISCO, ITS SUPPLIERS OR PARTNERS. USERS SHOULD CONSULT THEIR OWN
TECHNICAL ADVISORS BEFORE IMPLEMENTING THE DESIGNS. RESULTS MAY VARY DEPENDING ON
FACTORS NOT TESTED BY CISCO.
CCDE, CCENT, Cisco Eos, Cisco Lumin, Cisco Nexus, Cisco StadiumVision, Cisco TelePresence, Cisco
WebEx, the Cisco logo, DCE, and Welcome to the Human Network are trademarks; Changing the Way We
Work, Live, Play, and Learn and Cisco Store are service marks; and Access Registrar, Aironet, AsyncOS,
Bringing the Meeting To You, Catalyst, CCDA, CCDP, CCIE, CCIP, CCNA, CCNP, CCSP, CCVP, Cisco, the
Cisco Certified Internetwork Expert logo, Cisco IOS, Cisco Press, Cisco Systems, Cisco Systems Capital, the
Cisco Systems logo, Cisco Unified Computing System (Cisco UCS), Cisco UCS B-Series Blade Servers, Cisco
UCS C-Series Rack Servers, Cisco UCS S-Series Storage Servers, Cisco UCS Manager, Cisco UCS
Management Software, Cisco Unified Fabric, Cisco Application Centric Infrastructure, Cisco Nexus 9000
Series, Cisco Nexus 7000 Series. Cisco Prime Data Center Network Manager, Cisco NX-OS Software, Cisco
MDS Series, Cisco Unity, Collaboration Without Limitation, EtherFast, EtherSwitch, Event Center, Fast Step,
Follow Me Browsing, FormShare, GigaDrive, HomeLink, Internet Quotient, IOS, iPhone, iQuick Study,
LightStream, Linksys, MediaTone, MeetingPlace, MeetingPlace Chime Sound, MGX, Networkers, Networking
Academy, Network Registrar, PCNow, PIX, PowerPanels, ProConnect, ScriptShare, SenderBase, SMARTnet,
Spectrum Expert, StackWise, The Fastest Way to Increase Your Internet Quotient, TransPath, WebEx, and the
WebEx logo are registered trademarks of Cisco Systems, Inc. and/or its affiliates in the United States and
certain other countries.
All other trademarks mentioned in this document or website are the property of their respective owners. The
use of the word partner does not imply a partnership relationship between Cisco and any other company.
(0809R)
2
Table of Contents
Executive Summary ..................................................................................................................................................................... 8
Introduction ............................................................................................................................................................................. 9
Audience................................................................................................................................................................................. 9
What’s New?........................................................................................................................................................................... 9
Why FlashStack?............................................................................................................................................................... 31
3
VMware Horizon with View Design Fundamentals ................................................................................................................. 39
Products Deployed................................................................................................................................................................ 44
VLANs............................................................................................................................................................................... 46
VMware Clusters............................................................................................................................................................... 47
Validation .................................................................................................................................................................................. 48
4
Pure Storage FlashArray//m50 Configuration for Cisco Validated Design .............................................................................. 90
Volume and Data stores creation on Pure Storage FlashArray//m50 ................................................................................... 101
Configure User Profile Manager Share on Pure Storage FlashArray//m50 ........................................................................... 107
Install and Configure VSUM and Cisco Nexus 1000v .......................................................................................................... 123
Building the Virtual Machines and Environment for Workload Testing .................................................................................. 136
Creating the Golden Image for Horizon Linked Clone Desktops .......................................................................................... 142
Install RDSH Server Role for RDS Hosted Session Hosts Users ....................................................................................... 150
Configure the RDSH Server Roles for RDS Hosted Session User Hosts .......................................................................... 152
Cisco UCS B200 M4 Single Server Configuration and Sizing for Horizon RDSH User Sessions....................................... 155
Cisco UCS B200 M4 Configuration and Sizing for Horizon RDSH User Session Cluster .................................................. 156
5
Cisco UCS B200 M4 Single Server Configuration and Sizing for Horizon Linked Clones ................................................. 157
Cisco UCS B200 M4 Configuration and Sizing for Horizon Linked Clone Cluster ............................................................ 157
Cisco UCS B200 M4 Configuration and Sizing for 5000 User Mixed Workload Horizon RDSH and Linked Clone Scale Test
....................................................................................................................................................................................... 158
Single-Server Recommended Maximum Workload for Cisco UCS B200 M4 Blade Servers ................................................ 166
Single Server Testing with RDSH Remote Desktop Server Hosted Session Users ............................................................... 166
Single Server Testing with Horizon Linked Clone Users ...................................................................................................... 169
Horizon Linked Clone Cluster Testing with 3550 Users ....................................................................................................... 176
5000 Users Mixed Horizon RDSH and Linked Clone Workload Testing ............................................................................... 181
Perfmon Charts for 5000 Users Mixed Workload Test for VMware View Connection Server ........................................... 185
Pure Storage FlashArray//m50 Test Results for Full Scale, Mixed Workload Resiliency Testing .......................................... 186
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations ............................................................... 195
6
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results ..................................................................................... 249
Pure Storage FlashArray//m50 Test Results for 1450 RDSH Sessions ............................................................................ 249
Pure Storage FlashArray//m50 Test Results for 3550 Linked-Clone Windows 7 Sessions ............................................. 251
Pure Storage FlashArray//m50 Test Results for Full Scale, Mixed Workload Scalability................................................... 253
7
Executive Summary
Executive Summary
Cisco® Validated Designs include systems and solutions that are designed, tested, and documented to
facilitate and improve customer deployments. These designs incorporate a wide range of technologies and
products into a portfolio of solutions that have been developed to address the business needs of customers.
Cisco and Pure Storage have partnered to deliver this document, which serves as a specific step by step
guide for implementing this solution. This Cisco Validated Design provides an efficient architectural design that
is based on customer requirements. The solution that follows is a validated approach for deploying Cisco and
Pure Storage technologies as a shared, high performance, resilient, virtual desktop infrastructure.
This document provides architecture reference and design guide for up to 5000 seat mixed workload on Cisco
UCS and Pure Storage FlashArray//m with VMware Horizon 6.2 RDS server-based sessions and Linked Clone
Windows 7 virtual desktops on vSphere 6. The solution is a predesigned, best-practice data center
architecture built on the Cisco Unified Computing System (UCS), the Cisco Nexus® 9000 family of switches,
Cisco MDS 9000 family of Fibre Channel switches and Pure Storage all flash array.
This solution is 100 percent virtualized on Cisco UCS B200 M4 blade server booting via fibre channel SAN
from Pure Storage M50 storage array running VMware vSphere 6.0 U1 hypervisor. The virtual desktops are
configured with VMware Horizon with View 6.2 which now incorporated both tradition persistent and non-
persistent virtual Windows 7/8/10 desktops, hosted applications and Remote desktop service (RDS) server
2012 R2 based desktops provides unparalleled scale and management simplicity. VMware Horizon with View
pool linked-clone floating assignment Windows 7 desktops (3550) and RDS server based desktop sessions
(1450) provisioned desktops on Pure Storage. Where applicable the document provides best practice
recommendations and sizing guidelines for customer deployment of this solution.
The solution provides outstanding virtual desktop end user experience as measured by the Login VSI 4.1
Knowledge Worker workload running in benchmark mode.
The 5000 seat solution provides a large-scale building block that can be replicated to confidently scale-out to
tens of thousands of users.
Solution Overview
Solution Overview
Introduction
The current industry trend in data center design is towards shared infrastructures. By using virtualization along
with pre-validated IT platforms, enterprise customers have embarked on the journey to the cloud by moving
away from application silos and toward shared infrastructure that can be quickly deployed, thereby increasing
agility and reducing costs. Cisco and Pure Storage have partnered to deliver this Cisco Validated Design,
which uses best of breed storage, server and network components to serve as the foundation for desktop
virtualization workloads, enabling efficient architectural designs that can be quickly and confidently deployed.
Audience
The audience for this document includes, but is not limited to; sales engineers, field consultants, professional
services, IT managers, partner engineers, and customers who want to take advantage of an infrastructure built
to deliver IT efficiency and enable IT innovation.
What’s New?
This is the first Cisco Validated Design with Pure Storage. It incorporates the following features:
Validation of Cisco Nexus 9000 with a Pure Storage all flash storage array
Validation of Cisco MDS 9000 with a Pure Storage all flash storage array
Support for the Cisco UCS 3.1(1) release and Cisco UCS B200-M4 servers
Support for the latest release of Pure Storage FlashArray//m hardware and Purity Operating Environment
4.5.5 and 4.6.8.
VMware Horizon 6.2 VDI Linked Clones and RDSH shared server sessions
9
Solution Overview
The data center market segment is shifting toward heavily virtualized private, hybrid and public cloud
computing models running on industry-standard systems. These environments require uniform design points
that can be repeated for ease of management and scalability.
The factors have led to the need for predesigned computing, networking and storage building blocks
optimized to lower the initial design cost, simplify management, and enable horizontal scalability and high
levels of utilization.
10
Solution Summary
Solution Summary
This Cisco Validated Design prescribes a defined set of hardware and software that serves as an integrated
foundation for both VMware Horizon Microsoft Windows 7 virtual desktops and VMware Horizon RDSH server
desktop sessions based on Microsoft Server 2012 R2. The mixed workload solution includes Pure Storage
FlashArray//m storage array, Cisco Nexus® and MDS networking, the Cisco Unified Computing System (Cisco
UCS®), VMware Horizon and VMware vSphere software in a single package. The design is space optimized
such that the network, compute, and storage required can be housed in one data center rack. Switch port
density enables the networking components to accommodate multiple compute and storage configurations of
this kind.
One benefit of the Cisco Validated Design architecture is the ability to customize the environment to suit a
customer's requirements. A Cisco Validated Design can easily be scaled as requirements and demand change.
The unit can be scaled both up (adding resources to a Cisco Validated Design unit) and out (adding more
Cisco Validated Design units). The reference architecture detailed in this document highlights the resiliency,
cost benefit, and ease of deployment of a fibre channel storage solution. Figure 2 shows the VMware Horizon
6.2 on vSphere 6 built on Cisco Validated Design components and the network connections for a configuration
with fibre channel-based storage. This design uses the Cisco Nexus 9000, Cisco MDS 9000, Cisco UCS B-
Series blade servers and the Pure Storage FlashArray//m family of storage controllers connected in a highly
available modular design. This infrastructure is deployed to provide Fibre Channel-booted hosts with block-
level access to shared storage. The reference architecture reinforces the "wire-once" strategy, because as
additional storage is added to the architecture, no re-cabling is required from the hosts to the Cisco UCS
fabric interconnect.
This workforce is also increasingly mobile, conducting business in traditional offices, conference rooms across
the enterprise campus, home offices, on the road, in hotels, and at the local coffee shop. This workforce wants
to use a growing array of client computing and mobile devices that they can choose based on personal
preference. These trends are increasing pressure on IT to ensure protection of corporate data and prevent
data leakage or loss through any combination of user, endpoint device, and desktop access scenarios (Figure
1).
These challenges are compounded by desktop refresh cycles to accommodate aging PCs and bounded local
storage and migration to new operating systems, specifically Microsoft Windows 10.
11
Solution Summary
Some of the key drivers for desktop virtualization are increased data security and reduced TCO through
increased control and reduced management costs.
Simplified
Cisco UCS provides a radical new approach to industry-standard computing and provides the core of the data
center infrastructure for desktop virtualization. Among the many features and benefits of Cisco UCS are the
drastic reduction in the number of servers needed and in the number of cables used per server, and the
capability to rapidly deploy or reprovision servers through Cisco UCS service profiles. With fewer servers and
cables to manage and with streamlined server and virtual desktop provisioning, operations are significantly
simplified. Thousands of desktops can be provisioned in minutes with Cisco UCS Manager service profiles and
Cisco storage partners’ storage-based cloning. This approach accelerates the time to productivity for end
users, improves business agility, and allows IT resources to be allocated to other tasks.
Cisco UCS Manager automates many mundane, error-prone data center operations such as configuration and
provisioning of server, network, and storage access infrastructure. In addition, Cisco UCS B-Series Blade
Servers and C-Series Rack Servers with large memory footprints enable high desktop density that helps
reduce server infrastructure requirements.
Simplification also leads to more successful desktop virtualization implementation. Cisco and its technology
partners like VMware Technologies and Pure have developed integrated, validated architectures, including
predefined converged architecture infrastructure packages such as FlashStack. Cisco Desktop Virtualization
Solutions have been tested with all the leading hypervisors, including VMware vSphere, Citrix XenServer, and
Microsoft Hyper-V.
12
Solution Summary
Secure
Although virtual desktops are inherently more secure than their physical predecessors, they introduce new
security challenges. Mission-critical web and application servers using a common infrastructure such as virtual
desktops are now at a higher risk for security threats. Inter–virtual machine traffic now poses an important
security consideration that IT managers need to address, especially in dynamic environments in which virtual
machines, using VMware vMotion, move across the server infrastructure.
Desktop virtualization, therefore, significantly increases the need for virtual machine–level awareness of policy
and security, especially given the dynamic and fluid nature of virtual machine mobility across an extended
computing infrastructure. The ease with which new virtual desktops can proliferate magnifies the importance of
a virtualization-aware network and security infrastructure. Cisco data center infrastructure (Cisco UCS and
Cisco Nexus Family solutions) for desktop virtualization provides strong data center, network, and desktop
security, with comprehensive security from the desktop to the hypervisor. Security is enhanced with
segmentation of virtual desktops, virtual machine–aware policies and administration, and network security
across the LAN and WAN infrastructure.
Scalable
Growth of a desktop virtualization solution is all but inevitable, so a solution must be able to scale, and scale
predictably, with that growth. The Cisco Desktop Virtualization Solutions support high virtual-desktop density
(desktops per server), and additional servers scale with near-linear performance. Cisco data center
infrastructure provides a flexible platform for growth and improves business agility. Cisco UCS Manager
service profiles allow on-demand desktop provisioning and make it just as easy to deploy dozens of desktops
as it is to deploy thousands of desktops.
Cisco UCS servers provide near-linear performance and scale. Cisco UCS implements the patented Cisco
Extended Memory Technology to offer large memory footprints with fewer sockets (with scalability to up to 1
terabyte (TB) of memory with 2- and 4-socket servers). Using unified fabric technology as a building block,
Cisco UCS server aggregate bandwidth can scale to up to 80 Gbps per server, and the northbound Cisco UCS
fabric interconnect can output 2 terabits per second (Tbps) at line rate, helping prevent desktop virtualization
I/O and memory bottlenecks. Cisco UCS, with its high-performance, low-latency unified fabric-based
networking architecture, supports high volumes of virtual desktop traffic, including high-resolution video and
communications traffic. In addition, Cisco storage partners Pure help maintain data availability and optimal
performance during boot and login storms as part of the Cisco Desktop Virtualization Solutions. Recent Cisco
Validated Designs based on VMware Horizon View, Cisco UCS, and Pure joint solutions have demonstrated
scalability and performance, with up to 5000 desktops up and running in 30 minutes.
Cisco UCS and Cisco Nexus data center infrastructure provides an excellent platform for growth, with
transparent scaling of server, network, and storage resources to support desktop virtualization, data center
applications, and cloud computing.
The simplified deployment of Cisco UCS for desktop virtualization accelerates the time to productivity and
enhances business agility. IT staff and end users are more productive more quickly, and the business can
respond to new opportunities quickly by deploying virtual desktops whenever and wherever they are needed.
13
Solution Summary
The high-performance Cisco systems and network deliver a near-native end-user experience, allowing users
to be productive anytime and anywhere.
The ultimate measure of desktop virtualization for any organization is its efficiency and effectiveness in both
the near term and the long term. The Cisco Desktop Virtualization Solutions are very efficient, allowing rapid
deployment, requiring fewer devices and cables, and reducing costs. The solutions are also very effective,
providing the services that end users need on their devices of choice while improving IT operations, control,
and data security. Success is bolstered through Cisco’s best-in-class partnerships with leaders in virtualization
and storage, and through tested and validated designs and services to help customers throughout the solution
lifecycle. Long-term success is enabled through the use of Cisco’s scalable, flexible, and secure architecture
as the platform for desktop virtualization.
Physical Topology
Figure 2 illustrates the physical architecture.
Figure 2 Architecture
14
Solution Summary
For desktop virtualization, the deployment includes VMware Horizon 6.2 running on VMware vSphere 6. The
design is intended to provide a large scale building block for both RDSH and Linked Clone desktops in the
following ratio:
The data provided in this document will allow our customers to adjust the mix of RDSH and VDI desktops to
suite their environment. For example, additional blade chassis can be deployed to increase compute capacity,
additional disk shelves can be deployed to improve I/O capability and throughput, and special hardware or
software features can be added to introduce new features. This document guides you through the detailed
steps for deploying the base architecture. This procedure covers everything from physical cabling to network,
compute and storage device configurations.
Configuration Guidelines
This document provides details for configuring a fully redundant, highly available configuration for a Cisco
Validated Design for a 5000 seat mixed workload virtual desktop solution with Pure Storage. Configuration
guidelines are provided that refer to which redundant component is being configured with each step. For
example, node01 and node02 are used to identify the two Pure Storage controllers that are provisioned with
this document, Cisco Nexus A or Cisco Nexus B identifies the pair of Cisco Nexus switches that are configured
and Cisco MDS A or Cisco MDS B identifies the pair of Cisco MDS switches that are configured. The Cisco
UCS 6248UP Fabric Interconnects are similarly configured. Additionally, this document details the steps for
provisioning multiple Cisco UCS hosts, and these are identified sequentially: VM-Host-Infra-01, VM-Host-
Infra-02, VM-Host-RDSH-01, VM-Host-VDI-01 and so on. Finally, to indicate that you should include
information pertinent to your environment in a given step, <text> appears as part of the command structure.
15
Solution Components
Solution Components
This section describes the components used in the solution outlined in this study.
Cisco UCS is a next-generation data center platform that unites computing, networking, and storage access.
The platform, optimized for virtual environments, is designed using open industry-standard technologies and
aims to reduce total cost of ownership (TCO) and increase business agility. The system integrates a low-
latency; lossless 10 Gigabit Ethernet unified network fabric with enterprise-class, x86-architecture servers. It
is an integrated, scalable, multi-chassis platform in which all resources participate in a unified management
domain.
Compute: The system is based on an entirely new class of computing system that incorporates blade
servers based on Intel® Xeon® processor E5-2600/4600 v3 and E7-2800 v3 family CPUs.
Network: The system is integrated on a low-latency, lossless, 10-Gbps unified network fabric. This
network foundation consolidates LANs, SANs, and high-performance computing (HPC) networks, which
are separate networks today. The unified fabric lowers costs by reducing the number of network adapters,
switches, and cables needed, and by decreasing the power and cooling requirements.
Virtualization: The system unleashes the full potential of virtualization by enhancing the scalability,
performance, and operational control of virtual environments. Cisco security, policy enforcement, and
diagnostic features are now extended into virtualized environments to better support changing business
and IT requirements.
Storage access: The system provides consolidated access to local storage, SAN storage, and network-
attached storage (NAS) over the unified fabric. With storage access unified, Cisco UCS can access
storage over Ethernet, Fibre Channel, Fibre Channel over Ethernet (FCoE), and Small Computer System
Interface over IP (iSCSI) protocols. This capability provides customers with choice for storage access and
investment protection. In addition, server administrators can preassign storage-access policies for system
connectivity to storage resources, simplifying storage connectivity and management and helping increase
productivity.
Management: Cisco UCS uniquely integrates all system components, enabling the entire solution to be
managed as a single entity by Cisco UCS Manager. The manager has an intuitive GUI, a CLI, and a robust
API for managing all system configuration processes and operations.
16
Solution Components
A cohesive, integrated system that unifies the technology in the data center; the system is managed,
serviced, and tested as a whole
Scalability through a design for hundreds of discrete servers and thousands of virtual machines and the
capability to scale I/O bandwidth to match demand
Cisco UCS Manager provides unified, embedded management of all software and hardware components of
the Cisco Unified Computing System across multiple chassis, rack servers, and thousands of virtual machines.
Cisco UCS Manager manages Cisco UCS as a single entity through an intuitive GUI, a command-line interface
(CLI), or an XML API for comprehensive access to all Cisco UCS Manager Functions.
The fabric interconnects provide the management and communication backbone for the Cisco UCS B-Series
Blade Servers and Cisco UCS 5100 Series Blade Server Chassis. All chassis, and therefore all blades,
attached to the fabric interconnects become part of a single, highly available management domain. In addition,
by supporting unified fabric, the Cisco UCS 6200 Series provides both LAN and SAN connectivity for all blades
in the domain.
17
Solution Components
For networking, the Cisco UCS 6200 Series uses a cut-through architecture, supporting deterministic, low-
latency, line-rate 10 Gigabit Ethernet on all ports, 1-terabit (Tb) switching capacity, and 160 Gbps of
bandwidth per chassis, independent of packet size and enabled services. The product series supports Cisco
low-latency, lossless, 10 Gigabit Ethernet unified network fabric capabilities, increasing the reliability,
efficiency, and scalability of Ethernet networks. The fabric interconnects support multiple traffic classes over a
lossless Ethernet fabric, from the blade server through the interconnect. Significant TCO savings come from an
FCoE-optimized server design in which network interface cards (NICs), host bus adapters (HBAs), cables, and
switches can be consolidated.
18
Solution Components
Cisco UCS combines Cisco UCS B-Series Blade Servers and C-Series Rack Servers with networking and
storage access into a single converged system with simplified management, greater cost efficiency and agility,
and increased visibility and control. The Cisco UCS B200 M4 Blade Server is one of the newest servers in the
Cisco UCS portfolio.
The Cisco UCS B200 M4 delivers performance, flexibility, and optimization for data centers and remote sites.
This enterprise-class server offers market-leading performance, versatility, and density without compromise
for workloads ranging from web infrastructure to distributed databases. The Cisco UCS B200 M4 can quickly
deploy stateless physical and virtual workloads with the programmable ease of use of the Cisco UCS Manager
software and simplified server access with Cisco® Single Connect technology. Based on the Intel Xeon
processor E5-2600 v3 product family, it offers up to 768 GB of memory using 32-GB DIMMs, up to two disk
drives, and up to 80 Gbps of I/O throughput. The Cisco UCS B200 M4 offers exceptional levels of
performance, flexibility, and I/O throughput to run your most demanding applications.
In addition, Cisco UCS has the architectural advantage of not having to power and cool excess switches, NICs,
and HBAs in each blade server chassis. With a larger power budget per blade server, it provides
uncompromised expandability and capabilities, as in the new Cisco UCS B200 M4 server with its leading
memory-slot capacity and drive capacity.
Up to two multicore Intel Xeon processor E5-2600 v3 series CPUs for up to 36 processing cores
19
Solution Components
24 DIMM slots for industry-standard DDR4 memory at speeds up to 2133 MHz, and up to 768 GB of total
memory when using 32-GB DIMMs
Cisco UCS VIC 1340, a 2-port, 40 Gigabit Ethernet and FCoE–capable modular (mLOM) mezzanine
adapter
Provides two 40-Gbps unified I/O ports or two sets of four 10-Gbps unified I/O ports
Cisco FlexStorage local drive storage subsystem, with flexible boot and local storage
capabilities that allow you to:
Configure the Cisco UCS B200 M4 to meet your local storage requirements without hav-
ing to buy, power, and cool components that you do not need
The Cisco UCS B200 M4 server is a half-width blade. Up to eight can reside in the 6-rack-unit (6RU) Cisco
UCS 5108 Blade Server Chassis, offering one of the highest densities of servers per rack unit of blade chassis
in the industry.
Virtualized workloads
Consolidating applications
Virtual desktops
Middleware
20
Solution Components
Highest-performing CPU and memory options without constraints in configuration, power, or cooling
For more information about the Cisco UCS B200 B4, see http://www.cisco.com/c/en/us/support/servers-
unified-computing/ucs-b200-m4-blade-server/model.html
The Cisco UCS VIC 1340 enables a policy-based, stateless, agile server infrastructure that can present over
256 PCIe standards-compliant interfaces to the host that can be dynamically configured as either network
interface cards (NICs) or host bus adapters (HBAs). In addition, the Cisco UCS VIC 1340 supports Cisco® Data
Center Virtual Machine Fabric Extender (VM-FEX) technology, which extends the Cisco UCS fabric
interconnect ports to virtual machines, simplifying server virtualization deployment and management.
Figure 8 illustrates the Cisco UCS VIC 1340 Virtual Interface Cards Deployed in the Cisco UCS B-Series B200
M4 Blade Servers.
21
Solution Components
Figure 8 Cisco UCS VIC 1340 deployed in the Cisco UCS B200 M4
Cisco Switching
Architectural Flexibility
Includes top-of-rack or middle-of-row fiber-based server access connectivity for traditional and leaf-
spine architectures
Leaf node support for Cisco ACI architecture is provided in the roadmap
Increase scale and simplify management through Cisco Nexus 2000 Fabric Extender support
Feature Rich
Enhanced Cisco NX-OS Software is designed for performance, resiliency, scalability, manageability, and
programmability
ACI-ready infrastructure helps users take advantage of automated policy-based systems management
Cisco Nexus 9372PX-E supports IP-based endpoint group (EPG) classification in ACI mode
22
Solution Components
Simplified Operations
Power-On Auto Provisioning (POAP) support allows for simplified software upgrades and configuration file
installation
An intelligent API offers switch management through remote procedure calls (RPCs, JSON, or XML) over a
HTTP/HTTPS infrastructure
Python Scripting for programmatic access to the switch command-line interface (CLI)
Investment Protection
A Cisco 40 Gb bidirectional transceiver allows reuse of an existing 10 Gigabit Ethernet multimode cabling plant
for 40 Gigabit Ethernet Support for 1 Gb and 10 Gb access connectivity for data centers migrating access
switching infrastructure to faster speed. The following is supported:
6 fixed 40-Gbps QSFP+ for uplink connectivity that can be turned into 10 Gb ports through a QSFP to SFP
or SFP+ Adapter (QSA)
Latency of 1 to 2 microseconds
Forms the foundation of virtual network overlays for the Cisco Open Network Environment and Software
Defined Networking (SDN)
Important Differentiators for the Cisco Nexus 1000V for VMware vSphere
The following lists the benefits of the Cisco Nexus 1000V for VMware vSphere:
23
Solution Components
Extensive virtual network services built on Cisco advanced service insertion and routing technology
Feature and management consistency for easy integration with the physical infrastructure
Policy management and control by the networking team instead of the server virtualization team
(separation of duties)
Cisco vPath 2.0 supports service chaining so you can use multiple virtual network services as part of a single
traffic flow. For example, you can simply specify the network policy, and vPath 2.0 can direct traffic through
the Cisco Virtual Security Gateway for Nexus 1000V Switch for a zoning firewall.
Additionally, Cisco vPath works on VXLAN to support movement between servers in different Layer 2 domains.
Together, these features promote highly secure policy, application, and service delivery in the cloud.
MDS 9148S has a pay-as-you-grow model which helps you scale from a 12 port base license to a 48 port
with an incremental 12-port license. This helps customers to pay and activate only the required ports.
MDS 9148S has a dual power supply and FAN trays to provide physical redundancy. The software features,
like ISSU and ISSD, helps with upgrading and downgrading code without reloading the switch and without
interrupting the live traffic.
Features
In-Service Software Upgrade and dual redundant hot-swappable power supplies for high availability
24
Solution Components
Specifications at-a-Glance
Performance and Port Configuration
Up to 256 buffer credits per group of 4 ports (64 per port default, 253 maximum for a single port in the
group)
Supports configurations of 12, 24, 36, or 48 active ports, with pay-as-you-grow, on-demand licensing
Advanced Functions
VMware vSphere 6.0 introduces many enhancements to vSphere Hypervisor, VMware virtual machines,
vCenter Server, virtual storage, and virtual networking, further extending the core capabilities of the vSphere
platform.
Scalability Improvements
ESXi 6.0 dramatically increases the scalability of the platform. With vSphere Hypervisor 6.0, clusters can scale
to as many as 64 hosts, up from 32 in previous releases. With 64 hosts in a cluster, vSphere 6.0 can support
8000 virtual machines in a single cluster. This capability enables greater consolidation ratios, more efficient
25
Solution Components
use of VMware vSphere Distributed Resource Scheduler (DRS), and fewer clusters that must be separately
managed. Each vSphere Hypervisor 6.0 instance can support up to 480 logical CPUs, 12 terabytes (TB) of
RAM, and 1024 virtual machines. By using the newest hardware advances, ESXi 6.0 enables the virtualization
of applications that previously had been thought to be non-virtualizable.
Security Enhancements
ESXi 6.0 offers these security enhancements:
Account management: ESXi 6.0 enables management of local accounts on the ESXi server using new
ESXi CLI commands. The capability to add, list, remove, and modify accounts across all hosts in a
cluster can be centrally managed using a vCenter Server system. Previously, the account and
permission management functions for ESXi hosts were available only for direct host connections. The
setup, removal, and listing of local permissions on ESXi servers can also be centrally managed.
Account lockout: ESXi Host Advanced System Settings have two new options for the management of
failed local account login attempts and account lockout duration. These parameters affect Secure Shell
(SSH) and vSphere Web Services connections, but not ESXi direct console user interface (DCUI) or
console shell access.
Password complexity rules: In previous versions of ESXi, password complexity changes had to be
made by manually editing the /etc/pam.d/passwd file on each ESXi host. In vSphere 6.0, an entry in
Host Advanced System Settings enables changes to be centrally managed for all hosts in a cluster.
Improved auditability of ESXi administrator actions: Prior to vSphere 6.0, actions at the vCenter Server
level by a named user appeared in ESXi logs with the vpxuser username: for example, [user=vpxuser].
In vSphere 6.0, all actions at the vCenter Server level for an ESXi server appear in the ESXi logs with
the vCenter Server username: for example, [user=vpxuser: DOMAIN\User]. This approach provides a
better audit trail for actions run on a vCenter Server instance that conducted corresponding tasks on
the ESXi hosts.
Flexible lockdown modes: Prior to vSphere 6.0, only one lockdown mode was available. Feedback
from customers indicated that this lockdown mode was inflexible in some use cases. With vSphere
6.0, two lockdown modes are available:
In normal lockdown mode, DCUI access is not stopped, and users on the DCUI access
list can access the DCUI.
Exception users: vSphere 6.0 offers a new function called exception users. Exception users are local
accounts or Microsoft Active Directory accounts with permissions defined locally on the host to which
these users have host access. These exception users are not recommended for general user ac-
counts, but they are recommended for use by third-party applications—for service accounts, for ex-
ample—that need host access when either normal or strict lockdown mode is enabled. Permissions
on these accounts should be set to the bare minimum required for the application to perform its task
and with an account that needs only read-only permissions on the ESXi host.
Smart card authentication to DCUI: This function is for U.S. federal customers only. It enables DCUI
login access using a Common Access Card (CAC) and Personal Identity Verification (PIV). The ESXi
host must be part of an Active Directory domain
26
Solution Components
VMware Horizon View Virtual machines and RDSH known as server-based hosted sessions: These are
applications hosted from Microsoft Windows servers to any type of device, including Windows PCs, Macs,
smartphones, and tablets. Some VMware editions include technologies that further optimize the
experience of using Windows applications on a mobile device by automatically translating native mobile-
device display, navigation, and controls to Windows applications; enhancing performance over mobile
networks; and enabling developers to optimize any custom Windows application for any mobile
environment.
VMware Horizon View and RDSH session users also known as server-hosted desktops: These are
inexpensive, locked-down Windows virtual desktops hosted from Windows server operating systems.
They are well suited for users, such as call center employees, who perform a standard set of tasks.
VMware Horizon 6 version 6.2 provides the following new features and enhancements:
Windows 10
The View User Profile Migration tool migrates Windows 7, 8/8.1, Server 2008 R2, or Server 2012 R2
user profiles to Windows 10 user profiles.
View Composer. View Composer and linked clones provide automated and efficient management of
RDS server farms.
Graphics Support. Existing 3D vDGA and GRID vGPU graphics solutions on VDI desktops have been
extended to RDS hosts, enabling graphics-intensive applications to run on RDS desktops and Hosted
Apps.
Enhanced Load Balancing. A new capability provides load balancing of server farm applications based
on memory and CPU resources.
One-Way AD Trusts
One-way AD trust domains are now supported. This feature enables environments with limited trust
relationships between domains without requiring View Connection Server to be in an external domain.
27
Solution Components
Hosted App Support. Support for application remoting allows applications to be launched using global
entitlements across a pod federation.
HTML Access (Blast) Support. Users can use HTML Access to connect to remote desktops and
applications in a Cloud Pod Architecture deployment.
Access Point is a hardened Linux-based virtual appliance that protects virtual desktop and application
resources to allow secure remote access from the Internet. Access Point provides a new
authenticating DMZ gateway to View Connection Server. Smart card support on Access Point is
available as a Tech Preview. Security server will continue to be available as an alternative
configuration. For more information, see Deploying and Configuring Access Point.
FIPS
Install-time FIPS mode allows customers with high security requirements to deploy Horizon 6.
Graphics Enhancements
AMD vDGA enables vDGA pass-through graphics for AMD graphics hardware.
View Administrator shows additional licensing information, including license key, named user and
concurrent connection user count.
Several new features are supported on Horizon 6 for Linux desktops, including NVIDIA GRID vGPU,
vSGA, RHEL 7.1 and Ubuntu 14.04 guest operating systems, and View Agent installation of JRE 8 with
no user steps required.
An RDS host is a server computer that hosts applications and desktop sessions for remote access. An RDS
host can be a virtual machine or a physical server.
An RDS host has the Microsoft Remote Desktop Services role, the Microsoft Remote Desktop Session Host
service, and Horizon Agent installed. Remote Desktop Services was previously known as Terminal
Services. The Remote Desktop Session Host service allows a server to host applications and remote
desktop sessions. With Horizon Agent installed on an RDS host, users can connect to applications and
desktop sessions by using the display protocol PCoIP or Blast Extreme. Both protocols provide an
optimized user experience for the delivery of remote content, including images, audio and video.
28
Solution Components
The performance of an RDS host depends on many factors. For information on how to tune the
performance of different versions of Windows Server, see
http://msdn.microsoft.com/library/windows/hardware/gg463392.aspx.
Horizon 7 supports at most one desktop session and one application session per user on an RDS host.
When users submit print jobs concurrently from RDS desktops or applications that are hosted on the same
RDS host, the ThinPrint server on the RDS host processes the print requests serially rather than in parallel.
This can cause a delay for some users. Note that the print server does not wait for a print job to complete
before processing the next one. Print jobs that are sent to different printers will print in parallel.
If a user launches an application and also an RDS desktop, and both are hosted on the same RDS host,
they share the same user profile. If the user launches an application from the desktop, conflicts may result
if both applications try to access or modify the same parts of the user profile, and one of the applications
may fail to run properly.
The process of setting up applications or RDS desktops for remote access involves the following tasks:
Installing Applications
If you plan to create application pools, you must install the applications on the RDS hosts. If you want
Horizon 7 to automatically display the list of installed applications, you must install the applications so
that they are available to all users from the Start menu. You can install an application at any time before
you create the application pool. If you plan to manually specify an application, you can install the
application at any time, either before or after creating an application pool.
Important
When you install an application, you must install it on all the RDS hosts in a farm and in the same
location on each RDS host. If you do not, a health warning will appear on the View Administrator
dashboard. In such a situation, if you create an application pool, users might encounter an error when
they try to run the application.
When you create an application pool, Horizon 7 automatically displays the applications that are
available to all users rather than individual users from the Start menu on all of the RDS hosts in a farm.
You can choose any applications from that list. In addition, you can manually specify an application that
is not available to all users from the Start menu. There is no limit on the number of applications that
you can install on an RDS host.
RDS Hosts
RDS hosts are server computers that have Windows Remote Desktop Services and View Agent
installed. These servers host applications and desktop sessions that users can access remotely. To
use RDS desktop pools or applications, your end users must have access to Horizon Client 3.0 or later
software.
29
Solution Components
Desktop Pools
There are three types of desktop pools: automated, manual, and RDS. Automated desktop pools use a
vCenter Server virtual machine template or snapshot to create a pool of identical virtual machines.
Manual desktop pools are a collection of existing vCenter Server virtual machines, physical computers,
or third-party virtual machines. In automated or manual pools, each machine is available for one user
to access remotely at a time. RDS desktop pools are not a collection of machines, but instead, provide
users with desktop sessions on RDS hosts. Multiple users can have desktop sessions on an RDS host
simultaneously.
Application Pools
Application pools let you deliver applications to many users. The applications in application pools run
on a farm of RDS hosts.
Farms
Farms are collections of RDS hosts and facilitate the management of those hosts. Farms can have a
variable number of RDS hosts and provide a common set of applications or RDS desktops to users.
When you create an RDS desktop pool or an application pool, you must specify a farm. The RDS hosts
in the farm provide desktop and application sessions to users.
What is FlashStack?
FlashStack CI is a flexible, all-flash converged infrastructure solution that brings the flash revolution to your
data center, faster. It combines the latest in compute, network, storage hardware and virtualization software,
into a single, integrated architecture that speeds time to deployment, lowers overall IT costs and reduces
deployment risk. Highly efficient components reduce the costs associated with power, cooling and data center
space. Based on 100 percent flash storage, FlashStack CI provides the performance and reliability business-
critical applications demand.
The hardware foundation of FlashStack CI includes Pure Storage FlashArrays, Cisco UCS Blade Servers, Cisco
Nexus ethernet switches and Cisco MDS fibre channel switches. VMware vSphere provides the virtualization
technology.
30
Solution Components
FlashStack CI is available from qualified FlashStack Partners who help to provide an excellent converged
infrastructure ownership experience. FlashStack Partners have the knowledge and experience necessary to
help streamline the sizing, procurement, and delivery of your entire system.
Both the hardware and software components are combined into a single integrated unit that helps in faster
deployments and lowers overall IT costs.
Why FlashStack?
The following lists the benefits of FlashStack:
31
Solution Components
Operational Simplicity
Lowest TCO
Free FlashArray controller upgrades every three years with Forever Flash™
Suitable for even large resource-intensive workloads such as real-time analytics or heavy transactional
databases
32
Solution Components
FlashArray//m’s performance can also make your business smarter by unleashing the power of real-time
analytics, driving customer loyalty, and creating new, innovative customer experiences that simply weren’t
possible with disk. All by Transforming Your Storage with FlashArray//m.
FlashArray//m enables you to transform your data center, cloud, or entire business with an affordable all-flash
array capable of consolidating and accelerating all your key applications.
~1kW of power
6 cables
Up to 9 GB/s bandwidth
Modular Scale—Scale FlashArray//m inside and outside of the chassis for generations
The FlashArray//m expands upon the FlashArray’s modular, stateless architecture, designed to enable
expandability and upgradability for generations. The FlashArray//m leverages a chassis-based design with
customizable modules, enabling both capacity and performance to be independently improved over time with
advances in compute and flash, to meet your business’ needs today and tomorrow.
Accelerating Databases and Applications Speed transactions by 10x with consistent low latency, enable
online data analytics across wide datasets, and mix production, analytics, dev/test, and backup workloads
without fear.
Virtualizing and Consolidating Workloads Easily accommodate the most IO-hungry Tier 1 workloads,
increase consolidation rates (thereby reducing servers), simplify VI administration, and accelerate common
administrative tasks.
33
Solution Components
Delivering the Ultimate Virtual Desktop Experience Support demanding users with better performance than
physical desktops, scale without disruption from pilot to >1000’s of users, and experience all-flash
performance with simple management for under $50/desktop.
Protecting and Recovering Vital Data Assets Provide an always-on protection for business-critical data,
maintain performance even under failure conditions, and recover instantly with FlashRecover.
Pure Storage FlashArray sets the benchmark for all-flash enterprise storage arrays. It delivers:
Consistent Performance FlashArray delivers consistent <1ms average latency. Performance is optimized
for the real-world applications workloads that are dominated by I/O sizes of 32K or larger vs. 4K/8K hero
performance benchmarks. Full performance is maintained even under failures/updates.
Less Cost than Disk Inline de-duplication and compression deliver 5 – 10x space savings across a broad
set of I/O workloads including Databases, Virtual Machines and Virtual Desktop Infrastructure. With VDI
workloads data reduction is typically > 10:1.
Mission-Critical Resiliency FlashArray delivers >99.999% proven availability, as measured across the Pure
Storage installed base and does so with non-disruptive everything without performance impact.
Disaster Recovery Built-In FlashArray offers native, fully-integrated, data reduction-optimized backup and
disaster recovery at no additional cost. Setup disaster recovery with policy-based automation within
minutes. And, recover instantly from local, space-efficient snapshots or remote replicas.
Simplicity Built-In FlashArray offers game-changing management simplicity that makes storage installation,
configuration, provisioning and migration a snap. No more managing performance, RAID, tiers or caching.
Achieve optimal application performance without any tuning at any layer. Manage the FlashArray the way
you like it: Web-based GUI, CLI, VMware vCenter, Windows PowerShell, Python, REST API, or OpenStack.
FlashArray//m Specifications
Figure 11 Pure Storage FlashArray//m Portfolio
34
Solution Components
* Effective capacity assumes HA, RAID, and metadata overhead, GB-to-GiB conversion, and includes the
benefit of data reduction with always-on inline deduplication, compression, and pattern removal. Average data
reduction is calculated at 5-to-1.
** Why does Pure Storage quote 32K, not 4K IOPS? The industry commonly markets 4K IOPS benchmarks to
inflate performance numbers, but real-world environments are dominated by IO sizes of 32K or larger.
FlashArray adapts automatically to 512B-32KB IO for superior performance, scalability, and data reduction.
***//m20 can be expanded beyond the 3U base chassis with expansion shelves.
Storage Software Built for Flash—The FlashCare technology virtualizes the entire pool of flash within the
FlashArray, and allows Purity to both extend the life and ensure the maximum performance of consumer-
grade MLC flash.
Granular and Adaptive—Purity Core is based upon a 512-byte variable block size metadata layer. This fine-
grain metadata enables all of Purity’s data and flash management services to operate at the highest
efficiency.
Best Data Reduction Available—FlashReduce implements five forms of inline and post-process data
reduction to offer the most complete data reduction in the industry. Data reduction operates at a 512-byte
aligned variable block size, to enable effective reduction across a wide range of mixed workloads without
tuning.
Highly Available and Resilient—FlashProtect implements high availability, dual-parity RAID-3D, non-
disruptive upgrades, and encryption, all of which are designed to deliver full performance to the FlashArray
during any failure or maintenance event.
Backup and Disaster Recovery Built In—FlashRecover combines space-saving snapshots, replication, and
protection policies into an end-to-end data protection and recovery solution that protects data against loss
35
Solution Components
locally and globally. All FlashProtect services are fully-integrated in the FlashArray and leverage the native
data reduction capabilities.
Pure1
Pure1 Manage—By combining local web-based management with cloud-based monitoring, Pure1 Manage
allows you to manage your FlashArray wherever you are – with just a web browser.
Pure1 Connect—A rich set of APIs, plugin-is, application connectors, and automation toolkits enable you to
connect FlashArray//m to all your data center and cloud monitoring, management, and orchestration tools.
Pure1 Support—FlashArray//m is constantly cloud- connected, enabling Pure Storage to deliver the most
proactive support experience possible. Highly trained staff combined with big data analytics help resolve
problems before they start.
Pure1 Collaborate—Extend your development and support experience online, leveraging the Pure1
Collaborate community to get peer-based support, and to share tips, tricks, and scripts.
Purchase and deploy storage once and once only – then expand capacity and performance incrementally in
conjunction with your business needs and without downtime. You’ll get our maintenance and support for all
components of your system – including flash – and you’ll get modern and new controller upgrades included
every three years with your 3-year renewal. But your maintenance pricing will remain flat. Pure Storage’s vision
for Evergreen Storage is delivered by a combination of the FlashArray’s stateless, modular architecture and the
ForeverFlash business model, enabling you to extend the lifecycle of storage from 3-5 years to a decade or
more.
36
Architecture and Design Considerations for Desktop Virtualization
Knowledge Workers today do not just work in their offices all day – they attend meetings, visit branch
offices, work from home, and even coffee shops. These anywhere workers expect access to all of their
same applications and data wherever they are.
External Contractors are increasingly part of your everyday business. They need access to certain portions
of your applications and data, yet administrators still have little control over the devices they use and the
locations they work from. Consequently, IT is stuck making trade-offs on the cost of providing these
workers a device vs. the security risk of allowing them access from their own devices.
Task Workers perform a set of well-defined tasks. These workers access a small set of applications and
have limited requirements from their PCs. However, since these workers are interacting with your
customers, partners, and employees, they have access to your most critical data.
Mobile Workers need access to their virtual desktop from everywhere, regardless of their ability to connect
to a network. In addition, these workers expect the ability to personalize their PCs, by installing their own
applications and storing their own data, such as photos and music, on these devices.
Shared Workstation users are often found in state-of-the-art university and business computer labs,
conference rooms or training centers. Shared workstation environments have the constant requirement to
re-provision desktops with the latest operating systems and applications as the needs of the organization
change, tops the list.
After the user classifications have been identified and the business requirements for each user classification
have been defined, it becomes essential to evaluate the types of virtual desktops that are needed based on
user requirements. There are essentially five potential desktops environments for each user:
Traditional PC: A traditional PC is what typically constitutes a desktop environment: physical device with a
locally installed operating system.
Hosted Shared Desktop: A hosted, server-based desktop is a desktop where the user interacts through a
delivery protocol. With hosted, server-based desktops, a single installed instance of a server operating
system, such as Microsoft Windows Server 2012, is shared by multiple users simultaneously. Each user
receives a desktop "session" and works in an isolated memory space. Hosted Virtual Desktop: A hosted
virtual desktop is a virtual desktop running on a virtualization layer (ESX). The user does not work with and
sit in front of the desktop, but instead the user interacts through a delivery protocol.
Published Applications: Published applications run entirely on the VMware RDSH Session Hosts and the
user interacts through a delivery protocol. With published applications, a single installed instance of an
application, such as Microsoft Office, is shared by multiple users simultaneously. Each user receives an
application "session" and works in an isolated memory space.
37
Architecture and Design Considerations for Desktop Virtualization
Streamed Applications: Streamed desktops and applications run entirely on the user‘s local client device
and are sent from a server on demand. The user interacts with the application or desktop directly but the
resources may only available while they are connected to the network.
Local Virtual Desktop: A local virtual desktop is a desktop running entirely on the user‘s local device and
continues to operate when disconnected from the network. In this case, the user’s local device is used as
a type 1 hypervisor and is synced with the data center when the device is connected to the network.
For the purposes of the validation represented in this document, both Horizon View Virtual Desktops and
Remote Desktop Services sever Hosted Sessions were validated. Each of the sections provides some
fundamental design decisions for this environment.
The process of analyzing the variety of application and data pairs for an organization will likely be complicated
by the inclusion cloud applications, like SalesForce.com. This application and data analysis is beyond the
scope of this Cisco Validated Design, but should not be omitted from the planning process. There are a variety
of third party tools available to assist organizations with this crucial exercise.
Has a VDI pilot plan been created based on the business analysis of the desktop groups, applications and
data?
Are the required skill sets to execute the VDI project available? Can we hire or contract for them?
Do we have end user experience performance metrics identified for each desktop sub-group?
Below is a short, non-exhaustive list of sizing questions that should be addressed for each user sub-group:
How many virtual desktops will be deployed in the pilot? In production? All Windows 7/8/10?
38
Architecture and Design Considerations for Desktop Virtualization
Will VMware RDSH for Remote Desktop Server Hosted Sessions used?
Are there sufficient IOPS available for the write-intensive VDI workload?
Is user profile management (e.g., non-roaming profile based) part of the solution?
Users can select applications from an easy-to-use “store” that is accessible from tablets, smartphones, PCs,
Macs, and thin clients. VMware Horizon View delivers a native touch-optimized experience with HDX high-
definition performance, even over mobile networks.
In this CVD, linked clone VM provisioning relies on VMware View Composer aligning with VMware Horizon
View Connection Server and vCenter Server components.
Microsoft Windows Server 2012 R2 RDSH VMs are created in vCenter using traditional cloning techniques.
These VMs are created in a dedicated vCenter Cluster and are imported into a Horizon Farm. We used this
technique to have more granular control over the RDSH farm VMs compared with an auto-provisioned farm.
From there Horizon Desktop and RDSH Pools are created to consume the linked clone or server desktop
session resources. Figure 12 illustrates how users access desktops and applications through Unified
Workspace catalogs and Single Sign-On.
39
Architecture and Design Considerations for Desktop Virtualization
Figure 12 Server OS and Desktop OS Machines Configured to Support RDS Hosted Shared Desktops and VDI
Hosted Virtual Desktops
The VMware View Composer pooled desktops solution’s infrastructure is based on software-streaming
technology. After installing and configuring the composed pooled desktops, a single shared disk image
(Master Image) is taken a snapshot of the OS and application image, and then storing that snapshot file
accessible to host(s).
40
Architecture and Design Considerations for Desktop Virtualization
In Figure 15, The image depicting sites, a site was created in two data centers. Having two sites globally,
rather than just one, minimizes the amount of unnecessary WAN traffic. Two Cisco blade servers host the
required infrastructure services (Domain Controllers, DNS, DHCP, Profile, SQL, VMware Horizon View
Connection Servers, View Composer server and web servers).
Based on the requirement and no of data centers or remote location, you can chose any of the available Load
balancing software or tools accelerates the application performance, load balances servers, increases security,
41
Architecture and Design Considerations for Desktop Virtualization
and optimizes the user experience. In this example, two Big-IP Local Traffic Manager are used to provide a
high availability configuration.
BIG-IP Local Traffic Manager has been shown as example for presentation purpose.
Application types: Applications that might not work well with other applications
or might interact with the operating system, such as .NET framework. These
types of applications are ideal for hosting on virtual machines.
42
Architecture and Design Considerations for Desktop Virtualization
You want: Employees with secure remote access to a physical computer without
Remote PC Access
using a VPN. For example, the user may be accessing their physical desktop PC
from home or through a public WIFI hotspot. Depending upon the location, you
may want to restrict the ability to print or copy and paste outside of the desktop.
This method enables BYO device support without migrating desktop images into
the datacenter.
Your users: Employees or contractors that have the option to work from home,
but need access to specific software or data on their corporate desktops to per-
form their jobs remotely.
Application types: Applications that are delivered from an office computer and
display seamlessly in high definition on the remote user's device.
For this Cisco Validated Design, a mix of Remote Desktop Server Hosted sessions (RDSH) using RDS-based
Server OS machines and Hosted Virtual Desktops (HVDs) using VDI-based Desktop OS machines were
configured and tested. The following sections discuss design decisions relative to the VMware Horizon View
deployment, including the CVD test environment.
For RDSH server shared desktops and the linked clone Windows 7 virtual desktops, the following
recommendations are best practices for the delta disks, the internal disks, the disposable disks, user profiles, ,
user data, and application virtualization:
User Profiles: To make sure that user profiles and settings are preserved, you can leverage the profile
management server on the Pure storage Array. We did not deploy this feature in this project.
User data: We recommend hosting user data either on CIFS home directories to preserve data upon VM
reboot or redeploy.
Monitoring and management. We recommend using VMware vCenter Operations Manager for View to
provide monitoring and management of the solution.
43
Solution Hardware and Software
Products Deployed
The architecture deployed is highly modular. While each customer’s environment might vary in its exact
configuration, the reference architecture contained in this document once built, can easily be scaled as
requirements and demands change. This includes scaling both up (adding additional resources within a Cisco
UCS Domain) and out (adding additional Cisco UCS Domains and Pure Storage FlashArrays.) The solution
includes Cisco networking, Cisco UCS and Pure Storage FlashArray//m storage, which efficiently fit into a
single data center rack, including the access layer network switches.
This validated design document details the deployment of 5000 users for a mixed Horizon VMware desktop
workload featuring the following software:
Software Deployed VMware Horizon View 6.2 Remote Desktop Server Hosted sessions(RDSH) on Pure
Storage
VMware Horizon View 6.2 non-persistent linked clone Microsoft Windows 7 Virtual Desktops (VDI) on Pure
Storage
Microsoft Windows 2012 Server for Login VSI Management Console and LVSI Share data servers to
simulate real world VDI workload.
Microsoft Windows Server 2012 R2 Data Center Version for RDSH Servers & Microsoft Windows 7 32-bit
virtual machine operating systems for VDI virtual machines
VMware Horizon View 6.2 Connection Server and Replica Servers for redundancy and support 5000 seat
scale
Hardware Deployed
The workload contains the following hardware as shown in Figure 2:
Four Cisco UCS 5108 Blade Server Chassis with two UCS-IOM-2208XP IO Modules
Two Cisco UCS B200 M4 Blade servers with Intel Xeon E5-2660v3 2.6-GHz 10-core processors, 128GB
RAM 2133-MHz, and VIC1340 mezzanine cards for the hosted infrastructure with N+1 server fault
tolerance
44
Solution Hardware and Software
30 Thirty Cisco UCS B200 M4 Blade servers with Intel Xeon E5-2680v3 2.5-GHz 12-core processors,
384GB RAM 1866-MHz, and VIC1340 mezzanine cards for the virtual desktop workloads with N+1 server
fault tolerance
Pure Storage FlashArray//m50 dual controller storage system, one base disk shelf with 40TB raw space,
one external shelf with 44TB raw space for 88TB total raw space and 8 GB ports for Fibre Channel
connectivity respectively
Software Deployed
Table 2 lists the software and firmware version used in the study.
Logical Architecture
The logical architecture of this solution is designed to support up to 5000 users within four Cisco UCS 5108
Blade server chassis containing 32 blades, which provides physical redundancy for the blade servers for each
workload type.
45
Solution Hardware and Software
This document is intended to allow you to fully configure your environment. In this process, various
steps require you to insert customer-specific naming conventions, IP addresses, and VLAN
schemes, as well as to record appropriate MAC addresses. Table 3 through Table 5 lists the infor-
mation you need to configure your environment.
VLANs
The VLAN configuration recommended for the environment includes a total of seven VLANs as outlined in
Table 4.
46
Solution Hardware and Software
VMware Clusters
The following four VMware Clusters were used in one vCenter data center to support the solution and testing
environment:
Infrastructure Cluster: Infra VMs (vCenter, Active Directory, DNS, DHCP, SQL Server, VMware
Connection Servers, VMware Replica Servers, View Composer Server, and Nexus 1000v Virtual
Supervisor Modules ( VSMs,) etc.)
VDI Cluster 1: VMware Linked Clones VDI VM Pools (Windows 7 SP1 32-bit)*
VMware Supports 32 Hosts in single VMware cluster with latest Horizon View version.
Launcher Cluster 1 and 2: Login VSI Cluster (The Login VSI launcher infrastructure was connected
using the same set of switches and vCenter instance, but was hosted on separate local storage and
servers.)
47
Validation
Validation
This section details the configuration and tuning that was performed on the individual components to produce
a complete, validated solution. Figure 17 illustrates the configuration topology for this solution.
Figure 17 Configuration Topology for Scalable VMware Horizon View 6.2 Mixed Workload
Physical Infrastructure
Solution Cabling
The information in this section is provided as a reference for cabling the physical equipment in this Cisco
Validated Design environment. To simplify cabling requirements, the tables include both local and remote
device and port locations.
The tables in this section contain the details for the prescribed and supported configuration of the Pure
Storage FlashArray//m 50 to the Cisco 6248UP Fabric Interconnects via Cisco MDS 9148S FC switches.
This document assumes that out-of-band management ports are plugged into an existing manage-
ment infrastructure at the deployment site. These interfaces will be used in various configuration
steps.
48
Validation
Be sure to follow the cabling directions in this section. Failure to do so will result in necessary
changes to the deployment procedures that follow because specific port locations are mentioned.
Figure 17 shows a cabling diagram for a VMware Horizon View configuration using the Cisco Nexus 9000 and
Pure Storage FlashArray//m50. The Pure Storage controller and disk shelves should be connected according
to best practices for the specific storage controller and disk shelves.
For devices requiring GbE connectivity, use the GbE Copper SFP+s (GLC-T=).
49
Validation
Eth 1/17- 10GbE UCS 5108 Blade Chassis IOM-A, IOM 1-4
1/20 Chassis 4
50
Validation
Eth 1/17- 10GbE UCS 5108 Blade Chassis IOM-B, IOM 1-4
1/20 Chassis 4
Figure 18 Cable Connectivity Between Cisco Nexus 9372PX, Cisco UCS 6248UP Fabric Interconnects and
Cisco 2208 IO Modules in Cisco 5108AC Blade Chassis
Figure 18 show cable connectivity between the Cisco MDS 9148S and the Cisco 6248 Fabric Interconnects
and the Pure Storage FlashArray//m50.
We used two 8Gb FC connections from each Fabric Interconnect to each MDS switch.
We utilized two 16Gb FC connections from each Pure Storage FlashArray//m50 controller to each MDS switch.
51
Validation
52
Validation
Figure 19 FlashArray//m50 Controller-A and B Connection to MDS 9148S Switches using VSAN 20 for Fabric
A and VSAN 30 Configured for Fabric B Side
Figure 20 Fibre Channel Cable Connectivity from Pure FlashArray//m50 to Cisco MDS 9148S to Cisco 6248
Fabric Interconnects
53
Cisco Unified Computing System Configuration
Cisco UCS Manager Configuration Guides – GUI and Command Line Interface (CLI) Cisco UCS Manager -
Configuration Guides - Cisco
1. Connect a console cable to the console port on what will become the primary fabric interconnect.
2. If the fabric interconnect was previously deployed and you want to erase it to redeploy, follow these
steps:
54
Cisco Unified Computing System Configuration
6. Type “y” or “n” depending on your organization’s security policies, then press Enter.
55
Cisco Unified Computing System Configuration
10. Console onto second fabric interconnect, select console as the configuration method and provide the
following inputs.
11. Open a web browser and go to the Virtual IP address configured above.
1. Open a web browser and navigate to the Cisco UCS 6248 Fabric Interconnect cluster address.
2. Click the Launch UCS Manager link to download the Cisco UCS Manager software.
4. When prompted, enter admin as the user name and enter the administrative password.
56
Cisco Unified Computing System Configuration
1. On the Equipment tab, expand the Fabric Interconnects node and click Fabric Interconnect A.
2. On the General tab in the Actions pane, click Set FC End Host mode.
Both Fabric Interconnects automatically reboot sequentially when you confirm you want to operate
in this mode.
1. After the restarts are complete, from the General tab, Actions pane, click Configure Unified ports.
57
Cisco Unified Computing System Configuration
Ports to the right of the slider will become FC ports. For our study, we configured the last four ports on the
Expansion Module as FC ports.
58
Cisco Unified Computing System Configuration
5. Click Finish, then click Yes to confirm. This action will cause a reboot of the Expansion Module.
59
Cisco Unified Computing System Configuration
After the expansion module reboot, your FC Ports configuration should look like the figure below:
7. Insert Cisco SFP 8 Gbps FC (DS-SFP-FC8-SW) modules into ports 13 through 16 on both Fabric In-
terconnects and cable as prescribed later in this document.
1. In Cisco UCS Manager, in the navigation pane, click the Equipment node and select Equipment in the
list on the left.
60
Cisco Unified Computing System Configuration
6. Click OK.
1. In Cisco UCS Manager, in the navigation pane, click the Equipment tab.
61
Cisco Unified Computing System Configuration
1. In Cisco UCS Manager, in the navigation pane, click the LAN tab.
4. Enter the starting IP address of the block and the number of IP addresses required, and the subnet and
gateway information.
1. In Cisco UCS Manager, in the navigation pane, click the Admin tab.
3. In the Properties pane, select the appropriate time zone in the Timezone menu.
62
Cisco Unified Computing System Configuration
7. Click OK.
1. In Cisco UCS Manager, in the navigation pane, click the Equipment tab.
2. Select Equipment > Fabric Interconnects > Fabric Interconnect A > Fixed Module.
4. Select ports 1 through 28 (Ports highlighted 9, 10, 15 & 16 not configured) that are connected to the
Cisco IO Modules of the four B-Series 5108 Chassis, right-click them, and select Configure as Server
Port.
6. In the left pane, navigate to Fabric Interconnect A. In the right pane, navigate to the Physical Ports tab
> Ethernet Ports tab. Confirm that ports have been configured correctly in the in the Role column.
63
Cisco Unified Computing System Configuration
7. Repeat the above steps for Fabric Interconnect B. The screenshot below shows the server ports for
Fabric B.
To configure network ports used to uplink the Fabric Interconnects to the Cisco Nexus 9172PX switches,
follow these steps:
1. In Cisco UCS Manager, in the navigation pane, click the Equipment tab.
2. Select Equipment > Fabric Interconnects > Fabric Interconnect A > Fixed Module.
4. Select ports 29 through 32 that are connected to the Nexus 9172PX switches, right-click them, and
select Configure as Network Port.
6. In the left pane, navigate to Fabric Interconnect A. In the right pane, navigate to the Physical Ports tab
> Ethernet Ports tab. Confirm that ports have been configured correctly in the in the Role column.
7. Verify the Ports connected to Cisco Nexus upstream swithces are now configured as network ports.
8. Repeat the above steps for Fabric Interconnect B. The screenshot shows the network uplink ports for
Fabric B.
9. Successful configuration should result in ports 29-23 configured as network ports as shown in the
screen shot below:
64
Cisco Unified Computing System Configuration
To configure the necessary port channels in the Cisco UCS environment, complete the following steps:
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
7. Click Next.
9. Click Finish.
10. Repeat steps 1-9 for Fabric Interconnect B, substituting 31 for the port channel number and FI-B-Up-
link for the name. The resulting configuration should look like the screen shot below:
65
Cisco Unified Computing System Configuration
To configure the necessary port channels in the Cisco UCS environment, complete the following steps:
1. In Cisco UCS Manager, click the SAN tab in the navigation pane.
2. Under SAN > SAN Cloud, right-click Fabric A Create Resource Pools
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
7. Enter the seed MAC address and provide the number of MAC addresses to be provisioned.
66
Cisco Unified Computing System Configuration
1. Click the LAN tab in UCS Manager, expand the Pools node, expand the root node, right-click IP Pools,
then click Create IP Pool
67
Cisco Unified Computing System Configuration
4. Complete the starting IP address, size, subnet mask, default gateway, primary and secondary DNS val-
ues for your network, then click OK.
5. Click Next.
68
Cisco Unified Computing System Configuration
6. Click Finish.
1. In Cisco UCS Manager, click the SAN tab in the navigation pane.
3. Under WWPN Pools, right click WWPN Pools and select Create WWPN Pool.
69
Cisco Unified Computing System Configuration
6. Click Next.
9. Click Finish.
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
8. Click Next.
70
Cisco Unified Computing System Configuration
11. Specify a size for the UUID block that is sufficient to support the available blade or server resources .
Consider creating unique server pools to achieve the granularity that is required in your environ-
ment.
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
7. Click Next.
8. Select two (or more) servers to be used for the VMware management cluster and click >> to add them
to the Infra_Pool server pool.
9. Click Finish.
11. Create additional Server Pools for Horizon Linked Clone servers and Horizon RDSH servers
Create VLANs
To configure the necessary virtual local area networks (VLANs) for the Cisco UCS environment, complete the
following steps:
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
In this procedure, six unique VLANs are created. Refer to Table 12.
71
Cisco Unified Computing System Configuration
3. Right-click VLANs.
5. Enter MGMT as the name of the VLAN to be used for in-band management traffic.
6. Keep the Common/Global option selected for the scope of the VLAN.
10. Repeat the above steps to create all VLANs and configure the Default VLAN as native.
72
Cisco Unified Computing System Configuration
Create VSANs
To configure the necessary virtual storage area networks (VSANs) for the Cisco UCS environment, complete
the following steps:
1. In Cisco UCS Manager, click the SAN tab in the navigation pane.
In this procedure, two VSANs are created. When these VSANs are created, be sure to add them to
the port-channel uplink created earlier.
5. Enter VSAN20 as the name of the VSAN to be used for in-band management traffic.
9. Repeat the above steps on Fabric B with VSAN30 to create the VSANs necessary for this solution.
73
Cisco Unified Computing System Configuration
10. After configuring VSANs both sides, go into the port-channel created earlier in the section ‘Create up-
links for MDS 9148S and add the respective VSANs to their port channels. VSAN20 in this study is
assigned to Fabric A and VSAN30 is assigned to Fabric B. (VSAN20 Should only be on Fabric A and
VSAN30 on Fabric B).
11. Go to the Port-Channel for each Fabric and assign the VSAN appropriately.
74
Cisco Unified Computing System Configuration
To create a firmware management policy for a given server configuration in the Cisco UCS environment,
complete the following steps:
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
7. Select the version 2.2. (6c) for both the Blade Package
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
4. On the Best Effort row, enter 9216 in the box under the MTU column.
75
Cisco Unified Computing System Configuration
6. Click OK.
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
76
Cisco Unified Computing System Configuration
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
77
Cisco Unified Computing System Configuration
7. Click Finish.
1. In Cisco UCS Manager, click the Servers tab in the navigation pane.
78
Cisco Unified Computing System Configuration
1. In Cisco UCS Manager, click the LAN tab in the navigation pane.
10. Under VLANs, select the checkboxes for MGMT, Default, VDI, Infra, and vMotion.
79
Cisco Unified Computing System Configuration
26. Under VLANs, select the checkboxes for MGMT, Default, VDI, Infra, and vMotion.
80
Cisco Unified Computing System Configuration
1. In Cisco UCS Manager, click the SAN tab in the navigation pane.
81
Cisco Unified Computing System Configuration
6. Click Next.
9. Click Add.
82
Cisco Unified Computing System Configuration
83
Cisco Unified Computing System Configuration
84
Cisco Unified Computing System Configuration
26. Click Create Boot Policy to create a Boot From SAN policy.
85
Cisco Unified Computing System Configuration
To create a boot from SAN policy, go to Cisco UCS Manager, complete the following steps:
1. Right-click the Boot Policies option shown below and select Create Boot Policy.
86
Cisco Unified Computing System Configuration
2. Enter a name for the boot policy and expand the vHBA menu shown below:
4. After selecting Add SAN Boot, add the primary vHBA as shown below. Note that the vHBA name
needs to match exactly.
87
Cisco Unified Computing System Configuration
6. Add SAN Boot Targets to the primary and secondary. The SAN boot targets also include primary and
secondary options in order to maximize resiliency and number of paths.
7. Highlight the SAN primary and select Add SAN Boot Target to SAN Primary.
8. From the Pure Storage GUI, find and enter the Pure Storage WWN for Controller 0, Fibre Channel Port
0. This information can be found in the Pure Storage GUI under System > Host Connections at the
bottom of the screen under Target Ports:
9. When the Pure WWNs are recorded, use port CT0.FC0 for the first Boot Target WWPN:
10. Add a secondary SAN Boot Target by clicking the Add SAN Boot Target to SAN Primary while the pri-
mary SAN Boot option is highlighted. Enter the Pure Storage WWPN for CT1.FC0.
88
Cisco Unified Computing System Configuration
11. Repeat this procedure for the secondary SAN boot target and use WWPN for CT0.FC1 and CT1.FC1 in
the primary and secondary SAN boot options.
12. Below you can see a properly configured boot from SAN UCS policy. The next step is to create and
attach the boot volumes to the hosts from within the Pure GUI:
Provisioning a volume for SAN boot for Pure Storage is extremely simple. The only difference between a SAN
volume and a vCenter datastore is that you will connect the SAN boot volume to the single host that will be
booting from the volume.
13. Login in to Pure Storage GUI, select the + icon next to Volumes under the Storage tab.
14. Create multiple volumes since there are a large number of ESXi hosts and you need to create separate
boot volumes for each one.
89
Cisco Unified Computing System Configuration
15. To connect the LUN to an ESXi host, select a newly created volume under the Volumes tab and then
select Connect Hosts from the options under the gear located on the right:
16. Click the single host you to connect to this volume and click Confirm.
17. Repeat this procedure with all ESXi hosts. Remember, it is important to only connect a single host to
each boot LUN.
The expansion shelf is connected with four 12GB SAS cables (two per controller) in order to ensure full
redundancy, availability and performance of the shelf in the event of a single controller reboot or failure.
A Pure Storage Systems Engineer or authorized partner will perform the installation and initial setup of the
array. Setup (racking, power, cabling and array setup) is typically completed in less than one hour and will
only require between 12-16 cables (that number is variable based upon number of SAN FC connections) for
the base unit and expansion shelf. Arrays without an expansion shelf require as little as six cables in total.
The Cisco UCS hosts are redundantly connected to the array controllers (via the Cisco MDS 9148S) with two
FC connections to each controller from two HBAs on each Cisco UCS host over the 8GB Fibre Channel
protocol for a total of four logical paths for 32GB/s of total throughput with no single point of failure. 16GB is
also supported by default on the FlashArray//m50 but was not used within this Cisco Validated Design. The
FlashArray//m50 array will support up to eight FC connections in total in the baseline system and using all eight
connections is recommended in production deployments in order to maximize resiliency and throughput
capabilities of the array. However, we will show that a large scale VDI deployment can be accomplished with
only half of the available FC ports in use to further prove the resiliency and performance of the
FlashArray//m50.
90
Cisco Unified Computing System Configuration
Behind the bezel on the FlashArray//m50 the 2 NVRAM modules and 20 drive bays can be accessed as shown
in the screenshot below:
The next figure shows the back of the array which houses the two controllers. The rear of the array in this
example includes Fibre-Channel connectivity (8x 16GB/s ports), though 10GB iSCSI is also supported.
The table below lists the default port services associated with the rear view of the array. Pure Storage
FlashArray//m Default Ports Overview
Furthermore, default port configurations for SAN switch connectivity can be found below as well as further
expansion options. Note that the port configurations of the new //m10 array is identical to the //m20 array.
91
Cisco Unified Computing System Configuration
Lastly, the rear of the expansion shelf shown below includes redundant SAS connectivity to each controller
and the ability to daisy-chain additional shelfs for added capacity non-disruptively.
Figure 26 Pure Storage FlashAray External Shelf SAS Connectivity for each Controller Overview
Figure 27 Pure Storage FlashArray//m50 Storage Connectivity to Cisco MDS 9148S Switches Overview
92
Cisco Unified Computing System Configuration
All Pure Storage FlashArray//m systems require a minimum of three static IP addresses for management. One
IP address is assigned to each controller Ethernet port (eth0) and a virtual IP is assigned between the two
controllers. Initial setup of the array and assignment of these IP addresses is accomplished by connecting to
the console via either of the KVM ports shown above on the back of the array.
All GUI-based operations shown in this section can also be accomplished via the Purity command line. For the
sake of consistency, we will show all configuration steps via the GUI but the array is so simple to use that the
entire user manual fits on a double-sided folded business card as shown below in Figure 28. In addition,
Purity features a restful API for additional extensibility and usage across a multitude of platforms and languages
including PowerShell and the VMware vRealize Automation Suite, among many others.
1. To do so, navigate to an IP address assigned for array management from the web browser. The fol-
lowing login screen should appear:
93
Cisco Unified Computing System Configuration
After login, the Pure GUI should appear. Note that a newly deployed array will have 0% space utilization and
not be driving any IO.
Figure 29 through Figure 32 shows examples of datastores created for data and for SAN boot.
94
Cisco Unified Computing System Configuration
Figure 29 Data stores Created for Data and for SAN Boot
95
Cisco Unified Computing System Configuration
The Pure Storage GUI shows a variety of metrics and status for the array. On the left, the Array Status can be
seen which shows a top-level health status of the backplane, NVRam modules, controllers and individual solid
state drive bays. To the right of that, the amount of storage space being used, data reduction and percentage
of space in use is shown across the top. Below that, key metrics including Latency, IOPs and Bandwidth are
shown in real-time and the zoom can be adjusted to as fine an interval as 15 minutes or as wide of an interval
of 30 days. You have the capability to see additional performance and capacity metrics for individual, or
groups of datastores over a wider timespan under the ‘Analysis’ tab
When the array has been brought online, networking information can be confirmed by completing the following
steps:
1. Go to the System tab which will show the following window. Note that the initial view shows a much
more detailed health view for all array components.
2. Click Configuration to expand the menu and click Networking to bring up the management networking
configuration screen:
3. To update network information, hover the mouse cursor over the Ethernet port you wish to change and
click Edit on the gear that appears. The following screen shots an example of editing a network inter-
face:
96
Cisco Unified Computing System Configuration
When the management network has been confirmed as being properly setup, the next step is to complete the
SAN zoning of the Cisco MCS 9148S switches so that the storage array and Cisco UCS servers are able to
communicate. To do the SAN zoning, complete the following steps:
1. Record the WWN’s of the array from within the Pure GUI. This can be found under System > Configu-
ration > Host Connections as shown below:
2. At the bottom of the above screen, the WWN’s of the Pure Storage array are displayed. Note that only
6 out of 8 are being shown below. Record your unique WWN values so that they can be included in
the Cisco MDS zonesets along with the WWNs that were created when the Cisco UCS Service Profiles
were built.
When the Cisco MDS 9148S zonesets have been created and activated, the WWNs of the Cisco UCS servers
should automatically become visible to the Pure Storage array (note that it can sometimes take up to 30
minutes for the UCS WWNs to become visible.
97
Cisco Unified Computing System Configuration
1. To create a Host within the Pure Storage GUI, first navigate to the storage tab and you will see next to
the Hosts menu a ‘+’ sign:
2. Click the ‘+’ sign will spawn a window for you to create single, or multiple hosts. Since we are using
multiple hosts in our design we will show the method for creating several hosts at once.
3. After clicking the ‘Create Multiple..’ option above, the following screen displays which enables batch
creation of multiple hosts. In this example, we are creating two hosts.
Below you can see that our two hosts have been created and the next step is to connect their WWNs. WWNs
can be matched up with specific Cisco UCS servers from within Cisco UCS Manager.
98
Cisco Unified Computing System Configuration
Figure 41 Configuring the ESXi host WWNs from Cisco UCS for the Hosts
4. To configure the FC WWNs, highlight a host, select the ‘Host Ports’ tab and then select the ‘Configure
Fibre Channel WWNs’ button as shown in the below screenshot.
5. Click them to select and move them over to the right-hand column and click ‘Confirm’ when com-
pleted. In this Cisco Validated Design, each Cisco UCS host will have two separate vHBAs to provide
both performance and resiliency.
Below, you see a properly setup UCS host within the Pure GUI. Repeat the above procedure for host creation
and configuring the Fibre Channel WWNs for all Cisco UCS servers.
6. When the hosts have been created and the WWNs have been assigned to them, then click the ‘Hosts’
button again and this time select ‘Create Host Group’.
99
Cisco Unified Computing System Configuration
7. The next step is to add the two example hosts we previously created to this newly created Host
Group. To accomplish this, highlight the Host Group, click the gear to the right and select ‘Add Hosts’:
8. The following screenshots display; click the hosts you wish to add to the Host Group on the left column
in order to move them to the right. When all hosts have been included, click ‘Confirm.’
100
Cisco Unified Computing System Configuration
The following screenshot shows the populated Host Group with two hosts as members.
Repeat this procedure for all Host Groups that are required. For this Cisco Validated Design we created three
separate Host Groups for the following components:
With the Hosts and Host Group setup completed, the array is now ready to serve VMware Horizon desktops
and the associated infrastructure.
2. On the collapsible Volumes menu on the left, select the ‘+’ button:
101
Cisco Unified Computing System Configuration
4. Select the newly created Volume under the Volumes collapsible menu on the left-hand side of the GUI
and click on the gear icon to connect it to host(s) or host group(s):
5. Click the appropriate Host Groups that require access to the Volume so that they are under the ‘Select
Host Groups’ list and then click Confirm when they have been added:
Figure 54 Add Host Group or Group Hosts Configured for the Volume
6. Storage setup has been completed and the datastore now needs to be added from within vSphere.
From within the vSphere client, select a host from the Host Group that was connected to the volume in
the previous step, go to Related Objects and then Datastore and click the icon to add the new datas-
tore:
102
Cisco Unified Computing System Configuration
8. Name the datastore and select the appropriate volume from the list shown. It might be necessary to
rescan the vHBA on the host in order to see a newly created datastore.
9. Leave the partition configuration options at default values and click Next.
103
Cisco Unified Computing System Configuration
Repeat this process for each of the datastores listed in the Datastore Table 14.
We used the below backup schedule shown in Table 14 for the datastores defined in the previous section.
Important to note is that since the linked-clone desktops were set as being non-persistent, we elected not to
assign any snapshot schedule to either datastore since no persistent data was being saved between user
sessions. For environments with persistent desktops, it’s a recommended practice to use a separate
datastore with a defined snapshot schedule to protect against the loss of any user data.
104
Cisco Unified Computing System Configuration
Figure 60 Pure GUI to Create a Snapshot Protection for Host or Host Group
2. Using the Snapshot schedule for the Infrastructure datastore, assign the following values to create the
snapshot schedule and retention policy:
105
Cisco Unified Computing System Configuration
3. When the Snapshot and retention policy has been defined, add the Datastore to the policy by clicking
the ‘Members’ tab of the Source Group:
4. Click the Volume(s) you want to include in the Snapshot policy to move them to the right. Click Con-
firm when completed.
5. After clicking Confirm, the Datastore should be listed as a Snapshot policy member:
Figure 64 Infra-Datastore Volume Applied to the Newly Created Snapshot Retention Policy for Backup
Repeat these steps for the RDSH and ESXi Boot LUNs Data stores. It’s recommended to include all ESXi boot
LUNs in a single Snapshot policy for simplicity.
106
Cisco Unified Computing System Configuration
Therefore, all ESXi servers were configured to change the default PSP for Pure Storage FlashArray devices
from MRU to Round Robin (with an advanced configuration to alternate to logical paths after every I/O). The
following command was run on each ESXi server prior to the presentation of FlashArray devices:
esxcli storage nmp satp rule add -s "VMW_SATP_ALUA" -V "PURE" -M "FlashArray" - P
"VMW_PSP_RR" -O "iops=1"
A PowerShell script has been created to check for, and apply these best practices automatically to the entire
vCenter cluster. Please see the attachment in Appendix B for further information.
The profiles were stored on a separate D: partition on the Windows 2012 server.
107
Cisco Unified Computing System Configuration
1. In this solution we utilized the Cisco MDS 9148S Switches for Fiber Channel Switching. For racking,
cable and initial setup of the MDS switches, please refer to the Quick Start Guide:
http://www.cisco.com/c/en/us/td/docs/switches/datacenter/mds9000/hw/9148/quick/guide/MDS_91
48_QSG.pdf
2. When the MDS switch is racked and can be logged into it can now be configured to communicate with
the Cisco UCS Fabric Interconnects.
3. In this study, we used two separate fabrics each with their own unique VSAN. Fabric A is configured
for VSAN20 while Fabric B for VSAN30. In our initial Cisco UCS configuration you will see where we
configured fiber cables on ports 13 and 14 and configured a FC port-channel. FI-A’s FC port channel
is configured for VSAN20 and FI-B’s FC port-channel for VSAN30.
Physically, the Fabric Interconnects extended ports 13 and 14 run to the MDS switch ports 1 and 2.
108
Cisco Unified Computing System Configuration
After the ports and port channels are configured, the next steps are to configure the zones and Active Zoneset
Database in the MDS switches. The below commands show how to add in a single host on both MDS A and
B. You will need to configure all hosts that will access the Pure Array in these commands. Then entire MDS
switch configuration is included in this document in Appendix A.
9148S-A
Configure Terminal
Zoneset name VDI-Infra-A vsan 20
Zone name {ESXi hostname-fc0} vsan 20
Member pwwn {ESXi Host pwwn for fc0}
Member pwwn {PURE pwwn Controller A, Port 1}
Member pwwn {PURE pwwn Controller B, Port 1}
Zone commit vsan 20
Zoneset name VDI-Infra-A vsan 20
Member {ESXi hostname-fc0}
Exit
Zoneset activate name VDI-Infra-A vsan 20
Zone commit vsan 20
Exit
Copy running-config startup-config
9148S-B
Configure Terminal
Zoneset name VDI-Infra-B vsan 30
Zone name {ESXi hostname-fc1} vsan 30
Member pwwn {ESXi Host pwwn for fc1}
Member pwwn {PURE pwwn Controller A, Port 2}
Member pwwn {PURE pwwn Controller B, Port 2}
Zone commit vsan 30
Zoneset name VDI-Infra-B vsan 30
Member {ESXi hostname-fc1}
109
Cisco Unified Computing System Configuration
Exit
Zoneset activate name VDI-Infra-B vsan 30
Zone commit vsan 30
Exit
Copy running-config startup-config
Several methods exist for installing ESXi in a VMware environment. These procedures focus on how to use the
built-in keyboard, video, mouse (KVM) console and virtual media features in Cisco UCS Manager to map
remote installation media to individual servers and connect to their boot logical unit numbers (LUNs).
2. Type your email or customer number and the password and then click Log in.
This ESXi 6.0 Cisco custom image includes updates for the fnic and eNIC drivers. The versions that
are part of this image are: eNIC: 2.3.0.7; fNIC: 1.6.0.25
Install ESXi
To install VMware ESXi to the SAN-bootable LUN of the hosts, complete the following steps on each host:
1. On reboot, the machine detects the presence of the ESXi installation media. Select the ESXi installer
from the menu that is displayed.
2. After the installer is finished loading, press Enter to continue with the installation.
3. Read and accept the end-user license agreement (EULA). Press F11 to accept and continue.
4. Select the PURE LUN that was previously set up as the installation disk for ESXi and press Enter to
continue with the installation.
110
Cisco Unified Computing System Configuration
7. The installer issues a warning that existing partitions will be removed from the volume. Press F11 to
continue with the installation.
8. After the installation is complete, clear the Mapped checkbox (located in the Virtual Media tab of the
KVM console) to unmap the ESXi installation image.
The ESXi installation image must be unmapped to make sure that the server reboots into ESXi and
not into the installer.
9. The Virtual Media window might issue a warning stating that it is preferable to eject the media from the
guest. Because the media cannot be ejected and it is read-only, simply click Yes to unmap the image.
10. From the KVM tab, press Enter to reboot the server.
To configure the ESXi host with access to the management network, complete the following steps:
1. After the server has finished rebooting, press F2 to customize the system.
3. Select the Configure the Management Network option and press Enter.
6. From the Configure Management Network menu, select IP Configuration and press Enter.
7. Select the Set Static IP Address and Network Configuration option by using the space bar.
10. Enter the default gateway for the first ESXi host.
13. Using the spacebar, unselect Enable IPv6 (restart required) and press Enter.
Because the IP address is assigned manually, the DNS information must also be entered manually.
111
Cisco Unified Computing System Configuration
17. Enter the fully qualified domain name (FQDN) for the first ESXi host.
20. Press Y to confirm the changes and return to the main menu.
21. The ESXi host reboots. After reboot, press F2 and log back in as root.
22. Select Test Management Network to verify that the management network is set up correctly and press
Enter.
112
Cisco Unified Computing System Configuration
1. Open a web browser on the management workstation and navigate to the VM-Host-01 management
IP address.
This application is downloaded from the VMware website and Internet access is required on the
management workstation.
5. Click Next.
113
Cisco Unified Computing System Configuration
8. Click Install.
9. Click Finish.
1. Open the recently downloaded VMware vSphere Client and enter the IP address of VM-Host-01 as the
host you are trying to connect to
1. From the vSphere 6 download page on the VMware Web site, download the vCenter ISO file for the
vCenter Server appliance onto your system.
2. Open the vSphere ISO via Windows Explorer and double-click the vcsa-setup.htm file.
114
Cisco Unified Computing System Configuration
3. Click Install.
5. Enter the IP of the ESXi host the vCenter Appliance will reside. Click Next.
115
Cisco Unified Computing System Configuration
Figure 72 Provide Host IP or FQDN and User Name, Password Credentials of the Host to Connect
7. Provide a name for the vCenter appliance, then click Next to continue.
8. Select Install vCenter Server with and Embedded Platform Services Controller (unless your environ-
ment already has a PSC).
116
Cisco Unified Computing System Configuration
9. Create a new SSO domain (unless your environment already has and SSO domain. Multiple SSO do-
mains can co-exist).
10. Select the proper appliance size for your deployment. In our study, Large was selected.
117
Cisco Unified Computing System Configuration
Figure 79 PostgreSQL
It is important to note at this step that you should create a DNS A record for your appliance prior to
running the install. The services will fail to startup and your install will fail if it cannot resolve
properly.
118
Cisco Unified Computing System Configuration
Figure 80 Provide the Necessary Network Gateways and DNS Server Information
16. When your install completes successfully, you can now login to your Web Client and begin adding
hosts and configuring clusters.
119
Cisco Unified Computing System Configuration
120
Cisco Unified Computing System Configuration
24. To create a Datacenter, click the icon in the center pane which has the green plus symbol above it.
26. Click the vCenter server available in the list. Click OK to continue.
27. Right-click Datacenters > VDI-DC in the list in the center pane, then click New Cluster.
121
Cisco Unified Computing System Configuration
If mixing Cisco UCS B 200 M4 servers within a vCenter cluster, it is necessary to enable VMware
Enhanced vMotion Compatibility (EVC) mode. For more information about setting up EVC mode, re-
fer to Enhanced vMotion Compatibility (EVC) Processor Support.
33. Right-click Infra in the center pane and click Add Host.
35. Type root as the user name and root password as the password. Click Next to Continue.
122
Cisco Unified Computing System Configuration
41. Review the configuration parameters then click Finish to add the host.
To set up the primary Cisco Nexus 1000V VSM on the Cisco Nexus 1110-X A, complete the following steps:
Verifying the authenticity of an image is optional. You can still install the image without validating
its authenticity.
Nexus1000v-vsum.1.5.x-pkg.zip image
signature.txt file
cisco_n1k_image_validation_v_1_5_x script
123
Cisco Unified Computing System Configuration
5. Check the output. If the validation is successful, the following message displays:
5. Click Open.
6. Click Next.
124
Cisco Unified Computing System Configuration
10. Name the Virtual Machine, select the VDI-DC datacenter and click Next.
12. Select Infra-Datastore and the Thin Provision virtual disk format and click Next.
125
Cisco Unified Computing System Configuration
19. Select the Home button in VMware vSphere Web Client and select Hosts and Clusters.
20. Expand the Infrastructure cluster and select the Virtual Switch Update Manager VM.
21. In the center pane, select Launch Remote Console. If a security warning pops up, click Allow.
24. When the VM has completely booted up, log out and log back into the VMware vSphere Web Client.
25. Review and click Next to install the click Nexus 1000V.
126
Cisco Unified Computing System Configuration
Cisco VSUM is a virtual appliance that is registered as a plug-in to the VMware vCenter Server.
The Cisco VSUM is the GUI that you use to install, migrate, monitor, and upgrade the VSMs in high
availability (HA) or standalone mode and the VEMs on ESX/ESXi hosts.
127
Cisco Unified Computing System Configuration
Optionally, an existing VSM can be used that is provided by a Cisco Nexus Cloud Services Platform
(CSP).
1. Log in to VMware vSphere Web Client and choose Home > Cisco Virtual Switch Update Manager >
Nexus 1000V > Install, and then choose the data center. The installation screen appears.
2. In the Nexus 1000v Switch Deployment area, choose I want to deploy new control plane (VSM).
3. In the Cisco Nexus 1000V Switch Deployment Type area, install the switches as an HA pair. By default,
the High Availability Pair is selected.
128
Cisco Unified Computing System Configuration
The Cisco Nexus 1000V VSM uses the management network to communicate with vCenter Server
and ESXi. The management and control port group can use the same VLAN.
6. In the Host Selection area, click Suggest to choose two hosts based on the details provided in the
Cisco Nexus 1000V Switch Deployment Type area. The IP address of the hosts on which the switch
will be deployed.
7. The primary switch is deployed on Infrastructure Host 1 and the secondary switch is deployed on In-
frastructure Host 2. Click Pick a Host to override the system choices.
8. Choose the system-selected datastore that you want to override. Choose PURE Infra-Datastore as the
datastore for each host.
9. Provide Host IP address where the Virtual Ethernet Modules to be created. (note it requires two esxi
hosts for installing VEM primary and secondary modules for redundancy purpose)
10. In the Switch Configuration area, enter 70 as the domain ID for the switch.
11. The domain ID is common for both the primary and secondary switches and it should be unique for
every new switch. The range for the domain is from 1 to 1023.
12. In the Virtual Supervisor Module (VSM) configuration area, enter the Switch Name, IP Address, Subnet
Mask, and Gateway Address.
129
Cisco Unified Computing System Configuration
The Cisco Nexus 1000V installation is confirmed when the primary task Create Nexus 1000v Switch
has the status Completed. A typical installation of the switch takes about 4 minutes.
1. Using an SSH client, log in to the primary Cisco Nexus 1000V VSM as admin.
Any VLAN that has a VMKernal port should be assigned as a system VLAN on both the uplink and
the vEthernet ports of the virtual switch.
config t
ntp server <<var_switch_a_ntp_ip>> use-vrf management
ntp server <<var_switch_b_ntp_ip>> use-vrf management
vlan <<var_ib-mgmt_vlan_id>> 160
name IB-MGMT-VLAN
vlan <<var_vmotion_vlan_id>> 166
name vMotion-VLAN
The Cisco Nexus 1000V is currently limited to 1024 Max ports per profile. This solution is comprised
of 3500 plus virtual desktop machines for the user workload and requires four dedicated port-pro-
files(VDI,VDI-1,VDI-2,VDI-3).
130
Cisco Unified Computing System Configuration
131
Cisco Unified Computing System Configuration
state enabled
port-profile type vethernet n1kv-L3
capability l3control
vmware port-group
switchport mode access
port-profile type vethernet VDI
vmware port-group
switchport mode access
switchport access vlan <<var_vdi_1_vlan_id>> 162
no shutdown
max-ports 1024
system vlan <<var_vdi_1_vlan_id>> 162
state enabled
port-profile type vethernet VDI-1
vmware port-group
switchport mode access
switchport access vlan <<var_vdi_1_vlan_id>> 162
no shutdown
max-ports 1024
system vlan <<var_vdi_1_vlan_id>> 162
state enabled
port-profile type vethernet VDI-2
vmware port-group
switchport mode access
switchport access vlan <<var_vdi_1_vlan_id>> 162
no shutdown
max-ports 1024
system vlan <<var_vdi_1_vlan_id>> 162
state enabled
port-profile type vethernet VDI-3
vmware port-group
switchport mode access
switchport access vlan <<var_vdi_1_vlan_id>> 162
no shutdown
max-ports 1024
system vlan <<var_vdi_1_vlan_id>> 162
state enabled
switchport access vlan <<var_OB-MGMT_vlan_id>> 164
no shutdown
system vlan <<var_OB-MGMT_vlan_id>> 164
state enabled
exit
copy run start
132
Cisco Unified Computing System Configuration
1. Back in the VMware vSphere Web Client, from the Home tab, select Cisco Virtual Switch Update Man-
ager.
3. Select Configure.
6. Click Manage.
8. Expand the Infrastructure ESXi Cluster and select one of the Infrastructure Management Hosts.
9. Click Suggest.
10. Scroll down to Physical NIC Migration and expand each ESXi host.
11. On both hosts, unselect vmnic0, and select vmnic1. For vmnic1, select the system-uplink Profile.
12. Scroll down to VM Kernel NIC Setup and expand both ESXi hosts.
13. All VMkernel ports should already have the appropriate checkboxes selected.
15. Select the IB-MGMT-VLAN profile for the VSUM and vCenter Virtual Machines.
The progress of the virtual switch installation can be monitored from the c# interface.
1. In the VMware vSphere Web Client window, select Home > Hosts and Clusters.
2. On the left expand the Datacenter and cluster, and select the first VMware ESXi host.
3. In the center pane, select the Manage tab, then select Networking.
4. Select vSwitch0. All of the port groups on vSwitch0 should be empty. Click the red X under Virtual
switches to delete vSwitch0.
133
Cisco Unified Computing System Configuration
5. Click Yes to remove vSwitch0. It may be necessary to refresh the Web Client to see the deletion.
6. The Nexus 1000V VSM should now be the only virtual switch. Select it and select the third icon above
it under Virtual switches (Manage the physical network adapters connected to the selected switch).
8. For UpLink01, select the system-uplink port group and make sure vmnic0 is the Network adapter.
Click OK.
9. Click OK to complete adding the Uplink. It may be necessary to refresh the Web Client to see the addi-
tion.
11. From the SSH client that is connected to the Cisco Nexus 1000V, run show interface status to verify
that all interfaces and port channels have been correctly configured.
12. Run show module and verify that the one ESXi host is present as a module.
134
Cisco Unified Computing System Configuration
14. Repeat the above steps to migrate the remaining ESXi hosts to the Nexus 1000V.
Using vTracker show commands, you can access consolidated network information across the following views:
Upstream View—Provides information on all the virtual ports connected to an upstream physical switch. The
view is from top of the network to the bottom.
VM vNIC View—Provides information about the virtual machines (VMs) that are managed by the Cisco
Nexus 1000V switch. The vNIC view is from the bottom to the top of the network.
VM Info View—VM Info View—Provides information about all the VMs that run on each server module.
Module pNIC View—Provides information about the physical network interface cards (pNIC) that are
connected to each Virtual Ethernet Module (VEM).
VLAN View—Provides information about all the VMs that are connected to specific VLANs.
vMotion View—Provides information about all the ongoing and previous VM migration events.
The vTracker feature on the Cisco Nexus 1000V switch provides information about the virtual network.
environment. To connect SSH to the primary VSM, complete the following step:
1. From an SSH interface connected to the Cisco Nexus 1000V VSM, enter the following:
config t
feature vtracker
copy run start
show vtracker upstream-view
show vtracker vm-view vnic
135
Cisco Unified Computing System Configuration
1. Download the view composer installer from VMware and click install on the View composer Windows
server Image. In this study, View Composer 6.2.0.3001314 version used
136
Cisco Unified Computing System Configuration
137
Cisco Unified Computing System Configuration
8. Click Install.
138
Cisco Unified Computing System Configuration
10. Completed the installation and view composer is ready to serve the composer operations or desktops
provisioning
To view connection and replica server installation, complete the following step:
1. Download the View Composer server Installer. In this study, View Connection Server 6.2.0-3005368
version.
139
Cisco Unified Computing System Configuration
140
Cisco Unified Computing System Configuration
141
Cisco Unified Computing System Configuration
To install additional Horizon View Replica servers, complete the following steps:
3. Proceed with the rest of the procedure to complete Replica Servers installation.
1. Select ESXi host in Infrastructure cluster and create a virtual machine to use as the Golden Image with
windows 7 OS. We used windows 7 32 bit OS for our testing.
Memory : 2048MB
142
Cisco Unified Computing System Configuration
Processor : 2vCPU
Hard Disk : 20 GB
4. Select datastore.
143
Cisco Unified Computing System Configuration
144
Cisco Unified Computing System Configuration
10. Complete the remaining steps to install the Win7 Operating System.
www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf
https://my.vmware.com/web/vmware/details?productId=529&downloadGroup=VIEW-620-ENT-GA
145
Cisco Unified Computing System Configuration
4. Install service packs and hot fixes required for the additional software components that were added.
1. Shut down the Windows 7 Golden Image virtual machine to take a snapshot.
2. Right-click Windows 7 Golden Image Virtual Machine Properties to take a snapshot which is required
for the virtual desktop deployment.
3. Provide the name and description for the Snapshot and click OK.
4. Click OK.
146
Cisco Unified Computing System Configuration
1. Right-click the powered off virtual machine after taking a snapshot and select Template and click Con-
vert to template.
2. Provide a name to the template and provide the host /cluster, data store details.
3. Select Guest Customization and check the radio button for Customize using the Customization Wizard.
4. Click Next.
147
Cisco Unified Computing System Configuration
148
Cisco Unified Computing System Configuration
149
Cisco Unified Computing System Configuration
Now the Golden image is ready for VDI virtual machines deployment.
Install RDSH Server Role for RDS Hosted Session Hosts Users
This section describes the installation of RDS Role for RDSH server configuration.
Prerequisite: Configure a Windows 2012 or R2 server for configuring the RDSH server roles. In this study we
have configured 2012 R2 servers for RDS server roles.
To create a new Virtual machine as a Windows 2012 R2 server, complete the following steps:
2. Select configuration.
150
Cisco Unified Computing System Configuration
4. Select datastore.
6. Disk size.
151
Cisco Unified Computing System Configuration
10. Follow the link provided below to optimize windows server 2012.
Complete the required Windows server 2012 creation and provide the necessary IP address or DHCP and
reboot the server.
Now you need to customize the windows servers using VMware Customization Specs.
https://labs.vmware.com/flings/vmware-os-optimization-tool
Configure the RDSH Server Roles for RDS Hosted Session User Hosts
To configure the RDSH server roles for RDS hosted session user hosts, complete the following steps:
152
Cisco Unified Computing System Configuration
5. Select remote desktop services and select remote desktop session host.
6. Select remote desktops services tools and select licensing diagnostic tool.
7. RDSH Server has been created and ready for RDSH server role for users to connect.
153
Cisco Unified Computing System Configuration
8. Reboot the Server and install other updates and necessary software.
3. Install service packs and hot fixes required for the additional software components that were added.
We also tested a mixed workload scenario with 5000 VDI users running both RDS Hosted server sessions and
VDI virtual machines on Cisco B200 M4 B-Series servers. We separately tested single server RDS Hosted
sessions, RDSH cluster testing, single server VDI virtual machines along with VDI cluster testing all running on
Cisco UCS B200 M4 servers.
The tested system was comprised of the following hardware and software components:
Hardware components:
2 Cisco UCS B200 M4 Blade Servers (2 Intel Xeon processor E5-2660 v3 CPUs at 2.6 GHz, with 128 GB
of memory per blade server [16 GB x 8 DIMMs at 2133 MHz]) for two Infrastructure blades.
30 Cisco UCS B200 M4 Blade Servers (2 Intel Xeon processor E5-2680 v3 CPUs at 2.5 GHz, with 384 GB
of memory per blade server [16 GB x 24 DIMMs at 2133 MHz]) for RDS work load blade.
154
Cisco Unified Computing System Configuration
Pure Storage FlashArray//m50 + 44TB external shelf (Purity Operating Environment 4.5.5)
Software components:
Windows 2012 R2 Server for RDS Hosted Sessions. Total of 9 Servers Configured.
Microsoft Windows 7
Cisco UCS B200 M4 Single Server Configuration and Sizing for Horizon RDSH User
Sessions
Nine Windows 2012 R2 virtual servers on one Cisco UCS B200 M4 blade server configured for Hosted Shared
Desktops to host 180 User Sessions.
Figure 96 RDS Single Server Testing – Recommended workload: 180 User Sessions per Server
155
Cisco Unified Computing System Configuration
Cisco UCS B200 M4 Configuration and Sizing for Horizon RDSH User Session Cluster
Nine Cisco UCS B200 M4 blade servers running Windows 2012 R2 Servers configured for Hosted Shared
Desktops to host 1450 User Sessions. N+1 server fault tolerance is provided by this cluster.
156
Cisco Unified Computing System Configuration
Cisco UCS B200 M4 Single Server Configuration and Sizing for Horizon Linked Clones
181 Horizon virtual Windows 7 linked clone desktops on one Cisco UCS B200 M4 blade server represents our
maximum recommended workload for a single server.
Figure 98 Cisco UCS B200 M4 Server for Single Server Scalability Horizon View 6.2
Cisco UCS B200 M4 Configuration and Sizing for Horizon Linked Clone Cluster
3550 Horizon virtual Windows 7 linked clone desktops on 21 Cisco UCS B200 M4 blade server represents our
cluster scale point, providing N+1 server fault tolerance at the cluster level.
157
Cisco Unified Computing System Configuration
Figure 99 21x Cisco UCS B200 M4 Server for Blade Server VDI Scalability VMware Horizon View VDI Desktops
Cisco UCS B200 M4 Configuration and Sizing for 5000 User Mixed Workload Horizon
RDSH and Linked Clone Scale Test
1450 Horizon RDSH sessions on 9 Cisco UCS B200 M4 blade servers running Windows 2012 R2 virtual
servers configured for Hosted Shared Desktops.
3550 Horizon virtual Windows 7 linked clone desktops in two Horizon Linked Clone Desktop pools on 21 Cisco
UCS B200 M4 blade server represents our cluster scale point, providing N+1 server fault tolerance at the
cluster level.
158
Cisco Unified Computing System Configuration
Figure 100 30x Cisco UCS B200 M4 Servers for both RDSH and VDI mixed work load with server N+1 fault tol-
erance at the cluster level.
159
Testing Methodology and Success Criteria
The testing results focused on the entire process of the virtual desktop lifecycle by capturing metrics during
the desktop boot-up, user logon and virtual desktop acquisition (also referred to as ramp-up,) user workload
execution (also referred to as steady state), and user logoff for the RDSH Servers Session under test.
Test metrics were gathered from the virtual desktop, storage, and load generation software to assess the
overall success of an individual test cycle. Each test cycle was not considered passing unless all of the
planned test users completed the ramp-up and steady state phases (described below) and unless all metrics
were within the permissible thresholds as noted as success criteria.
Three successfully completed test cycles were conducted for each hardware configuration and results were
found to be relatively consistent from one test to the next.
You can obtain additional information and a free test license from http://www.loginvsi.com.
Testing Procedure
The following protocol was used for each test cycle in this study to insure consistent results.
All Launchers for the test were shut down. They were then restarted in groups of 10 each minute until the
required number of launchers was running with the Login VSI Agent at a “waiting for test to start” state.
In addition, Cisco requires that the Login VSI Benchmark method is used for all single server and scale testing.
This assures that our tests represent real-world scenarios. For each of the three consecutive runs on single
server tests, the same process was followed. Complete the following steps:
All Infrastructure VMs used in test run (AD, SQL, View Connection brokers, image
mgmt., etc.)
3. Time 0:05: Boot RDS Machines using VMware Horizon View 6.2 Administrator Console.
5. Time 0:35 Single Server or Scale target number of RDS Servers registered on XD.
160
Testing Methodology and Success Criteria
No more than 60 Minutes of rest time is allowed after the last desktop is registered and available on
VMware Horizon View 6.2 Administrator Console dashboard. Typically a 20-30 minute rest period
for Windows 7 desktops and 10 minutes for RDS VMs is sufficient.
6. Time 1:35 Start Login VSI 4.1.4 Office Worker Benchmark Mode Test, setting auto-logoff time at 900
seconds, with Single Server or Scale target number of desktop VMs utilizing sufficient number of
Launchers (at 20-25 sessions/Launcher).
7. Time 2:23 Single Server or Scale target number of desktop VMs desktops launched (48 minute
benchmark launch rate).
All sessions launched must become active for a valid test run within this window.
9. Time 2:40 Login VSI Test Ends (based on Auto Logoff 900 Second period designated above).
All sessions launched and active must be logged off for a valid test run. The VMware Horizon View
6.2 Administrator Dashboard must show that all desktops have been returned to the regis-
tered/available state as evidence of this condition being met.
12. Time 3:15 Copy all log files off to archive; Set virtual desktops to maintenance mode through broker;
Shutdown all Windows 7 machines.
Success Criteria
Our “pass” criteria for this testing follows is Cisco will run tests at a session count level that effectively utilizes
the blade capacity measured by CPU utilization, memory utilization, storage utilization, and network utilization.
We will use Login VSI to launch version 4.1 Knowledge Worker workloads. The number of launched sessions
must equal active sessions within two minutes of the last session launched in a test as observed on the VSI
Management console.
The VMware Horizon with View Connection Server Dashboard will be monitored throughout the steady state to
make sure of the following:
No sessions move to unregistered, unavailable or available state at any time during steady state
Within 20 minutes of the end of the test, all sessions on all launchers must have logged out automatically and
the Login VSI Agent must have shut down. Stuck sessions define a test failure condition.
161
Testing Methodology and Success Criteria
Cisco requires three consecutive runs with results within +/-1% variability to pass the Cisco Validated Design
performance criteria. For white papers written by partners, two consecutive runs within +/-1% variability are
accepted. (All test data from partner run testing must be supplied along with proposed white paper.)
We will publish Cisco Validated Designs with our recommended workload following the process above and will
note that we did not reach a VSImax dynamic in our testing.
The purpose of this testing is to provide the data needed to validate VMware Horizon View 6.2 Hosted Shared
Desktop with VMware View 6.2 Composer provisioning using Microsoft Windows Server 2012 R2 sessions on
Cisco UCS B200 M4
The information contained in this section provides data points that a customer may reference in designing their
own implementations. These validation results are an example of what is possible under the specific
environment conditions outlined here, and do not represent the full characterization of VMware products.
Four test sequences, each containing three consecutive test runs generating the same result, were performed
to establish single blade performance and multi-blade, linear scalability.
Login VSI is different in approach. Login VSI is not primarily designed to be a steady state benchmark
(however, if needed, Login VSI can act like one). Login VSI was designed to perform benchmarks for SBC or
VDI workloads through system saturation. Login VSI loads the system with simulated user workloads using well
known desktop applications like Microsoft Office, Internet Explorer and Adobe PDF reader. By gradually
increasing the amount of simulated users, the system will eventually be saturated. Once the system is
saturated, the response time of the applications will increase significantly. This latency in application response
times show a clear indication whether the system is (close to being) overloaded. As a result, by nearly
overloading a system it is possible to find out what its true maximum user capacity is.
After a test is performed, the response times can be analyzed to calculate the maximum active
session/desktop capacity. Within Login VSI this is calculated as VSImax. When the system is coming closer to
its saturation point, response times will rise. When reviewing the average response time it will be clear the
response times escalate at saturation point.
This VSImax is the “Virtual Session Index (VSI)”. With Virtual Desktop Infrastructure (VDI) and Terminal Services
(RDS) workloads this is valid and useful information. This index simplifies comparisons and makes it possible to
understand the true impact of configuration changes on hypervisor host or guest level.
An alternative to the Login VSI method would be to generate user actions client side through the remoting
protocol. These methods are always specific to a product and vendor dependent. More importantly, some
protocols simply do not have a method to script user actions client side.
For Login VSI the choice has been made to execute the scripts completely server side. This is the only
practical and platform independent solutions, for a benchmark like Login VSI.
162
Testing Methodology and Success Criteria
The five operations from which the response times are measured are:
Loading and initiating VSINotepad.exe and opening the openfile dialog. This operation is handled by the
OS and by the VSINotepad.exe itself through execution. This operation seems almost instant from an end-
user’s point of view.
Loading and initiating VSINotepad.exe and opening a file. This operation is also handled by the OS and by
the VSINotepad.exe itself through execution. This operation seems almost instant from an end-user’s point
of view.
This action copy's a random file and compresses it (with 7zip) with high compression enabled. The
compression will very briefly spike CPU and disk IO.
This action copy's a random file and compresses it (with 7zip) with low compression enabled. The
compression will very briefly disk IO and creates some load on the CPU.
CPU
Calculates a large array of random data and spikes the CPU for a short period of time.
These measured operations within Login VSI do hit considerably different subsystems such as CPU (user and
kernel), Memory, Disk, the OS in general, the application itself, print, GDI, etc. These operations are specifically
short by nature. When such operations become consistently long: the system is saturated because of
excessive queuing on any kind of resource. As a result, the average response times will then escalate. This
effect is clearly visible to end-users. If such operations consistently consume multiple seconds the user will
regard the system as slow and unresponsive.
Figure 101 Sample of a VSI Max Response Time Graph, Representing a Normal Test
163
Testing Methodology and Success Criteria
Figure 102 Sample of a VSI Test Response Time Graph with a Clear Performance Issue
When the test is finished, VSImax can be calculated. When the system is not saturated, and it could complete
the full test without exceeding the average response time latency threshold, VSImax is not reached and the
amount of sessions ran successfully.
The response times are very different per measurement type, for instance Zip with compression can be around
2800 ms, while the Zip action without compression can only take 75ms. This response time of these actions
are weighted before they are added to the total. This ensures that each activity has an equal impact on the
total response time.
In comparison to previous VSImax models, this weighting much better represent system performance. All
actions have very similar weight in the VSImax total. The following weighting of the response times are applied.
The following actions are part of the VSImax v4.1 calculation and are weighted as follows (US notation):
CPU: 0.75
This weighting is applied on the baseline and normal Login VSI response times.
With the introduction of Login VSI 4.1 we also created a new method to calculate the base phase of an
environment. With the new workloads (Taskworker, Powerworker, etc.) enabling 'base phase' for a more
reliable baseline has become obsolete. The calculation is explained below. In total 15 lowest VSI response
time samples are taken from the entire test, the lowest 2 samples are removed and the 13 remaining samples
are averaged. The result is the Baseline. The calculation is as follows:
The VSImax average response time in Login VSI 4.1.x is calculated on the amount of active users that are
logged on the system.
164
Testing Methodology and Success Criteria
Always a 5 Login VSI response time samples are averaged + 40% of the amount of “active” sessions. For
example, if the active sessions is 60, then latest 5 + 24 (=40% of 60) = 31 response time measurement are
used for the average calculation.
To remove noise (accidental spikes) from the calculation, the top 5% and bottom 5% of the VSI response time
samples are removed from the average calculation, with a minimum of 1 top and 1 bottom sample. As a result,
with 60 active users, the last 31 VSI response time sample are taken. From those 31 samples the top 2
samples are removed and lowest 2 results are removed (5% of 31 = 1.55, rounded to 2). At 60 users the
average is then calculated over the 27 remaining results.
VSImax v4.1.x is reached when the VSIbase + a 1000 ms latency threshold is not reached by the average VSI
response time result. Depending on the tested system, VSImax response time can grow 2 - 3x the baseline
average. In end-user computing, a 3x increase in response time in comparison to the baseline is typically
regarded as the maximum performance degradation to be considered acceptable.
In VSImax v4.1.x this latency threshold is fixed to 1000ms, this allows better and fairer comparisons between
two different systems, especially when they have different baseline results. Ultimately, in VSImax v4.1.x, the
performance of the system is not decided by the total average response time, but by the latency is has under
load. For all systems, this is now 1000ms (weighted).
The threshold for the total response time is: average weighted baseline response time + 1000ms.
When the system has a weighted baseline response time average of 1500ms, the maximum average response
time may not be greater than 2500ms (1500+1000). If the average baseline is 3000 the maximum average
response time may not be greater than 4000ms (3000+1000).
When the threshold is not exceeded by the average VSI response time during the test, VSImax is not hit and
the amount of sessions ran successfully. This approach is fundamentally different in comparison to previous
VSImax methods, as it was always required to saturate the system beyond VSImax threshold.
Lastly, VSImax v4.1.x is now always reported with the average baseline VSI response time result. For example:
“The VSImax v4.1 was 125 with a baseline of 1526ms”. This helps considerably in the comparison of systems
and gives a more complete understanding of the system. The baseline performance helps to understand the
best performance the system can give to an individual user. VSImax indicates what the total user capacity is
for the system. These two are not automatically connected and related:
When a server with a very fast dual core CPU, running at 3.6 GHZ, is compared to a 10 core CPU, running at
2,26 GHZ, the dual core machine will give and individual user better performance than the 10 core machine.
This is indicated by the baseline VSI response time. The lower this score is, the better performance an
individual user can expect.
However, the server with the slower 10 core CPU will easily have a larger capacity than the faster dual core
system. This is indicated by VSImax v4.1.x, and the higher VSImax is, the larger overall user capacity can be
expected.
With Login VSI 4.1.x a new VSImax method is introduced: VSImax v4.1. This methodology gives much better
insight in system performance and scales to extremely large systems.
165
Test Results
Test Results
This recommended maximum workload approach allows you to determine the server N+1 fault tolerance load
the blade can successfully support in the event of a server outage for maintenance or upgrade.
Our recommendation is that the Login VSI Average Response and VSI Index Average should not exceed the
Baseline plus 2000 milliseconds to insure that end user experience is outstanding. Additionally, during steady
state, the processor utilization should average no more than 90-95% when running the maximum
recommended load.
Callouts have been added throughout the data charts to indicate each phase of testing.
Test
Phase Description
Boot Start all RDS and VDI virtual machines at the same time
The Login VSI phase of test is where sessions are launched and start executing the work-
Login load over a 48 minutes duration
The steady state phase is where all users are logged in and performing various workload
Steady tasks such as using Microsoft Office, Web browsing, PDF printing, playing videos, and
state compressing files
Logoff Sessions finish executing the Login VSI workload and logoff
Single Server Testing with RDSH Remote Desktop Server Hosted Session
Users
The maximum recommended load for Horizon RDSH user sessions on a single Cisco UCS B200 M4 blade
server is 180 sessions running on 9 virtual Windows Server 2012 R2 machines.
166
Test Results
Figure 103 Login VSI Chart for RDSH Single Server Testing
167
Test Results
Figure 104 ESXTOP Core Utilization for RDSH Single Server testing
5:17:55 PM
5:26:10 PM
5:01:26 PM
5:04:11 PM
5:06:56 PM
5:12:26 PM
5:15:11 PM
5:20:40 PM
5:23:25 PM
5:28:55 PM
5:31:40 PM
5:34:26 PM
5:37:11 PM
5:39:57 PM
5:42:43 PM
5:45:29 PM
5:48:15 PM
5:51:01 PM
5:53:48 PM
5:56:33 PM
5:59:18 PM
6:02:03 PM
6:04:48 PM
6:07:33 PM
Figure 105 ESXTOP Utilization Time for RDSH Single Server testing
\\ESXi-RDSH-03\Physical Cpu(_Total)\% Util Time
100
90
80
70 \\RVDI-
60 03\Physical
50 Cpu(_Total)\%
Util Time
40
30
20
10
0
5:55:42 PM
5:01:26 PM
5:04:01 PM
5:06:35 PM
5:09:10 PM
5:11:44 PM
5:14:19 PM
5:16:54 PM
5:19:28 PM
5:22:03 PM
5:24:38 PM
5:27:12 PM
5:29:47 PM
5:32:22 PM
5:34:57 PM
5:37:32 PM
5:40:07 PM
5:42:43 PM
5:45:19 PM
5:47:54 PM
5:50:30 PM
5:53:06 PM
5:58:17 PM
6:00:51 PM
6:03:26 PM
6:06:00 PM
6:08:35 PM
Figure 106 Network Adapter (VMNICs) MB received /Transmitted for Sec for the RDSH Single Server testing
160
\\RVDI-03\Network
Port(vSwitch0:335544
140 34:vmnic0)\MBits
Received/sec
120
100 \\RVDI-03\Network
Port(vSwitch0:335544
80 36:vmnic1)\MBits
Received/sec
60
\\RVDI-03\Network
40 Port(vSwitch0:335544
34:vmnic0)\MBits
20 Transmitted/sec
0
5:01:2…
5:04:1…
5:06:5…
5:09:4…
5:12:2…
5:15:1…
5:17:5…
5:20:4…
5:23:2…
5:26:1…
5:28:5…
5:31:4…
5:34:2…
5:37:1…
5:39:5…
5:42:4…
5:45:2…
5:48:1…
5:51:0…
5:53:4…
5:56:3…
5:59:1…
6:02:0…
6:04:4…
6:07:3…
\\RVDI-03\Network
Port(vSwitch0:335544
36:vmnic1)\MBits
Transmitted/sec
168
Test Results
Figure 107 Disk Adapter MBytes Read / Write for Sec for the RDSH Single Server testing
45 \\RVDI-03\Physical Disk
Adapter(vmhba2)\MByt
40 es Read/sec
35
\\RVDI-03\Physical Disk
30 Adapter(vmhba1)\MByt
es Read/sec
25
\\RVDI-03\Physical Disk
20 Adapter(vmhba0)\MByt
es Read/sec
15
10 \\RVDI-03\Physical Disk
Adapter(vmhba2)\MByt
5 es Written/sec
0 \\RVDI-03\Physical Disk
5:01:26 PM
5:04:32 PM
5:07:37 PM
5:10:43 PM
5:13:48 PM
5:16:54 PM
5:19:59 PM
5:23:05 PM
5:26:10 PM
5:29:16 PM
5:32:22 PM
5:35:28 PM
5:38:34 PM
5:41:40 PM
5:44:47 PM
5:47:54 PM
5:51:01 PM
5:54:08 PM
5:57:15 PM
6:00:20 PM
6:03:26 PM
6:06:31 PM
6:09:37 PM
Adapter(vmhba1)\MByt
es Written/sec
\\RVDI-03\Physical Disk
Adapter(vmhba0)\MByt
es Written/sec
Figure 108 Non Kernel Mbytes for RDSH Single Server testing
\\ESXi-RDSH-03\Memory\NonKernel MBytes
160000
140000
120000
100000 \\RVDI-
03\Mem
80000 ory\NonK
60000 ernel
MBytes
40000
20000
0
5:04:21…
5:01:26…
5:07:16…
5:10:12…
5:13:07…
5:16:02…
5:18:57…
5:21:53…
5:24:48…
5:27:43…
5:30:38…
5:33:34…
5:36:30…
5:39:26…
5:42:22…
5:45:19…
5:48:15…
5:51:12…
5:54:08…
5:57:04…
6:00:00…
6:02:55…
6:05:50…
6:08:45…
169
Test Results
Figure 109 Login VSI graph for Non-persistent VDI users Single Server
170
Test Results
Figure 110 ESXTOP CPU Core utilization for Non-persistent VDI users Single Server testing.
\\VDI-15\Physical Cpu(_Total)\% Core Util Time
100
80
\\VDI-
60 15\Physical
Cpu(_Total)
40 \% Core Util
Time
20
0
9:16:05…
9:47:28…
8:52:26…
8:55:49…
8:59:11…
9:02:33…
9:05:56…
9:09:18…
9:12:41…
9:19:30…
9:22:56…
9:26:22…
9:29:50…
9:33:19…
9:36:49…
9:40:22…
9:43:56…
9:50:57…
9:54:21…
9:57:43…
10:01:0…
10:04:2…
10:07:5…
Figure 111 ESXTOP CPU Core Total utilization for Non-persistent VDI users Single Server testing
80
60 \\VDI-
15\Physical
40 Cpu(_Total)
\% Util
Time
20
0
9:11:34…
9:47:40…
8:52:26…
8:55:37…
8:58:48…
9:02:00…
9:05:11…
9:08:22…
9:14:46…
9:17:59…
9:21:13…
9:24:27…
9:27:43…
9:31:00…
9:34:17…
9:37:36…
9:40:57…
9:44:20…
9:50:57…
9:54:10…
9:57:21…
10:00:32…
10:03:44…
10:06:55…
171
Test Results
Figure 112 Esxtop for Non-persistent VDI users Single Server testing Network Adapter MBits Received /Sent
/Sec
\\VDI-
160 15\Network
Port(DvsPortset-
140 0:50331662:vmni
c1)\MBits
120 Received/sec
\\VDI-
100 15\Network
Port(DvsPortset-
80 0:50331661:vmni
c0)\MBits
60 Received/sec
\\VDI-
40 15\Network
Port(DvsPortset-
20 0:50331664:vmk
1)\MBits
0
9:43:08…
8:52:26…
8:56:00…
8:59:33…
9:03:07…
9:06:41…
9:10:15…
9:13:49…
9:17:25…
9:21:02…
9:24:38…
9:28:17…
9:31:58…
9:35:39…
9:39:23…
9:46:53…
9:50:34…
9:54:10…
9:57:43…
10:01:17…
10:04:52…
10:08:26…
Transmitted/sec
\\VDI-
15\Network
Port(DvsPortset-
0:50331663:vmk
0)\MBits
Transmitted/sec
Figure 113 Esxtop for Non-persistent VDI users Single Server testing Disk Adapter Mbytes Received /Sent /Sec
25 \\VDI-
15\Physical Disk
Adapter(vmhba
20 1)\MBytes
Read/sec
\\VDI-
15 15\Physical Disk
Adapter(vmhba
0)\MBytes
10 Read/sec
\\VDI-
15\Physical Disk
5 Adapter(vmhba
1)\MBytes
Written/sec
0
9:29:5…
10:01:…
8:52:2…
8:55:4…
8:59:1…
9:02:3…
9:05:5…
9:09:1…
9:12:4…
9:16:0…
9:19:3…
9:22:5…
9:26:2…
9:33:1…
9:36:4…
9:40:2…
9:43:5…
9:47:2…
9:50:5…
9:54:2…
9:57:4…
10:04:…
10:07:…
\\VDI-
15\Physical Disk
Adapter(vmhba
0)\MBytes
Written/sec
Figure 114 ESXTOP for Non-persistent VDI users Single Server testing Non-Kernel Mbytes
371000 \\VDI-
15\Mem
370000 ory\NonK
ernel
MBytes
369000
368000
367000
8:52:26…
8:55:49…
8:59:11…
9:02:33…
9:05:56…
9:09:18…
9:12:41…
9:16:05…
9:19:30…
9:22:56…
9:26:22…
9:29:50…
9:33:19…
9:36:49…
9:40:22…
9:43:56…
9:47:28…
9:50:57…
9:54:21…
9:57:43…
10:01:0…
10:04:2…
10:07:5…
172
Test Results
Figure 115 RDS cluster configured with 72 Windows 2012 R2 RDSH virtual servers hosting 1450 User Sessions
on 9 B200 M4 Servers (N+1)
173
Test Results
80
60
40 \\VDI-
20 01\Physic
al…
0
9:12:13 PM
9:18:23 PM
8:56:46 PM
8:59:51 PM
9:02:57 PM
9:06:02 PM
9:09:07 PM
9:15:18 PM
9:21:29 PM
9:24:34 PM
9:27:39 PM
9:30:45 PM
9:33:50 PM
9:36:56 PM
9:40:01 PM
9:43:07 PM
9:46:13 PM
9:49:19 PM
9:52:25 PM
9:55:31 PM
9:58:36 PM
10:01:41 PM
10:04:46 PM
10:07:52 PM
Figure 117 ESXTOP CPU Total Utilization for one of RDSH-Server on RDSH Cluster testing
\\RDSH-01\Physical Cpu(_Total)\% Util Time
100
80
60
40 \\VDI-
20 01\Physical
Cpu(_Total)
0 \% Util…
9:26:07 PM
8:56:46 PM
9:00:02 PM
9:03:17 PM
9:06:33 PM
9:09:48 PM
9:13:04 PM
9:16:20 PM
9:19:35 PM
9:22:51 PM
9:29:22 PM
9:32:38 PM
9:35:54 PM
9:39:10 PM
9:42:26 PM
9:45:42 PM
9:48:59 PM
9:52:15 PM
9:55:31 PM
9:58:46 PM
10:02:02 PM
10:05:17 PM
10:08:33 PM
Figure 118 ESXTOP Network Adapter Mbytes Received /Sent /Sec for one of RDSH Server on Cluster testing
140 \\VDI-
01\Network
Port(vSwitch0:3
120 3554434:vmnic1
)\MBits
100 Received/sec
\\VDI-
80 01\Network
Port(vSwitch0:3
3554436:vmnic0
60 )\MBits
Received/sec
40 \\VDI-
01\Network
20 Port(vSwitch0:3
3554434:vmnic1
)\MBits
0 Transmitted/sec
9:03:17 PM
9:29:22 PM
9:55:31 PM
8:56:46 PM
9:00:02 PM
9:06:33 PM
9:09:48 PM
9:13:04 PM
9:16:20 PM
9:19:35 PM
9:22:51 PM
9:26:07 PM
9:32:38 PM
9:35:54 PM
9:39:10 PM
9:42:26 PM
9:45:42 PM
9:48:59 PM
9:52:15 PM
9:58:46 PM
10:02:02 PM
10:05:17 PM
10:08:33 PM
\\VDI-
01\Network
Port(vSwitch0:3
3554436:vmnic0
)\MBits
Transmitted/sec
174
Test Results
Figure 119 ESXTOP Physical Disc Adapter Mbytes Read/Written /Sec for one of RDSH-Server on RDSH Cluster
testing
30 \\VDI-01\Physical
Disk
Adapter(vmhba0)\M
25 Bytes Read/sec
\\VDI-01\Physical
20 Disk
Adapter(vmhba1)\M
Bytes Read/sec
15 \\VDI-01\Physical
Disk
10 Adapter(vmhba2)\M
Bytes Read/sec
5 \\VDI-01\Physical
Disk
Adapter(vmhba0)\M
0 Bytes Written/sec
8:59:51 PM
9:46:13 PM
8:56:46 PM
9:02:57 PM
9:06:02 PM
9:09:07 PM
9:12:13 PM
9:15:18 PM
9:18:23 PM
9:21:29 PM
9:24:34 PM
9:27:39 PM
9:30:45 PM
9:33:50 PM
9:36:56 PM
9:40:01 PM
9:43:07 PM
9:49:19 PM
9:52:25 PM
9:55:31 PM
9:58:36 PM
10:01:41 PM
10:04:46 PM
10:07:52 PM
\\VDI-01\Physical
Disk
Adapter(vmhba1)\M
Bytes Written/sec
\\VDI-01\Physical
Disk
Adapter(vmhba2)\M
Bytes Written/sec
Figure 120 ESXTOP Non-Kernel Mbytes for one of RDSH-Server on RDSH Cluster testing
120000 \\VDI-01\Memory\NonKernel MBytes
100000
80000
\\VDI-
60000 01\Mem
ory\Non
40000 Kernel
MBytes
20000
0
9:03:38…
8:56:46…
9:00:12…
9:07:04…
9:10:30…
9:13:55…
9:17:21…
9:20:47…
9:24:13…
9:27:39…
9:31:05…
9:34:31…
9:37:58…
9:41:24…
9:44:51…
9:48:17…
9:51:44…
9:55:10…
9:58:36…
10:02:0…
10:05:2…
10:08:5…
175
Test Results
Figure 123 Read /Write Bandwidth for 1450 Users RDSH testing
176
Test Results
Figure 124 Login VSI graph for 3550 Users Linked Clones Cluster testing
Figure 125 3550 Users Linked Clones Cluster testing Login VSI Chart
Figure 126 3550 Users CPU Core Utilization for one of VDI server
\\VDI-11\Physical Cpu(_Total)\% Core Util Time
100
80
60 \\VDI-
11\Physic
40 al
Cpu(_Tot
20 al)\%
Core…
0
11:27:07 AM
10:29:01 AM
10:32:21 AM
10:35:42 AM
10:39:03 AM
10:42:24 AM
10:45:46 AM
10:49:08 AM
10:52:30 AM
10:55:54 AM
10:59:18 AM
11:02:44 AM
11:06:11 AM
11:09:38 AM
11:13:07 AM
11:16:38 AM
11:20:09 AM
11:23:40 AM
11:30:31 AM
11:33:51 AM
11:37:12 AM
11:40:32 AM
11:43:53 AM
11:47:14 AM
177
20
40
60
80
0
100
20
40
60
80
0
100
120
140
10:29:01 AM
Test Results
10:29:01 AM
10:31:48 AM
10:32:21 AM
10:34:35 AM 10:35:42 AM
10:37:22 AM 10:39:03 AM
10:40:10 AM 10:42:24 AM
10:42:58 AM
10:45:46 AM
10:45:46 AM
10:48:35 AM 10:49:08 AM
10:51:23 AM 10:52:30 AM
10:54:11 AM 10:55:54 AM
10:57:02 AM 10:59:18 AM
10:59:52 AM
11:02:44 AM
11:02:44 AM
11:05:37 AM 11:06:11 AM
11:08:29 AM 11:09:38 AM
11:11:23 AM 11:13:07 AM
11:14:18 AM
11:17:13 AM
11:16:38 AM
11:20:09 AM 11:20:09 AM
11:23:05 AM 11:23:40 AM
11:25:59 AM 11:27:07 AM
11:28:49 AM
178
11:30:31 AM
\\VDI-11\Physical Cpu(_Total)\% Util Time
11:31:38 AM
11:34:25 AM 11:33:51 AM
11:37:12 AM 11:37:12 AM
11:39:59 AM 11:40:32 AM
Figure 127 3550 Users CPU Total Utilization for one of VDI server
11:42:46 AM 11:43:53 AM
11:45:34 AM
11:48:21 AM
11:47:14 AM
cal
\\VDI-
11\Physi
Cpu(_T…
1)\MBits
0)\MBits
1)\MBits
0)\MBits
Received/sec
Received/sec
Transmitted/sec
Transmitted/sec
Port(DvsPortset-
Port(DvsPortset-
Port(DvsPortset-
Port(DvsPortset-
\\VDI-11\Network
\\VDI-11\Network
\\VDI-11\Network
\\VDI-11\Network
0:50331662:vmnic
0:50331661:vmnic
0:50331662:vmnic
0:50331661:vmnic
Figure 128 3550 Users network Mbytes received /transmitted per sec for one of VDI server
Test Results
Figure 129 3550 Users vHBA disk adapter Mbytes read /written per sec for one of VDI server
\\VDI-
25 11\Physical Disk
Adapter(vmhba0
)\MBytes
20 Read/sec
\\VDI-
11\Physical Disk
15 Adapter(vmhba1
)\MBytes
Read/sec
\\VDI-
10 11\Physical Disk
Adapter(vmhba2
)\MBytes
5 Read/sec
\\VDI-
11\Physical Disk
0 Adapter(vmhba0
)\MBytes
10:29:01 AM
10:31:59 AM
10:34:57 AM
10:37:56 AM
10:40:55 AM
10:43:54 AM
10:46:54 AM
10:49:53 AM
10:52:52 AM
10:55:54 AM
10:58:56 AM
11:01:58 AM
11:05:02 AM
11:08:06 AM
11:11:11 AM
11:14:18 AM
11:17:25 AM
11:20:33 AM
11:23:40 AM
11:26:44 AM
11:29:46 AM
11:32:44 AM
11:35:42 AM
11:38:41 AM
11:41:39 AM
11:44:38 AM
Written/sec
\\VDI-
11\Physical Disk
Adapter(vmhba1
)\MBytes
Written/sec
Figure 130 3550 Users Non-Kernel Mbytes for one VDI servers
\\VDI-11\Memory\NonKernel MBytes
360000
350000
340000
330000 \\VDI-
11\Memory\
320000 NonKernel
MBytes
310000
300000
290000
10:29:01 AM
10:32:21 AM
10:35:42 AM
10:39:03 AM
10:42:24 AM
10:45:46 AM
10:49:08 AM
10:52:30 AM
10:55:54 AM
10:59:18 AM
11:02:44 AM
11:06:11 AM
11:09:38 AM
11:13:07 AM
11:16:38 AM
11:20:09 AM
11:23:40 AM
11:27:07 AM
11:30:31 AM
11:33:51 AM
11:37:12 AM
11:40:32 AM
11:43:53 AM
11:47:14 AM
179
Test Results
Figure 132 Read / Write IOPs for 3550 Users Linked Clone VDI testing
Figure 133 Read /Write bandwidth for 3550 Users Liked clones Testing
180
Test Results
5000 Users Mixed Horizon RDSH and Linked Clone Workload Testing
1450 RDS Hosted Server Sessions along with 3550 Non-Persistent VDI-Linked Clone Virtual Machine
Tested Concurrently as 5000 Users Combined Mixed Workload
Figure 134 5000 Users Mixed Work Load Login VSI Chart
181
20
40
60
80
0
100
20
40
60
80
0
100
120
140
160
20
40
60
80
0
100
11:37:37 AM 11:37:37… 11:37:37 AM
Test Results
182
\\VDI-11\Physical Cpu(_Total)\% Util Time
12:54:01 PM
\\VDI-11\Physical Cpu(_Total)\% Core Util Time
12:53:16 PM 12:54:01…
12:57:45… 12:57:45 PM
12:57:22 PM 1:01:28 PM
1:01:28…
1:01:28 PM 1:05:12 PM
1:05:12…
1:05:35 PM 1:08:56 PM
1:08:56…
1:09:41 PM
Time
\\VDI-
\\VDI-
\\VDI-
\\VDI-
\\VDI-
Figure 137 5000 Users Mixed workload CPU Total Util Time for VDI-Server
Figure 136 5000 Users Mixed workload CPU Core Util Time for VDI-Server
\\VDI-
c1)\MBits
c0)\MBits
c1)\MBits
c0)\MBits
11\Network
11\Network
11\Network
11\Network
% Core Util
Received/sec
Received/sec
11\Physical
11\Physical
Cpu(_Total)\
Transmitted/sec
Transmitted/sec
Port(DvsPortset-
Port(DvsPortset-
Port(DvsPortset-
Port(DvsPortset-
0:50331662:vmni
0:50331661:vmni
0:50331662:vmni
0:50331661:vmni
Cpu(_Total)\%…
Figure 138 5000 Users Mixed workload Network Received /Transmitted sec for VDI-Server
Test Results
Figure 139 5000 Users Mixed workload Disk Adapter Mbytes for VDI-Server
Disk Adapter
35 \\VDI-11\Physical
Disk
30 Adapter(vmhba0)\
MBytes Read/sec
25
\\VDI-11\Physical
20 Disk
Adapter(vmhba1)\
15 MBytes Read/sec
10 \\VDI-11\Physical
Disk
5 Adapter(vmhba2)\
MBytes Read/sec
0
11:46:1…
11:37:3…
11:41:5…
11:50:2…
11:54:4…
11:59:0…
12:03:2…
12:07:4…
12:12:0…
12:16:3…
12:21:0…
12:25:3…
12:30:0…
12:34:3…
12:39:0…
12:43:3…
12:48:0…
12:52:2…
12:56:3…
1:00:55…
1:05:12…
1:09:29…
\\VDI-11\Physical
Disk
Adapter(vmhba0)\
MBytes
Written/sec
Figure 140 5000 Users Mixed workload Non-Kernel Memory Mbytes for VDI Server
\\VDI-11\Memory\NonKernel MBytes
350000
345000
340000
335000 \\VDI-
11\Memor
330000 y\NonKern
325000 el MBytes
320000
315000
11:37:3…
11:41:4…
11:45:4…
11:49:5…
11:54:0…
11:58:0…
12:02:1…
12:06:2…
12:10:3…
12:14:5…
12:19:0…
12:23:2…
12:27:4…
12:32:0…
12:36:2…
12:40:4…
12:44:5…
12:49:1…
12:53:1…
12:57:2…
1:01:28…
1:05:35…
1:09:41…
183
Test Results
Figure 142 Read / Write IOPs for 5000 Users Mixed Workload
Figure 143 Read / write Bandwidth for 5000 Users Mixed Workload
184
Test Results
Perfmon Charts for 5000 Users Mixed Workload Test for VMware View Connection
Server
80
60
40
\\VHVIEWCS-
20 6\Processor(…
1:00:17…
1:04:17…
11:44:17…
11:48:17…
11:52:17…
11:56:17…
12:00:17…
12:04:17…
12:08:17…
12:12:17…
12:16:17…
12:20:17…
12:24:17…
12:28:17…
12:32:17…
12:36:17…
12:40:17…
12:44:17…
12:48:17…
12:52:17…
12:56:17…
Figure 145 5000 Users Processor time View Connection Server Perfmon
\\VHVIEWCS-6\Processor(_Total)\% Processor Time
100
80
60
40
20 \\VHVIEWCS-
6\Processor(_
0 Total)\%
1:01:17 PM
1:04:57 PM
11:51:37 AM
12:20:57 PM
11:44:17 AM
11:47:57 AM
11:55:17 AM
11:58:57 AM
12:02:37 PM
12:06:17 PM
12:09:57 PM
12:13:37 PM
12:17:17 PM
12:24:37 PM
12:28:17 PM
12:31:57 PM
12:35:37 PM
12:39:17 PM
12:42:57 PM
12:46:37 PM
12:50:17 PM
12:53:57 PM
12:57:37 PM
Processor…
Figure 146 5000 Users Available Mbytes View Connection Server Perfmon
12:28:0…
12:32:0…
12:36:0…
12:40:0…
12:44:0…
12:48:0…
12:52:0…
12:56:0…
1:00:07…
1:04:07…
185
Test Results
Figure 147 5000 Users View Connection Server Network Bytes Received /sec Perfmon
\\VHVIEWCS-6\Network Interface(vmxnet3 Ethernet Adapter)\Bytes
80000000
Received/sec
60000000
\\VHVIEW
40000000 CS-
6\Network
20000000 Interface(v
mxnet3…
0
12:32:1…
11:44:1…
11:48:1…
11:52:1…
11:56:1…
12:00:1…
12:04:1…
12:08:1…
12:12:1…
12:16:1…
12:20:1…
12:24:1…
12:28:1…
12:36:1…
12:40:1…
12:44:1…
12:48:1…
12:52:1…
12:56:1…
1:00:17…
1:04:17…
Figure 148 5000 Users View Connection Server Network Bytes Sent /sec
40000000 \\VHVI
EWCS-
6\Net
20000000 work
Inter…
0
1:01:17…
1:04:57…
11:44:17…
11:47:57…
11:51:37…
11:55:17…
11:58:57…
12:02:37…
12:06:17…
12:09:57…
12:13:37…
12:17:17…
12:20:57…
12:24:37…
12:28:17…
12:31:57…
12:35:37…
12:39:17…
12:42:57…
12:46:37…
12:50:17…
12:53:57…
12:57:37…
Pure Storage FlashArray//m50 Test Results for Full Scale, Mixed Workload
Resiliency Testing
For the final simulation, we elected to highlight the tremendous resiliency of the Pure Storage FlashArray//m50
by performing an upgrade of the Purity Operating Environment in parallel with repeating the previous 5000
desktop mixed-workload simulation. In this test, we upgraded from Purity v4.5.5 to Purity v4.6.8 in the middle
of the Login VSI simulation while desktops were still logging in to the environment and running the Login VSI
Knowledge Worker workload.
Worth noting is that this exact procedure would also be followed during a controller upgrade as part of the
Evergreen Storage business model where customers receive the latest Pure Storage controllers for free every
three years so long as a valid maintenance agreement is in place. Another supported scenario would be non-
disruptively upgrading to the more performant FlashArray//m70 following this same procedure. Pure Storage
controllers are stateless and do not require any additional setup other than inserting them into the chassis and
connecting cables. Relatedly, capacity expansions (adding a shelf or denser SSD drives) are also
accomplished without any downtime in the middle of production operations.
Upgrading the Purity Operating Environment is typically handled by Pure Storage support for arrays that are
managed via Cloud Assist. Upgrades can also be accomplished locally by KVM or console connection to the
array as well by a Pure Storage or other authorized support customer. To accomplish the upgrade, the process
is extremely simple, usually takes less than 20 minutes and achieved with the following steps:
186
Test Results
5. Reboot primary controller, array non-disruptively fails over to upgraded secondary controller which be-
comes primary and new version of Purity OE is in use
The following charts show the array behavior during the Purity code upgrade. Immediately worth noting is that
the drop in reported metrics to 0 during the upgrade process was a bug in the Purity code that has since been
correct in version 4.7.0 (for this test we upgraded from Purity 4.5.5 to Purity 4.6.8). The Login VSI results that
will be shown in the section provide independent verification that the upgrade was non-disruptive and non-
impactful to the end-users running in the environment.
The chart below shows the latency of the array during the Purity upgrade process window shown by the
orange box. Overall the upgrade process took approximately 14 minutes to complete including controller
reboots.
Next we show IOPs being served during the 5000 mixed workload simulation with the Purity upgrade.
187
Test Results
Lastly, this chart shows our bandwidth during the mixed workload + Pure Storage FlashArray//m50 upgrade
process.
Login VSI provides impartial verification that the Purity upgrade was non-disruptive. You can see from the
below performance summary chart from the tool that no end-user performance degradation or large number
188
Test Results
of disconnected sessions occurred while the Pure Storage array was upgraded.
Lastly, you can see from the Login VSI performance charts that relative performance from the 5000 Mixed
Workload simulation and the 5000 Mixed Workload simulation that included the Purity upgrade are essentially
identical with only brief spikes in response time during the upgrade process that likely would not be noticeable
by the end-user:
The ability of Pure Storage to provide non-disruptive operations such as the above example cannot be
understated. No longer do storage administrators need to schedule downtime with the VMware Horizon team
for maintenance operations and as we have clearly demonstrated above – maintenance operations can now
be accomplished in the middle of the workday without causing any interruption to a huge amount of active
VMware Horizon users.
189
Summary
Summary
FlashStack delivers a platform for Enterprise VDI deployments and cloud datacenters using Cisco UCS Blade
and Rack Servers, Cisco Fabric Interconnects, Cisco Nexus 9000 switches, Cisco MDS switches and fibre
channel-attached Pure Storage FlashArray//m. FlashStack is designed and validated using compute, net-
work and storage best practices and high availability to reduce deployment time, project risk and IT costs
while maintaining scalability and flexibility for addressing a multitude of IT initiatives. This CVD validates the
design, performance, management, scalability, and resilience that FlashStack provides to customers wish-
ing to deploy enterprise-class VDI for 5000 users at a time.
Deploy and transition services to implement validated architectures and prepare your storage environment
Operations services to deliver continuous operations while driving operational excellence and efficiency.
In addition, Pure provides in-depth knowledge transfer and education services that give you access to our
global technical resources and intellectual property.
190
About the Authors
Ramesh Guduru is a Technical Marketing Engineer at Cisco Systems with over 12 years of experience in the
field of information technology. Ramesh’s focus primarily on design and implementation of Virtual Desktop
Infrastructure projects for data center, VMware View thin client administration, configuration and optimization of
virtual desktop environment, Data center reference architectures, Cisco Unified Computing System and
Storage design and implementation. Ramesh’s skill set include core VMware applications in the virtual
environment focusing in system design and implementation of virtualization components, solutions validation,
technical content creation and testing/benchmarking.
Acknowledgements
For their support and contribution to the design, validation, and creation of this Cisco Validated Design, we
would like to acknowledge the significant contribution and expertise that resulted in developing this document:
Kyle Grossmiller, Solutions Architect, Customer Solutions Group, Pure Storage, Inc.
Mike Brennan, Manager, Technical Marketing, Desktop Virtualization Solutions Team, Computing Systems
Product Group, Cisco Systems, Inc.
191
References
References
This section provides links to additional information for each partner’s solution component of this document.
http://www.cisco.com/c/en/us/products/servers-unified-computing/ucs-b-series-blade-
servers/index.html
http://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/release/notes/ucs_2_2_rn.html
http://www.cisco.com/c/en/us/products/collateral/interfaces-modules/ucs-virtual-interface-card-
1340/datasheet-c78-732517.html
http://www.cisco.com/c/en/us/products/interfaces-modules/ucs-virtual-interface-card-1340/index.html
http://www.cisco.com/c/en/us/products/switches/nexus-9372px-switch/index.html
http://www.cisco.com/c/en/us/products/switches/nexus-1000v-switch-vmware-vsphere/index.html
http://www.cisco.com/c/en/us/support/storage-networking/mds-9148s-16g-multilayer-fabric-
switch/model.html
http://www.cisco.com/c/en/us/products/collateral/storage-networking/mds-9148-multilayer-fabric-
switch/data_sheet_c78-571411.html
http://www.cisco.com/c/en/us/products/collateral/storage-networking/mds-9148s-16g-multilayer-
fabric-switch/datasheet-c78-731523.html
VMware References
https://www.vmware.com/support/pubs/vsphere-esxi-vcenter-server-6-pubs.html
http://pubs.vmware.com/Release_Notes/en/horizon-6-view/horizon-62-view-release-notes.html
https://pubs.vmware.com/horizon-62-view/index.jsp
192
References
https://pubs.vmware.com/horizon-62-view/index.jsp?topic=%2Fcom.vmware.horizon-
view.desktops.doc%2FGUID-DFAD071A-7F60-4720-86AB-8F1597BFC95C.html
http://pubs.vmware.com/vsphere-60/index.jsp
http://www.vmware.com/files/pdf/view/vmware-horizon-view-best-practices-performance-study.pdf
https://www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf
https://labs.vmware.com/flings/vmware-os-optimization-tool
Microsoft References
https://technet.microsoft.com/en-us/library/hh831620(v=ws.11).aspx
https://technet.microsoft.com/en-us/library/dn281793(v=ws.11).aspx
https://support.microsoft.com/en-us/kb/2833839
https://technet.microsoft.com/en-us/library/hh831447(v=ws.11).aspx
https://www.loginvsi.com/documentation/Start_your_first_test
http://www.purestorage.com/content/dam/purestorage/pdf/datasheets/PureStorage_FlashArraym-
Brochure.pdf
http://www.purestorage.com/solutions/infrastructure/flashstack.html
http://www.purestorage.com/resources/type-a/WP-
PureStorageandVMwarevSphereBestPracticesGuide_Request.html
http://www.purestorage.com/resources/type-b/WP-PureStorageandVAAI_Request.html
http://www.purestorage.com/resources/type-a/WP-FlashStackRefArch-VSI_Request.html
193
References
http://info.purestorage.com/WP-FlashStackRefArch-VDI_Request.html
http://www.purestorage.com/content/dam/purestorage/pdf/datasheets/ESG_Lab_Validation_Summar
y_Pure_Storage_Sep_2015.pdf
194
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
version 6.1(2)I3(3a)
switchname N9K-A
vdc N9K-A id 1
feature telnet
feature udld
feature interface-vlan
feature hsrp
feature lacp
feature dhcp
feature vpc
feature lldp
195
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
ip domain-lookup
vlan 1,160-164,166
vlan 160
name IB_Mgmt
vlan 161
name infra
vlan 162
name VDI
vlan 164
name OOB-Mgmt
vlan 166
name vMotion
service dhcp
ip dhcp relay
196
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
vpc domain 10
peer-switch
role priority 10
peer-gateway
auto-recovery
interface Vlan1
no ip redirects
no ipv6 redirects
no ip redirects
ip address 10.10.160.2/24
hsrp version 2
hsrp 11
preempt
priority 90
ip 10.10.160.1
no shutdown
interface Vlan161
no ip redirects
ip address 10.10.161.3/24
no ipv6 redirects
hsrp version 2
hsrp 12
preempt
priority 80
ip 10.10.161.1
no shutdown
197
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Vlan162
no ip redirects
ip address 10.10.192.2/19
hsrp version 2
hsrp 13
preempt
priority 82
ip 10.10.192.1
no shutdown
interface Vlan164
no ip redirects
ip address 10.29.164.3/24
no ipv6 redirects
hsrp version 2
hsrp 14
preempt
priority 80
ip 10.29.164.1
no shutdown
interface Vlan166
no ip redirects
ip address 10.29.166.3/24
no ipv6 redirects
hsrp version 2
hsrp 14
preempt
priority 80
ip 10.29.166.1
198
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
no shutdown
vpc 29
interface port-channel31
vpc 31
interface Ethernet1/1
interface Ethernet1/2
interface Ethernet1/3
interface Ethernet1/4
interface Ethernet1/5
interface Ethernet1/6
interface Ethernet1/7
interface Ethernet1/8
interface Ethernet1/9
interface Ethernet1/10
interface Ethernet1/11
interface Ethernet1/12
interface Ethernet1/13
interface Ethernet1/14
interface Ethernet1/15
interface Ethernet1/16
interface Ethernet1/17
interface Ethernet1/18
interface Ethernet1/19
speed 1000
interface Ethernet1/20
199
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Ethernet1/21
interface Ethernet1/22
interface Ethernet1/23
interface Ethernet1/24
interface Ethernet1/25
interface Ethernet1/26
interface Ethernet1/27
interface Ethernet1/28
interface Ethernet1/29
interface Ethernet1/30
interface Ethernet1/31
interface Ethernet1/32
interface Ethernet1/33
interface Ethernet1/34
interface Ethernet1/35
interface Ethernet1/36
interface Ethernet1/37
interface Ethernet1/38
200
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Ethernet1/39
interface Ethernet1/40
interface Ethernet1/41
interface Ethernet1/42
interface Ethernet1/43
interface Ethernet1/44
interface Ethernet1/45
interface Ethernet1/46
interface Ethernet1/47
interface Ethernet1/48
interface Ethernet1/49
interface Ethernet1/50
interface Ethernet1/51
interface Ethernet1/52
interface Ethernet1/53
interface Ethernet1/54
interface mgmt0
ip address 10.29.164.11/24
line console
line vty
N9K-A#
201
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
version 6.1(2)I3(3a)
switchname N9K-B
vdc N9K-B id 1
feature telnet
feature udld
feature interface-vlan
feature hsrp
feature lacp
feature dhcp
feature vpc
feature lldp
ip domain-lookup
202
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
vlan 1,160-164,166
vlan 160
vlan 161
vlan 162
vlan 164
vlan 166
service dhcp
ip dhcp relay
vpc domain 10
role priority 20
peer-gateway
auto-recovery
203
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Vlan1
no ip redirects
no ipv6 redirects
interface Vlan160
no ip redirects
ip address 10.10.160.3/24
hsrp version 2
hsrp 11
preempt
priority 80
ip 10.10.160.1
no shutdown
interface Vlan161
no ip redirects
ip address 10.10.161.2/24
no ipv6 redirects
hsrp version 2
hsrp 12
preempt
priority 90
ip 10.10.161.1
no shutdown
interface Vlan162
no ip redirects
ip address 10.10.192.3/19
no ipv6 redirects
hsrp version 2
hsrp 13
preempt
204
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
priority 92
ip 10.10.192.1
hsrp 192
preempt
priority 121
no shutdown
interface Vlan164
no ip redirects
ip address 10.29.164.2/24
no ipv6 redirects
hsrp version 2
hsrp 14
preempt
priority 90
ip 10.29.164.1
no shutdown
interface port-channel5
interface port-channel10
vpc peer-link
interface port-channel29
vpc 29
205
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface port-channel31
vpc 31
interface Ethernet1/1
no shutdown
interface Ethernet1/2
no shutdown
interface Ethernet1/3
no shutdown
interface Ethernet1/4
no shutdown
interface Ethernet1/5
no shutdown
interface Ethernet1/6
no shutdown
interface Ethernet1/7
no shutdown
interface Ethernet1/8
no shutdown
interface Ethernet1/9
no shutdown
interface Ethernet1/10
no shutdown
interface Ethernet1/11
no shutdown
interface Ethernet1/12
no shutdown
interface Ethernet1/13
no shutdown
interface Ethernet1/14
206
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
no shutdown
interface Ethernet1/15
no shutdown
interface Ethernet1/16
no shutdown
interface Ethernet1/17
no shutdown
interface Ethernet1/18
no shutdown
interface Ethernet1/19
speed 1000
no shutdown
interface Ethernet1/20
no shutdown
interface Ethernet1/21
no shutdown
interface Ethernet1/22
no shutdown
interface Ethernet1/23
no shutdown
interface Ethernet1/24
no shutdown
interface Ethernet1/25
no shutdown
interface Ethernet1/26
no shutdown
interface Ethernet1/27
no shutdown
207
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Ethernet1/28
no shutdown
interface Ethernet1/29
no shutdown
interface Ethernet1/30
no shutdown
interface Ethernet1/31
no shutdown
interface Ethernet1/32
no shutdown
interface Ethernet1/33
no shutdown
interface Ethernet1/34
no shutdown
interface Ethernet1/35
no shutdown
interface Ethernet1/36
no shutdown
interface Ethernet1/37
no shutdown
208
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Ethernet1/38
no shutdown
interface Ethernet1/39
no shutdown
interface Ethernet1/40
no shutdown
interface Ethernet1/41
no shutdown
interface Ethernet1/42
no shutdown
interface Ethernet1/43
no shutdown
interface Ethernet1/44
no shutdown
interface Ethernet1/45
no shutdown
interface Ethernet1/46
no shutdown
interface Ethernet1/47
no shutdown
interface Ethernet1/48
no shutdown
interface Ethernet1/49
no shutdown
209
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface Ethernet1/50
no shutdown
interface Ethernet1/51
no shutdown
interface Ethernet1/52
no shutdown
interface Ethernet1/53
no shutdown
interface Ethernet1/54
no shutdown
interface mgmt0
ip address 10.29.164.12/24
line console
line vty
N9K-B#
version 6.2(9a)
feature npiv
feature fport-channel-trunk
210
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
ip domain-lookup
vsan database
vsan 20
device-alias database
211
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
212
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface port-channel1
vsan database
switchname MDS-A
line console
line vty
interface fc1/1
interface fc1/2
interface fc1/11
interface fc1/12
interface fc1/3
interface fc1/4
interface fc1/5
interface fc1/6
interface fc1/7
interface fc1/8
interface fc1/9
interface fc1/10
213
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/13
interface fc1/14
interface fc1/15
interface fc1/16
interface fc1/17
interface fc1/18
interface fc1/19
interface fc1/20
interface fc1/21
interface fc1/22
interface fc1/23
interface fc1/24
interface fc1/25
interface fc1/26
interface fc1/27
interface fc1/28
interface fc1/29
interface fc1/30
interface fc1/31
interface fc1/32
interface fc1/33
interface fc1/34
interface fc1/35
interface fc1/36
interface fc1/37
interface fc1/38
interface fc1/39
interface fc1/40
interface fc1/41
interface fc1/42
interface fc1/43
interface fc1/44
214
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/45
interface fc1/46
interface fc1/47
interface fc1/48
interface fc1/1
interface fc1/2
interface fc1/11
interface fc1/12
215
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
216
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
217
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
218
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
member Infra-serv1-fc0
member Infra-serv2-fc0
member pure-vdi-serv4-fc0
member pure-vdi-serv1-fc0
member pure-vdi-serv2-fc0
member pure-vdi-serv3-fc0
member pure-vdi-serv5-fc0
member pure-vdi-serv6-fc0
member pure-vdi-serv8-fc0
member pure-vdi-serv7-fc0
member pure-vdi-serv9-fc0
member pure-vdi-serv10-fc0
member pure-vdi-serv11-fc0
member pure-vdi-serv12-fc0
member pure-vdi-serv13-fc0
member pure-vdi-serv16-fc0
member pure-vdi-serv17-fc0
member pure-vdi-serv18-fc0
member pure-vdi-serv19-fc0
member pure-vdi-serv20-fc0
member pure-vdi-serv21-fc0
member pure-vdi-serv24-fc0
member pure-vdi-serv25-fc0
member pure-vdi-serv26-fc0
member pure-vdi-serv27-fc0
member pure-vdi-serv28-fc0
219
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
member pure-vdi-serv29-fc0
member pure-vdi-serv30-fc0
member pure-vdi-serv15-fc0
member pure-vdi-serv14-fc0
member pure-vdi-serv23-fc0
member pure-vdi-serv22-fc0
interface fc1/1
port-license acquire
no shutdown
interface fc1/2
port-license acquire
no shutdown
interface fc1/3
port-license acquire
no shutdown
interface fc1/4
port-license acquire
no shutdown
interface fc1/5
port-license acquire
interface fc1/6
port-license acquire
interface fc1/7
no port-license
no shutdown
220
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/8
no port-license
no shutdown
interface fc1/9
port-license acquire
no shutdown
interface fc1/10
port-license acquire
no shutdown
interface fc1/11
port-license acquire
channel-group 1 force
no shutdown
interface fc1/12
port-license acquire
channel-group 1 force
no shutdown
interface fc1/13
port-license acquire
no shutdown
interface fc1/14
port-license acquire
no shutdown
interface fc1/15
port-license acquire
no shutdown
interface fc1/16
port-license acquire
no shutdown
interface fc1/17
port-license acquire
interface fc1/18
221
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
port-license acquire
interface fc1/19
port-license acquire
interface fc1/20
port-license acquire
interface fc1/21
port-license acquire
interface fc1/22
port-license acquire
interface fc1/23
port-license acquire
interface fc1/24
port-license acquire
interface fc1/25
port-license acquire
interface fc1/26
port-license acquire
interface fc1/27
port-license acquire
interface fc1/28
port-license acquire
interface fc1/29
port-license acquire
interface fc1/30
port-license acquire
interface fc1/31
port-license acquire
interface fc1/32
interface fc1/33
port-license acquire
interface fc1/34
port-license acquire
222
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/35
port-license acquire
interface fc1/36
port-license acquire
interface fc1/37
port-license acquire
interface fc1/38
port-license acquire
interface fc1/39
port-license acquire
interface fc1/40
port-license acquire
interface fc1/41
port-license acquire
interface fc1/42
port-license acquire
interface fc1/43
port-license acquire
interface fc1/44
port-license acquire
interface fc1/45
port-license acquire
interface fc1/46
port-license acquire
interface fc1/47
port-license acquire
interface fc1/48
port-license acquire
interface mgmt0
ip default-gateway 10.29.164.1
MDS-A#
223
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
version 6.2(9a)
feature npiv
feature fport-channel-trunk
no password strength-check
ip domain-lookup
vsan database
vsan 30
device-alias database
device-alias commit
224
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
225
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface port-channel1
vsan database
226
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
switchname MDS-B
line console
line vty
interface fc1/1
interface fc1/2
interface fc1/11
interface fc1/12
interface fc1/3
interface fc1/4
interface fc1/5
interface fc1/6
interface fc1/7
interface fc1/8
interface fc1/9
interface fc1/10
interface fc1/13
interface fc1/14
interface fc1/15
interface fc1/16
interface fc1/17
interface fc1/18
interface fc1/19
interface fc1/20
interface fc1/21
interface fc1/22
interface fc1/23
interface fc1/24
interface fc1/25
interface fc1/26
227
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/27
interface fc1/28
interface fc1/29
interface fc1/30
interface fc1/31
interface fc1/32
interface fc1/33
interface fc1/34
interface fc1/35
interface fc1/36
interface fc1/37
interface fc1/38
interface fc1/39
interface fc1/40
interface fc1/41
interface fc1/42
interface fc1/43
interface fc1/44
interface fc1/45
interface fc1/46
interface fc1/47
interface fc1/48
interface fc1/1
interface fc1/2
interface fc1/11
228
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
229
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
230
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
231
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
member Infra-serv1-fc1
member Infra-serv2-fc1
member Infra-serv3-fc1
member pure-vdi-serv1-fc1
member pure-vdi-serv2-fc1
232
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
member pure-vdi-serv3-fc1
member pure-vdi-serv4-fc1
member pure-vdi-serv5-fc1
member pure-vdi-serv6-fc1
member pure-vdi-serv7-fc1
member pure-vdi-serv8-fc1
member pure-vdi-serv9-fc1
member pure-vdi-serv10-fc1
member pure-vdi-serv11-fc1
member pure-vdi-serv12-fc1
member pure-vdi-serv13-fc1
member pure-vdi-serv16-fc1
member pure-vdi-serv17-fc1
member pure-vdi-serv18-fc1
member pure-vdi-serv19-fc1
member pure-vdi-serv20-fc1
member pure-vdi-serv21-fc1
member pure-vdi-serv24-fc1
member pure-vdi-serv25-fc1
member pure-vdi-serv26-fc1
member pure-vdi-serv27-fc1
member pure-vdi-serv28-fc1
member pure-vdi-serv29-fc1
member pure-vdi-serv30-fc1
member pure-vdi-serv15-fc1
member pure-vdi-serv23-fc1
member pure-vdi-serv22-fc1
233
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/1
port-license acquire
no shutdown
interface fc1/2
port-license acquire
no shutdown
interface fc1/3
port-license acquire
no shutdown
interface fc1/4
port-license acquire
no shutdown
interface fc1/5
234
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
port-license acquire
interface fc1/6
port-license acquire
interface fc1/7
port-license acquire
no shutdown
interface fc1/8
port-license acquire
no shutdown
interface fc1/9
interface fc1/10
port-license acquire
no shutdown
interface fc1/11
port-license acquire
channel-group 1 force
no shutdown
interface fc1/12
port-license acquire
channel-group 1 force
no shutdown
interface fc1/13
port-license acquire
interface fc1/14
port-license acquire
no shutdown
interface fc1/15
port-license acquire
no shutdown
235
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
interface fc1/16
port-license acquire
no shutdown
interface fc1/17
port-license acquire
interface fc1/18
port-license acquire
interface fc1/19
port-license acquire
interface fc1/20
port-license acquire
interface fc1/21
port-license acquire
interface fc1/22
port-license acquire
interface fc1/23
port-license acquire
interface fc1/24
port-license acquire
interface fc1/25
port-license acquire
interface fc1/26
port-license acquire
interface fc1/27
port-license acquire
interface fc1/28
port-license acquire
interface fc1/29
port-license acquire
interface fc1/30
port-license acquire
interface fc1/31
236
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
port-license acquire
interface fc1/32
port-license acquire
interface fc1/33
port-license acquire
interface fc1/34
port-license acquire
interface fc1/35
port-license acquire
interface fc1/36
port-license acquire
interface fc1/37
port-license acquire
interface fc1/38
port-license acquire
interface fc1/39
port-license acquire
interface fc1/40
port-license acquire
interface fc1/41
port-license acquire
interface fc1/42
port-license acquire
interface fc1/43
port-license acquire
interface fc1/44
port-license acquire
interface fc1/45
port-license acquire
interface fc1/46
port-license acquire
interface fc1/47
237
Appendix A – Cisco Nexus Ethernet and MDS Fibre Channel Switch Configurations
port-license acquire
interface fc1/48
port-license acquire
interface mgmt0
ip default-gateway 10.29.164.1
MDS-B#
238
Appendix B - Pure Storage Configuration and Scripts
Future updates to this script as well as other useful PowerShell scripts built for use with Pure Storage can be
found here:
https://github.com/codyhosterman/powercli
Below is the ESXi Best Practice PowerShell script used in this Cisco Validated Design:
#Enter the following required parameters. Log folder directory is just an example, change as needed.
#**********************************
$vcenter = ""
$vcuser = ""
$vcpass = ""
$logfolder = "C:\folder\folder\etc\"
#**********************************
<#
Optional parameters. Keep these values at default unless necessary and understood
For a different IO Operations limit beside the Pure Storage recommended value of 1, change $iopsvalue
to another integer value 1-1000.
To skip changing host-wide settings for XCOPY Transfer Size and In-Guest UNMAP change $hostwideset-
tings to $false
#>
$iopsvalue = 1
$hostwidesettings = $true
<#
*******Disclaimer:******************************************************
This scripts are offered "as is" with no warranty. While this
this script in a test lab before using in a production environment. Everyone can
use the scripts/commands provided here without any written permission but I
************************************************************************
239
Appendix B - Pure Storage Configuration and Scripts
-Create a SATP rule for Round Robin and IO Operations Limit of 1 only for FlashArrays
-Configure any existing devices properly (Pure Storage FlashArray devices only)
This can be run directly from PowerCLI or from a standard PowerShell prompt. PowerCLI must be in-
stalled on the local host regardless.
Supports:
#>
$logfile = $logfolder + (Get-Date -Format o |ForEach-Object {$_ -Replace ":", "."}) + "setbestprac-
tices.txt"
write-host "Checking and setting Pure Storage FlashArray Best Practices for VMware on the ESXi hosts
in this vCenter. No further information is printed to the screen."
240
Appendix B - Pure Storage Configuration and Scripts
add-content $logfile 'Pure Storage FlashArray VMware ESXi Best Practices Script v3.1'
write-host "This version of PowerCLI is too old, version 6.3 Release 1 or later is required
(Build 3737840)" -BackgroundColor Red
write-host (Get-PowerCLIVersion).build
add-content $logfile "This version of PowerCLI is too old, version 6.3 Release 1 or later is
required (Build 3737840)"
return
try
241
Appendix B - Pure Storage Configuration and Scripts
connect-viserver -Server $vcenter -username $vcuser -password $vcpass -ErrorAction Stop |out-
null
catch
write-host $Error
return
$hosts= get-vmhost
add-content $logfile "Checking host-wide setting for XCOPY and In-Guest UNMAP"
242
Appendix B - Pure Storage Configuration and Scripts
add-content $logfile "The VAAI XCOPY MaxHWTransferSize for this host is incorrect:"
add-content $logfile "This should be set to 16386 (16 MB). Changing to 16384..."
add-content $logfile "The VAAI XCOPY MaxHWTransferSize for this host is now 16 MB"
else
add-content $logfile "The VAAI XCOPY MaxHWTransferSize for this host is correct at 16
MB and will not be altered."
if ($enableblockdelete.Value -eq 0)
else
add-content $logfile "EnableBlockDelete for this host is correctly enabled and will
not be altered."
else
add-content $logfile "The current host is not version 6.0. Skipping EnableBlockDelete
check."
else
add-content $logfile "Not checking host wide settings for XCOPY and In-Guest UNMAP due to in-
script override"
243
Appendix B - Pure Storage Configuration and Scripts
$correctrule = 0
if ($rules.Count -ge 1)
add-content $logfile "Found the following existing Pure Storage SATP rules"
$issuecount = 0
add-content $logfile "The existing Pure Storage FlashArray rule is configured with the
correct Path Selection Policy:"
else
add-content $logfile "The existing Pure Storage FlashArray rule is NOT configured with
the correct Path Selection Policy:"
$issuecount = 1
add-content $logfile "The existing Pure Storage FlashArray rule is configured with the
correct IO Operations Limit:"
244
Appendix B - Pure Storage Configuration and Scripts
else
add-content $logfile "The existing Pure Storage FlashArray rule is NOT configured with
the correct IO Operations Limit:"
$issuecount = $issuecount + 1
add-content $logfile "The existing Pure Storage FlashArray rule is configured with the
correct model:"
else
add-content $logfile "The existing Pure Storage FlashArray rule is NOT configured with
the correct model:"
add-content $logfile "The rule should be configured with the model of FlashArray"
$issuecount = $issuecount + 1
if ($issuecount -ge 1)
$satpArgs = $esxcli.storage.nmp.satp.rule.remove.createArgs()
$satpArgs.model = $rule.Model
$satpArgs.vendor = "PURE"
$satpArgs.satp = $rule.Name
$satpArgs.psp = $rule.DefaultPSP
$satpArgs.pspoption = $rule.PSPOptions
$esxcli.storage.nmp.satp.rule.remove.invoke($satpArgs)
else
245
Appendix B - Pure Storage Configuration and Scripts
$correctrule = 1
if ($correctrule -eq 0)
add-content $logfile "No correct SATP rule for the Pure Storage FlashArray is found. Creating
a new rule to set Round Robin and an IO Operations Limit of $iopsvalue"
$satpArgs = $esxcli.storage.nmp.satp.rule.remove.createArgs()
$satpArgs.model = "FlashArray"
$satpArgs.vendor = "PURE"
$satpArgs.satp = "VMW_SATP_ALUA"
$satpArgs.psp = "VMW_PSP_RR"
$satpArgs.pspoption = $iopsoption
$result = $esxcli.storage.nmp.satp.rule.add.invoke($satpArgs)
else
add-content $logfile "ERROR: The rule failed to create. Manual intervention might be
required."
else
add-content $logfile "A correct SATP rule for the FlashArray exists. No need to create a new
one on this host."
246
Appendix B - Pure Storage Configuration and Scripts
if ($devices.count -ge 1)
add-content $logfile "Looking for existing Pure Storage volumes on this host"
add-content $logfile "Found the following number of existing Pure Storage volumes on this
host."
add-content $logfile "This device does not have the correct Path Selection Policy, it
is set to:"
else
add-content $logfile "This device's Path Selection Policy is correctly set to Round
Robin. No need to change."
$deviceargs = $esxcli.storage.nmp.psp.roundrobin.deviceconfig.get.createargs()
$deviceargs.device = $device.CanonicalName
$deviceconfig = $esxcli.storage.nmp.psp.roundrobin.deviceconfig.get.invoke($de-
viceargs)
$nmpargs = $esxcli.storage.nmp.psp.roundrobin.deviceconfig.set.createargs()
$nmpargs.iops = $iopsvalue
$nmpargs.type = "iops"
247
Appendix B - Pure Storage Configuration and Scripts
add-content $logfile "The current IO Operation limit for this device is:"
add-content $logfile "This device's IO Operation Limit is unset or is not set to the
value of $iopsvalue. Changing..."
$nmpargs.device = $device.CanonicalName
$esxcli.storage.nmp.psp.roundrobin.deviceconfig.set.invoke($nmpargs) |out-null
else
add-content $logfile "This device's IO Operation Limit matches the value of $iop-
svalue. No need to change."
else
add-content $logfile "No existing Pure Storage volumes found on this host."
248
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
From a storage perspective, it is critical to maintain a latency of near to or less than a millisecond in order to
guarantee a good end-user experience. As we will see, Pure Storage delivers that level of performance
despite driving a substantial amount of IOPs and bandwidth for the thousands of desktops hosted on the single
FlashArray//m50.
The following charts were compiled from extracting front-end array telemetry data from the storage logs and
are equivalent to values shown in the Pure GUI. Results were plotted in this format to highlight individual
storage performance metrics of interest during each simulation as well as clearly show the various phases of
each simulation. Please note that across the top of each graph we have identified and broken up the Login
VSI simulation into the three separate phases of the simulation run. The first phase (green arrows and text box)
is the 2880 second Login VSI simulation phase when all sessions are ramping up and logging in. Next, the all
sessions in the simulation steady-states for 600 seconds which is denoted by the yellow arrows and text box,
and finally the black arrow to the right shows the end of the simulation when users begin logging out of the
environment. For brevity, we generally did not show the entire logout operation as array activity is minimal
during that time and the Login VSI simulation had completed.
The first chart below shows the latency of the Pure Storage FlashArray//m50 during the 1450 Horizon RDSH
sessions running on top of 72 Windows 2012 R2 servers. There were three separate 1450 Horizon RDSH
simulation runs completed in total, all with very similar results. As we can see in the below chart, we
maintained latency of less than or close to one millisecond for both read and write operations throughout this
entire run, which resulted in a confirmed outstanding end-user experience for the simulated RDSH users.
249
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
The next chart shows our read and write IOPs during the selected 1450 Horizon RDSH simulation.
Finally, we can see bandwidth against the array ramping up, hitting steady-state and then dropping as
sessions logout in parallel with the 1450 Horizon RDSH sessions in the last chart from this simulation below.
250
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
Below we see the utilization of the Horizon RDSH datastore after the Login VSI simulation. Worth noting is that
the data reduction numbers are artificially low since Login VSI writes primarily non-reducible data to the
desktop.
Pure Storage FlashArray//m50 Test Results for 3550 Linked-Clone Windows 7 Sessions
The next cluster-level simulation was to run 3550 Horizon linked-clone Windows 7 desktops against the same
FlashArray//m50. All Login VSI parameters were kept consistent with the previous RDSH test with the only
change being to use 3550 linked-clone desktops. As can be seen in the below storage metrics, the Pure
Storage FlashArray//m50 was clearly able to handle this workload and continued to provide sub-millisecond
latency for an impressive Login VSI result.
Firstly, latency was consistently sub-millisecond throughout all phases of the Login VSI simulation despite
driving tens of thousands of IOPs and hundreds of megabytes of read and write bandwidth.
251
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
This next chart shows IOPS ramping up as linked-clone sessions are added to the simulation. The brief dip
occurs when all sessions login and begin to steady-state and then increases again as linked-clone desktops
log out and are refreshed back to the template image.
The bandwidth against the array followed a similar performance curve throughout the simulation with it
ramping up initially, temporarily dropping due to simulation steady-state and then write bandwidth increasing
dramatically as desktops are refreshed at logout.
252
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
Pure Storage FlashArray//m50 Test Results for Full Scale, Mixed Workload Scalability
The next simulation shows the results of combining the earlier 1450 Horizon RDSH sessions with the 3550
Horizon linked-clone sessions for a 5000 user VMware Horizon simulation all on the same Pure Storage
FlashArray//m50 array. Yet again we see performant and consistent results that prove an outstanding user
experience – even at this large of a user scale.
The below two screenshots shows the Pure Storage GUI with the cursor providing more detailed metrics at
both the start of the simulation and at the end. Despite driving nearly 1GB/s in bandwidth we maintain the
responsiveness of low latency throughout the entire test.
253
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
Latency results shown in the above GUI are exactly mirrored from logs pulled from the //m50 array and plotted
in similar fashion to earlier tests. We can see yet again consistent low latency throughout the test run that
show clear evidence of great VDI experience for the end-user.
Read and Write IOPS can be seen dramatically increasing as linked-clone and RSDH sessions are added
during the 2880 second simulation.
254
Appendix C - Pure Storage FlashArray//m50 Expanded Test Results
Similarly, bandwidth increases during the 2880 ramp-up timeframe of the simulation, then falls briefly as the
test hits steady-state and then begins to drop as sessions are logged out.
255