Have an account?

  •   Personalized content
  •   Your products and support

Need an account?

Create an account

SwiftStack, Veeam, and Cisco Solution for Cloud- Scale Storage for Backup and Recovery

White Paper

Available Languages

Download Options

  • PDF
    (7.7 MB)
    View with Adobe Reader on a variety of devices
Updated:April 6, 2020

Available Languages

Download Options

  • PDF
    (7.7 MB)
    View with Adobe Reader on a variety of devices
Updated:April 6, 2020
 

 

This document describes how to use the Cisco Unified Computing System™ (Cisco UCS®) and Cisco HyperFlex™ platforms together with SwiftStack storage and Veeam Backup and Replication software to build and manage a cloud-scale storage solution for backup and recovery data. This document provides an overview of the design, deployment, configuration procedures, and day-2 infrastructure lifecycle management of this combined solution for cloud-scale storage for backup and recovery data.

Executive summary

Cisco, SwiftStack, and Veeam have partnered to deliver a joint solution that enables backup, restoration, and replication for virtualized workloads using SwiftStack storage, Veeam Backup and Replication, and the Cisco Unified Computing System (Cisco UCS®) and Cisco HyperFlex™ platforms. This document describes how to design, deploy, and manage a cloud-scale storage solution for long-term retention of backup and recovery data. This document [OW(1] provides a framework for deploying virtual infrastructure in a Cisco HyperFlex system, a performance storage tier on Cisco UCS S3260 Storage Servers, and a scale-out cloud storage tier also on Cisco UCS S3260 Storage Servers. Veeam Backup and Replication software runs on the performance tier, and SwiftStack storage software runs on the capacity tier. Two additional Cisco UCS C220 Rack Servers are used to run the SwiftStack controller software. Figure 1 provides an overview of this solution. 

A screenshot of a cell phoneDescription automatically generated 

Solution overview

 


 

Introduction

The hyper-growth of data brings a corresponding hyper-growth of backup data, which needs to be managed, stored, recovered, and used. Changing compliance requirements demand longer-term data retention, which puts additional strain on traditional data storage architectures. Organizations trying to maintain control over their rapidly growing data storage and retention needs face capacity, reliability, and cost challenges.

Cisco, SwiftStack, and Veeam together provide a reliable, scalable, cloud-native data storage and management solution that dramatically reduces IT costs while improving data protection. Veeam Cloud Tier, SwiftStack software on Cisco UCS, and Cisco HyperFlex systems work together to simplify backup and recovery at scale and long-term data retention. Data is automatically moved between performance and capacity storage tiers according to user-defined policies. Data remains active and accessible throughout its life.

This solution offers your organization these benefits:

      Drastically reduce data retention costs with flexible, scalable cloud-native storage.

      Scale out seamlessly, from terabytes (TB) to petabytes (PB), with pay-as-you-grow cloud economics.

      Simplify data storage across a single data center, multiple regions, and multiple clouds.

      Retain immediate access to all data, whether it is onsite or offsite.

      Maintain alignment with compliance requirements by storing data as long as needed.

      Reduce capital expenditures (CapEx) and operating expenses (OpEx) by consolidating servers, reducing software licensing costs, automating procedures, and unifying architecture.

      Optimize infrastructure with integrated analytics and open programmable interfaces.

Cisco Unified Computing System

Cisco brings 30 years of leadership and vision to guide businesses through networking and infrastructure challenges. Cisco UCS continues Cisco's long history of innovation in providing integrated systems that deliver business results. Cisco UCS integrated infrastructure solutions speed up IT operations today and create the modern technology foundation needed for the critical business initiatives of tomorrow.

Cisco UCS is a next-generation data center platform that unites computing, network, and storage access resources. The platform is optimized for virtual environments and is designed using open industry-standard technologies, helping 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, multichassis platform in which all resources participate in a unified management domain.

These are the main components of Cisco UCS:

      Computing: The system is based on an entirely new class of computing system that incorporates rack-mount and blade servers based on Intel® Xeon® processors.

      Network: The system is integrated onto a low-latency, lossless, 10-Gbps unified network fabric. This network foundation consolidates LANs, SANs, and high-performance computing networks, which are separate networks today. The unified fabric lowers costs by reducing the number of network adapters, switches, and cables and by decreasing 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 both 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). This capability provides customers with choice of storage access and investment protection. In addition, server administrators can pre-assign storage-access policies for system connectivity to storage resources, simplifying storage connectivity and management for increased productivity.

      Management: The system uniquely integrates all system components, enabling the entire solution to be managed as a single entity by Cisco UCS Manager. Cisco UCS Manager has an intuitive GUI, a command-line interface (CLI), and a robust API to manage all system configuration and operations.

Cisco UCS is designed to deliver these benefits:

      Reduced TCO and increased business agility

      Increased IT staff productivity through just-in-time provisioning and mobility support

      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 that supports hundreds of discrete servers and thousands of virtual machines and the capability to scale I/O bandwidth to match demand

      Industry standards supported by a partner ecosystem of industry leaders

Cisco HyperFlex systems

Cisco HyperFlex systems deliver next-generation hyperconvergence in a data platform that offers end-to-end simplicity for faster IT deployments, unifying computing, networking, and storage resources. Cisco HyperFlex systems are built on Cisco UCS and use a data center architecture that supports traditional, converged, and hyperconverged systems with common policies and infrastructure management. The Cisco HyperFlex HX Data Platform is a purpose-built, high-performance, distributed file system that delivers a wide range of enterprise-class data management and optimization services. This platform redefines distributed storage technology, expanding the boundaries of hyperconverged infrastructure with its independent scaling, continuous data optimization, simplified data management, and dynamic data distribution for increased data availability. This agile system is easy to deploy and manage, scales as your business needs change, and provides top-level data availability

 

SwiftStack storage

Software-defined storage replaces traditional, purpose-built storage. It can be managed as a single platform and automates delivery of services based on built-in intelligence and best practices. The preconfigured, standardized storage components that are typical of today’s software-defined storage solutions deliver many of the same task-specific capabilities as standalone systems. They just use a different approach to achieve them: a faster, easier, more cost-effective approach that relies on software-based specialization and automation to meet different requirements quickly and economically. This approach leads to more consistent, predictable storage solutions that don’t require proprietary expertise to build or maintain them.

With SwiftStack software running on Cisco UCS S-Series Storage Servers, you get hybrid cloud storage that gives you the freedom to move workloads between clouds with universal access to data across on-premises and public infrastructure. SwiftStack was built from the start with the fundamental attributes of the cloud, such as a single namespace across multiple geographic locations, policy-based placement of data, and consumption-based pricing.

SwiftStack storage is optimized for unstructured data, which is growing at an ever-increasing rate in most thriving enterprises. When media assets, scientific research data, and even backup archives reside in a multitenant storage cloud, use of this valuable data increases and unnecessary costs are reduced.

SwiftStack is a fully distributed storage system that horizontally scales to hold your data today and tomorrow. It scales linearly, allowing you to add capacity and performance independently—whatever your applications need.

Although scaling storage typically is a complex process, this is not the case with SwiftStack. No advanced configuration is required. Only a few simple commands are needed to install software on a new Cisco UCS S3260 server and deploy it in the cluster. Load-balancing capabilities are fully integrated, allowing applications to take advantage of the distributed cluster automatically.

Powered by OpenStack Swift at the core, SwiftStack enables you get to use technology that powers some of the largest storage clouds. SwiftStack is the lead contributor to the Swift project, which has more than 220 additional contributors worldwide. With an engine backed by this community and deployed in demanding customer environments, SwiftStack the most proven, enterprise-class object storage software available.

SwiftStack offers these critical features for an active archive:

      Starts as small as 150 TB of usable capacity and scales to hundreds of petabytes

      Spans multiple data centers while presenting a single namespace

      Handles data according to defined policies that align with the needs of different applications

      Uses erasure coding and replicas in the same cluster to protect data

      Offers multitenant support with authentication using Microsoft Active Directory (AD), Lightweight Directory Access Protocol (LDAP), and Keystone

      Supports file protocols (Server Message Block [SMB] and Network File System [NFS]) and object APIs (Amazon Simple Storage Service [S3] and Swift) simultaneously

      Automatically synchronizes with Amazon S3, Google Cloud Storage, and Microsoft Azure with SwiftStack 1space

      Encrypts data and metadata at rest

      Manages highly scalable storage infrastructure through a centralized out-of-band controller

      Optimizes TCO with pay-as-you-grow licensing with support and maintenance included

Veeam Backup and Replication for Cloud Tier

Veeam is an industry leader in the data-protection market. In the era of digital transformation, Veeam recognizes the new challenges companies around the world face in enabling the Always-On Enterprise™, a business that must operate 24.7.365—that is, all the time. To address this need, Veeam has pioneered a new market of Availability for the Always-On Enterprise by helping organizations meet today’s service-level objectives, enabling recovery of any IT service and related applications and data within seconds and minutes. Veeam consistently leads in bringing sophisticated backup and disaster recovery functions to enterprises and cloud providers

Veeam delivers efficient virtual machine backup and replication to dramatically lower the recovery-time objective (RTO) and recovery-point objective (RPO), for a recovery time and point objective (RTPO) of less than 15 minutes for all applications and data. Veeam replication between Cisco HyperFlex clusters, both local and distributed, provides site-level disaster recovery. Veeam also provides backup and recovery at the virtual machine and item levels for instant recovery from more common, day-to-day problems. These isolated Veeam managed backups, stored on secondary storage, in the cloud or on tape, allow organizations to meet both internal and external data-protection and recovery requirements.

Veeam Backup and Replication integrates with Cisco HyperFlex systems by calling native Cisco® snapshot APIs, which improve the performance of backup and replication of VMware vSphere virtual machines hosted on Cisco HyperFlex systems. Cisco HyperFlex snapshots use VMware vStorage API for Array Integration (VAAI), which enables VMware vSphere ESXi hosts to communicate with storage devices and offload storage operations such as snapshot creation and cloning to the storage controller. Veeam Backup and Replication can use Cisco HyperFlex snapshots for virtual machine data processing, which helps speed up backup and replication operations, reduce the impact of backup and replication activities on the production environment, and improve RPOs. During the backup or replication process, Veeam provides application-aware consistency with agentless virtual machine ingest processing and uses Cisco HyperFlex snapshots to preserve this stage for backup. Cisco’s integration with VMware allows Veeam to completely avoid the use of VMware virtual machine snapshots. 


 

SwiftStack, Veeam Cloud Tier, and Cisco UCS data-protection solution

This section describes the requirements and design of the SwiftStack, Veeam Cloud Tier, and Cisco UCS data-protection solution.

Hardware and software requirements

The solution uses the following infrastructure and software components:

      Cisco Unified Computing System

    Cisco UCS S3260 M5 Storage Servers

    Cisco UCS C220 M5SX Rack Server

    Cisco UCS Virtual Interface Card (VIC) 1400

      Cisco HyperFlex HX220c M5SX All Flash Node

      Cisco UCS 6400 Series Fabric Interconnects

      Cisco Nexus® 9000 Series Switches

      Veeam Backup and Replication Version 10 (v10)

      Microsoft Windows 2019 Datacenter Edition for Veeam v10

      SwiftStack storage software Version 7

      Red Hat Enterprise Linux (RHEL) Release 7.7 for SwiftStack storage

Solution design

The data backup and recovery solution for Cisco HyperFlex systems with Veeam and SwiftStack is designed to deliver reliable backup and recovery with low RTOs and RPOs for all applications and data residing in virtual machines in the Cisco HyperFlex environment. The application virtual machines can reside in multiple Cisco HyperFlex nodes across data centers.

Application virtual machines residing in Cisco HyperFlex systems are backed up and replicated through Veeam Backup and Recovery software deployed on a single-node Cisco UCS S3260 Storage Server. This server functions as the performance tier. After the data is backed up in performance tier, it can be moved to the capacity tier either immediately or over time, based on the policy applied.

Veeam Backup and Replication software functions as the performance tier and is deployed on a single-node Cisco UCS S3260 M5 Storage Server.

SwiftStack functions as the capacity tier. SwiftStack controllers are deployed on Cisco UCS C2250 M5 Rack Servers. One server node acts as the primary controller, and another node acts as the backup controller. SwiftStack storage nodes are deployed on three-chassis, six-node Cisco UCS S3260 servers.

Figure 2 shows the design of the solution. Requirements and connectivity details are discussed in the following sections.

A screenshot of a computerDescription automatically generated

Figure 2.       

Solution design

Connectivity details

The connectivity of the solution is based on 25 Gigabit. All servers are connected using 25 Gigabit Ethernet Small Form-Factor Pluggable (SFP) cables. Between each Cisco UCS 6454 Fabric Interconnect and both Cisco Nexus C9336C-FX2 Switches is one virtual port channel (vPC). vPCs allow links that are physically connected to two different Cisco Nexus 9000 Series Switches to appear to the fabric interconnect to be coming from a single device and as part of a single port channel. Between both Cisco Nexus 9336C-FX2 Switches are two 40 Gigabit Ethernet cables. Each Cisco UCS 6332 Fabric Interconnect is connected using one 40 Gigabit Ethernet cable to each Cisco UCS C9336C-FX2 Switch.

The Cisco UCS S3260 servers are connected using two 25 Gigabit Ethernet cables to each fabric interconnect, and the Cisco UCS C220 M5 servers are connected using one 25 Gigabit Ethernet cable.

Software versions

Table 1 lists the software versions used in this solution.

Table 1.        Software versions

 

Component

Software version

Comments

Computing and storage

Cisco UCS S3260 M5 Storage Server

 

Release 4.0(4h)

  Directly managed through fabric interconnect
  Veeam installed on one server, and SwiftStack software installed on three chassis
  Provides storage repository

Cisco HyperFlex HX220c M5SX All Flash Node

 

Release 4.0(4e)

Hyperconverged node for Cisco HyperFlex cluster

Cisco UCS C220 M5SX Rack Server

Release 4.0(4h)

Controller nodes for SwiftStack storage

Management

Cisco UCS Manager

Release 4.0(4h)

Cisco UCS management for all servers directly attached to fabric interconnects

Performance tier

Veeam Backup and Replication software

Version v10

 

Operating system

Microsoft Windows Server 2019

 

Capacity tier

SwiftStack storage

Version 7

 

Operating system

RHEL Server Release 7.7[OW(2] 

 

Hyperconverged software

Cisco HyperFlex HX Data Platform

Release 4.0(1b)

 

Virtualization

VMware vSphere

VMware ESXi Release 6.5.0 Update 2

 

Network

Cisco UCS 6454 Fabric Interconnect

Release 7.0(3)N2(4.04b)

Fabric Interconnect with embedded Cisco UCS Manager

Cisco Nexus C9336C-FX2 Switch chassis

Release 9.3(2)[OW(3] 

  Cisco platform switch for top-of-rack (ToR), middle-of-row (MoR), and end-of-row (EoR) deployments
  Provides connectivity to users and other networks and deployed in Cisco NX-OS standalone mode

 

Table 2 provides information about the hardware configuration for servers.

Table 2.        Table 2  Hardware configuration

Component

Model

Quantity

Comments

Veeam Backup and Replication

 

Cisco UCS S3260 M5 Storage Server

1

  Per server node
  2 x Intel Xeon Gold 6240 processors

   2.60 GHz, with 18 cores

   384 GB of RAM

  Cisco 12G Modular RAID Controller with 2-GB cache
  2 x 480-GB boot solid-state disks (SSDs) for OS
  28 x 12-TB NL-SAS 7200-RPM hard-disk drives (HDDs)
  Cisco UCS VIC 1455

SwiftStack controller node

Cisco UCS C220 M5S Rack Server

2

  2 x Intel Xeon Silver 4210 processors

   2.2 GHz, with 10 cores

   96 GB of RAM

  Cisco 12G SAS RAID Controller
  2 x 600-GB SAS for OS
  Cisco UCS VIC 1457

SwiftStack storage node

Cisco UCS S3260 M5 Storage Server

6

  Per server node
  2 x Intel Xeon Gold 5220 processors

   2.2 GHz, with 18 cores

   192 GB of RAM

  Cisco 12G Modular RAID Controller with 2-GB cache
  2 x 480-GB boot SSDs for OS
  28 x 12-TB NL-SAS 7200-RPM HDDs
  1 X 2-TB Non-Volatile Memory Express (NVMe) for metadata
  Cisco UCS VIC 1455

Cisco HyperFlex node

Cisco HyperFlex HX220c M5SX All Flash Node

3

 

Cisco UCS fabric interconnect

Cisco UCS 6454 Fabric Interconnect

2

 

Cisco Nexus 9000 Series Switches

Cisco Nexus C9336C-FX2 Switch

2

 

Detailed design of SwiftStack, Veeam Cloud Tier, and Cisco UCS solution

The Veeam Cloud Tier feature includes scale-out backup repository architecture. This architecture enables organizations to move older backup files to more cost-effective storage, such as cloud or on-premises object storage. Veeam reports that this new architecture can result in up to 10X savings on long-term data retention costs. Veeam Cloud Tier works with SwiftStack S3-compatible storage systems. Figure 3 shows the Veeam Cloud Tier architecture.

A picture containing screenshotDescription automatically generated

Figure 3.       

Veeam Cloud Tier architecture

When Veeam archives data, it retains stubs—which represent the backup files—in the performance tier along with associated metadata, while most of the data is moved to SwiftStack. This approach makes highly efficient use of resources by storing most of the data on lower-cost storage.

Required components for Veeam Cloud Tier

The Veeam Backup and Replication Server is the core component needed for the Cloud Tier feature (Figure 4). It controls the scale-out repository, tiering, scheduling, settings, restore points, etc. This server is where most of the configuration is performed.

 

A screenshot of a cell phoneDescription automatically generated

Figure 4.       

Veeam Cloud Tier required components

Main features of Veeam Cloud Tier

These are some of the main features of Veeam Cloud Tier:

      Veeam does not charge subscription fees for data residing in object storage.

      Offloaded backup files remain on the performance-tier extents.

    The files exist as a dehydrated shell (metadata and pointers to contents in object storage).

    Shells also serve as local caches to reduce cloud I/O charges.

    Shells can be rehydrated on demand and can be rebuilt from object storage if lost.

    Per-job based block indexing is used (effectively, source-side deduplication).

      All Veeam restore functions work transparently with shells.

    Instant-recovery operations (including the instant virtualization of physical computer backups)

    Entire computer and disk-level restore operations (including restoration to different platforms)

    File-level and item-level restore operations

      Intelligent block recovery is used.

    Restore operations use the closest non-offloaded restore point to source data blocks that are unchanged between the two points.

    This approach improves the efficiency and speed of restore functions and reduces egress charges.

Data-offload process

Prior to release of Veeam v10, two conditions needed to be met to offload data:

      Operational restore window

    Points must be located outside the operational restore window.

    The setting is dictated by policy defined in scale-out backup repository (SOBR) configuration.

      Sealed backup chain

    Points must belong to a sealed backup chain (a chain that will not be transformed).

Veeam v10 adds a new feature that immediately performs a copy operation, providing data that can be immediately offloaded to the capacity tier as soon as first backup is completed. Data is immediately moved to the capacity tier, but optionally you can also keep data on the performance tier as defined for the operational restore window. This new feature helps reduce the storage footprint of the performance tier. It also provides high availability because data is available in SwiftStack storage if the performance tier becomes unavailable during a restore-point window.

The offload process proceeds as follows after backup data is collected from the extents and jobs:

      Veeam determines which data is a candidate to be moved based on conditions.

      Veeam uses veeamagent.exe to offload data to the capacity tier.

      The offload process is performed according to the SOBR Offload job.

    This job runs automatically every 4 hours.

    This job can be run manually using the console or Microsoft PowerShell.

1.     Figure 5 shows the SOBR configuration screen.

A screenshot of a social media postDescription automatically generated

Figure 5.       

Data offload in Veeam v10

Sealed Veeam backup chains are of two types (Figure 6):

      Forward incremental chain with periodic full backups (.vbk and .vib)

      Reverse incremental chain .vib (excluding the latest two)

 

A picture containing wall, sittingDescription automatically generated

Figure 6.       

Sealed Veeam backup chains

Process to move data to the capacity tier and dehydrate the data

The process for moving data to the capacity tier and dehydrating the data is as follows (Figures 7 and 8):

1.     When the sealed chain exits the restore window, the data can be moved.

2.     Object storage metadata is created for each backup file.

3.     The metadata is separated from the data blocks. A dehydration index is created to determine which blocks are copied so that these blocks are not copied again.

4.     When the Veeam incremental backup (VIB) is processed, the dehydration index is checked to find existing blocks. In the case shown in Figures 7 and 8, only block d goes up.

5.     The index is checked to verify that blocks a and b, shown in Figures 7 and 8, are present. Then a metadata file is created and copied, similar to a Resilient File System (ReFS) fast clone. Then the pointer is updated.

A close up of a signDescription automatically generated

Figure 7.       

Process of moving to capacity tier

A picture containing wallDescription automatically generated

Figure 8.       

Dehydration process

Process to move data back to the performance tier and rehydrate the data

In the case shown here, all the data is moved back to the performance tier because the operation relies on “This backup and its dependencies.”

The rehydration process consists of the following steps (Figures 9 and 10):

1.     File 2 is the oldest file, and file 1 is the file that has been placed in the capacity tier.

2.     So that the file can be rehydrated, a rehydration index is created. It identifies the files that are present that are stored locally.

3.     A check is performed to determine whether the blocks are present in the backup file that needs rehydration.

4.     After the blocks are found, blocks a and b don’t need to be copied locally. Only block c needs to be copied in this case.

A green and black signDescription automatically generated

Figure 9.          

Rehydration process

A picture containing wallDescription automatically generated

Figure 10.     

Process of moving to performance tier

Data restore process

All Veeam restore functions work transparently with shells. Veeam supports the following restore features:

      Instant recovery

      Entire computer and disk-level restore

      File-level and item-level restore

Veeam also supports intelligent block-level recovery features:

      Restorations use the closest non-offloaded restore point to source data blocks that are unchanged between the two points.

      This approach improves the efficiency and speed of restore functions and reduces egress charges.

Figure 11 shows the intelligent restore process.

A screenshot of a video gameDescription automatically generated

Figure 11.     

Data restoration using the intelligent restore function

Deploying the solution

The deployment of the combined solution consists of several main steps:

1.     Define the networks required.

2.     Configure Cisco Nexus C9336C-FX2 Switches.

3.     Configure the fabric interconnects.

4.     Deploy server profiles.

5.     Install the operating system.

6.     Install drivers and verify hardware and software compatibility.

7.     Deploy SwiftStack software.

8.     Install Veeam.

9.     Configure Veeam and SwiftStack.

Network topology

You need to segregate the network traffic with separate virtual network interface cards (NICs) and VLANs for Cisco HyperFlex systems, SwiftStack, and Veeam.

Figure 12 shows the network topology used in each application.[OW(4] 

A screenshot of a cell phoneDescription automatically generated

Figure 12.     

Network topology

 

Configuring Cisco Nexus C9336C-FX2 Switches

For detailed steps for configuring Cisco Nexus 9000 Series Switches, see the following link: https://www.cisco.com/c/en/us/support/switches/nexus-9336c-fx2-switch/model.html#ConfigurationGuides.

Here are the high-level configurations required in Cisco Nexus switches:

1.     Enable features: interface VLAN, Hot Standby Router Protocol (HSRP), Link Aggregation Control Protocol (LACP), and vPC.

2.     Set the jumbo maximum transmission unit (jumbomtu; Cisco Nexus 9216).

3.     Configure all the required VLANs.

4.     Configure HSRP in each VLAN.

5.     Configure vPC.

Configuring fabric interconnects

Perform the initial configuration of the Cisco UCS 6454 Fabric Interconnect. Then log in to Cisco UCS Manager and configure the settings listed in the following sections.

Configure the Network Time Protocol server

To configure the Network Time Protocol (NTP) server for the Cisco UCS environment, complete the following steps:

1.     Select the Admin tab on the left side of the window.

2.     Select Time Zone Management.

3.     Select Time Zone.

4.     Under Properties, select your time zone.

5.     Select Add NTP Server.

6.     Enter the IP address and Domain Name System (DNS) name of the NTP server.

7.     Click OK.

Configure global policies

To configure the global policies, complete the following steps (Figure 13):

1.     Select the Equipment tab on the left side of the window.

2.     Select Policies on the right side of the window.

3.     Select Global Policies.

4.     Under Chassis/FEX Discovery Policy, for Action select Platform Max.

5.     Under Rack Server Discovery Policy, for Action select Immediate.

6.     Under Rack Management Connection Policy, for Action select Auto Acknowledged.

7.     Under Power Policy, select Redundancy N+1.

8.     Under Global Power Allocation Policy, select Policy Driven Chassis Group Cap.

9.     Click Save Changes.

A screenshot of a cell phoneDescription automatically generated

Global policies

Enable fabric interconnect server ports

To enable the server ports, complete the following steps:

1.     Select the Equipment tab on the left side of the window.

2.     Choose Equipment > Policies > Port-Auto Discovery Policy.

3.     For Properties, select Enabled.

4.     Click Save Changes to configure server ports automatically for Fabric Interconnects A and B.

5.     Verify the server port on Fabric Interconnect A.

6.     Choose Equipment > Fabric Interconnects > Fabric Interconnect A (primary) > Fixed Module.

7.     Select the Ethernet Ports section.

8.     Verify the server port on Fabric Interconnects A and B.

Enable fabric interconnect ports for uplinks

To enable the uplink ports, complete the following steps:

1.     Select the Equipment tab on the left side of the window.

2.     Choose Equipment > Fabric Interconnects > Fabric Interconnect A (primary) > Fixed Module.

3.     Select the Ethernet Ports section.

4.     Select ports 53 to 54, right-click, and then choose Configure as Uplink Port. Click Yes and then OK.

5.     Repeat the same steps for Fabric Interconnect B.

Label the servers

For easier identification, label each server by completing the following steps:

1.     Select the Equipment tab on the left side of the window.

2.     Choose Chassis > Chassis 1 > Server 1.

3.     In the Properties section on the right, go to User Label and add veeam-node to the field.

4.     Repeat the previous steps for Chassis 2, 3, and 4 using Table 3 for guidance.

5.     Choose Servers > Rack-Mounts > Servers > and repeat the step for all servers using Table 3 for guidance.

Table 3.        Server identification

Server

Name

Chassis 1 and Server 1

Veeam-node

Chassis 2 and Server 1

storage-node1

Chassis 2 and Server 2

storage-node2

Chassis 3 and Server 1

storage-node3

Chassis 3 and Server 2

storage-node4

Chassis 4 and Server 1

storage-node5

Chassis 4 and Server 2

storage-node6

Rack Server 1

hx-node1

Rack Server 2

hx-node1

Rack Server 3

hx-node1

Rack Server 4

swiftstack-controller-1

Rack Server 5

swiftstack-controller-2

Create a MAC address pool

To create a MAC address pool, complete the following steps:

1.     Select the LAN tab on the left side of the window.

2.     Choose LAN > Pools > root > Mac Pools and right-click Create MAC Pool.

3.     For Name, enter SOBR-MAC-Pools.

4.     (Optional) Enter a description for the MAC address pool.

5.     For Assignment Order, select Sequential.

6.     Click Next.

7.     Click Add.

8.     Specify a starting MAC address.

9.     Specify a size for the MAC address pool that is sufficient to support the available server resources: for example, 100.

Create a Kernel-based Virtual Machine IP address pool

To create a Kernel-based Virtual Machine (KVM) IP address pool, complete the following steps (Figure 14):

1.     Select the LAN tab on the left side of the window.

2.     Choose LAN > Pools > root > IP Pools > IP Pool ext-mgmt.

3.     Click Create Block of IPv4 Addresses.

4.     In the From field, enter an IP address.

5.     For Size, enter 50.

6.     Enter your subnet mask.

7.     Enter your default gateway.

8.     Enter your primary DNS and secondary DNS if needed.

9.     Click OK

A screenshot of a cell phoneDescription automatically generated

Creating a KVM IP address pool

Create a universally unique ID pool

To create a universally unique ID (UUID) pool, complete the following steps:

1.     Select the Servers tab on the left side of the window.

2.     Choose Servers > Pools > root > UUID Suffix Pools and right-click Create UUID Suffix Pool.

3.     For Name, enter SOBR-UUID-Pools.

4.     (Optional) Enter a description for the UUID pool.

5.     Set Assignment Order to Sequential and click Next.

6.     Click Add.

7.     Specify a starting UUID suffix.

8.     Specify a size for the UUID suffix pool that is sufficient to support the available server resources: for example, 100.

9.     Click OK.

10.  Click Finish and then OK.

Enable Cisco Discovery Protocol

To enable network control policies, complete the following steps (Figure 15):

1.     Select the LAN tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose LAN > Policies > root > Network Control Policies and right-click Create Network-Control Policy.

3.     In the Name field, enter Enable-CDP.

4.     (Optional) Enter a description in the Description field.

5.     For CDP, select Enabled.

6.     For MAC Register Mode, select All Host Vlans.

7.     Leave everything else unchanged and click OK.

8.     Click OK.

A screenshot of a cell phoneDescription automatically generated

Figure 15.     

Setting network control policy

Create VLANs

As mentioned earlier, you should separate the network traffic with VLANs. Table 4 summarizes the VLAN configurations for SwiftStack storage.

Table 4.        VLAN configurations

Server

Name

Cluster_Network

120

Replication_Network

130

Client_Network

140

Storage_Data_Network

150

vMotion_Network

160

VM_Network

170

Management_Network

220

 

To configure VLANs in the Cisco UCS Manager GUI, complete the following steps (Figure 16):

1.     Select LAN in the left pane in the Cisco UCS Manager GUI.

2.     Choose LAN > LAN Cloud > VLANs and right click Create VLANs.

3.     Enter Management_Network as the VLAN name.

4.     Keep Multicast Policy Name set to <not set>.

5.     Select Common/Global.

6.     Enter 220 in the VLAN IDs field.

7.     Click OK and then click Finish.

8.     Repeat the steps for rest of the VLANs using Table 4 for guidance.

A screenshot of a cell phoneDescription automatically generated

Creating VLANs

Create the virtual NIC template

Using the previous section in which you created VLANs as the basis, you now need to create the appropriate virtual NIC (vNIC) templates. For this solution, you need to create four vNICs for SwiftStack storage nodes and three vNICs for Veeam. Table 5 provides an overview of the configurations. vNICs for Cisco HyperFlex systems are created automatically when you deploy Cisco HyperFlex and thus you do not need to create them.

Table 5.        vNIC template configuration details

 

vNIC name

Fabric

VLAN name

Failover

MTU

Cisco HyperFlex system

hv-mgmt-a

0

Management_Network

No

1500

hv-mgmt-b

1

Management_Network

No

1500

Hv-vmotion-a

0

VM_Network

No

9000

Hv-vmotion-b

1

vMotion_Network

No

9000

storage-data-a

0

Storage_Data_Network

No

9000

storage-data-b

1

Storage_Data_Network

No

9000

vm-network-a

0

VM_Network

No

9000

vm-network-b

1

VM_Network

No

9000

 

Veeam

 

Client

0

Management_Network

Yes

9000

Management

1

Storage_Data_Network

Yes

9000

Storage

0

Client_Network

Yes

9000

SwiftStack

Client

1

Management_Network

Yes

9000

Management

 

Cluster_Network

Yes

9000

Cluster

0

Replication_Network

Yes

9000

Replication

1

Client_Network

Yes

9000

 

To create the appropriate vNICs, complete the following steps (Figure 17):

1.     Select the LAN tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose LAN > Policies > root > vNIC Templates and right-click Create vNIC Template.

3.     In the Name field, enter Management.

4.     (Optional) Enter a description in the Description field.

5.     For Fabric ID, select Fabric A and enable failover.

6.     For Template Type, select Updating Template.

7.     Select the Management VLAN and click Native VLAN.

8.     For MAC Pool, select SOBR-MAC-Pools.

9.     For Network Control Policy, select Enable-CDP.

10.  Click OK and then click OK again.

11.  Repeat these steps for the Storage, Client, Cluster, Replication vNICs. Make sure you select the correct fabric ID, VLAN, and MTU size as listed in Table 5.

A screenshot of a cell phoneDescription automatically generated

Figure 17.     

Creating a vNIC template

Create Ethernet adapter policy  

By default, Cisco UCS provides a set of Ethernet adapter policies. These policies include the recommended settings for each supported server operating system. Operating systems are sensitive to the settings in these policies.

A Cisco UCS best practice is to enable Jumbo Frames MTU 9000 for any storage-facing networks (Storage-Mgmt and Storage-Cluster). Enabling jumbo frames on specific interfaces and modifying transmit (Tx) and receive (Rx) values helps ensure 39-Gbps bandwidth on the Cisco UCS fabric.

To create a specific adapter policy for Red Hat Enterprise Linux, complete the following steps (Figure 18):

1.     Select the Server tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Servers > Policies > root > Adapter Policies and right-click Create Ethernet Adapter Policy.

3.     In the Name field, enter SOBR.

4.     (Optional) Enter a description in the Description field.

5.     Under Resources, enter the following values:

a. Transmit Queues: 8

b. Ring Size: 4096

c. Receive Queues: 8

d. Ring Size: 4096

e. Completion Queues: 16

f. Interrupts: 32

6.     Under Options, enable Receive Side Scaling (RSS).

7.     Click OK and then click OK again.

A screenshot of a cell phoneDescription automatically generated

Figure 18.       

Ethernet adapter settings

Create boot policy

To create a boot policy, complete the following steps (Figure 19):

1.     Select the Servers tab in the left pane.

2.     Choose Servers > Policies > root > Boot Policies and right-click Create Boot Policy.

3.     In the Name field, enter SOBR.

4.     (Optional) Enter a description in the Description field.

5.     Click Add CD/DVD and click OK.

6.     Click Add Local Disk

7.     Click OK.

A screenshot of a cell phoneDescription automatically generated

Figure 19.       

Creating a boot policy

Create LAN connectivity policy for SwiftStack

To create a LAN connectivity policy for SwiftStack, complete the following steps (Figures 20 and 21):

1.     Select the LAN tab in the left pane.

2.     Choose LAN > Policies > root > LAN Connectivity Policies and right-click Create LAN Connectivity Policy.

3.     In the Name, field, enter SwiftStack.

4.     (Optional) Enter a description in the Description field.

5.     Click Add.

6.     In the Name field, enter Management.

7.     Select Use vNIC Template.

8.     For vNIC Template, choose Management from the drop-down list.

9.     If you are using the Jumbo Frame MTU 9000 setting, choose the default adapter policy, previously created as SOBR, from the drop-down list.

10.  Repeat these steps for the remaining networks: Cluster, Replication, and Client. Make sure you choose SOBR for Adapter Policy.

A screenshot of a cell phoneDescription automatically generated

Figure 20.     

Creating LAN connectivity policy

A screenshot of a cell phoneDescription automatically generated

Figure 21.     

Creating a vNIC

Create LAN connectivity policy for Veeam

To create a LAN connectivity policy for SwiftStack, complete the following steps:

1.     Select the LAN tab in the left pane.

2.     Choose LAN > Policies > root > LAN Connectivity Policies and right-click Create LAN Connectivity Policy.

3.     In the Name field, enter Veeam.

4.     (Optional) Enter a description in the Description field.

5.     Click Add.

6.     In the Name field, enter Management.

7.     Select Use vNIC Template.

8.     For vNIC Template, choose Management from the drop-down list.

9.     If you are using the Jumbo Frame MTU 9000 setting, select the default adapter policy, previously created as SOBR, from the drop-down list.

10.  Repeat these steps for the remaining networks: Storage and Client. Make sure you choose SOBR for Adapter Policy.

Create maintenance policy

To create a maintenance policy, complete the following steps (Figure 22):

1.     Select the Servers tab in the left pane.

2.     Choose Servers > Policies > root > Maintenance Policies and right-click Create Maintenance Policy.

3.     In the Name field, enter SOBR.

4.     (Optional) Enter a description in the Description field.

5.     For Reboot Policy, select User Ack.

6.     Click OK and then click OK again.

7.     Create maintenance policy.

A screenshot of a cell phoneDescription automatically generated

Creating maintenance policy

Creating a chassis profile

You need a chassis profile to assign specific disks to a particular server node in a Cisco UCS S3260 Storage Server as well as to upgrade to a specific chassis firmware package.

Create a chassis firmware package

To create a chassis firmware package, complete the following steps (Figure 23):

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Policies > root > Chassis Firmware Package and right-click Create Chassis Firmware Package.

3.     In the Name field, enter SOBR.

4.     (Optional) Enter a description in the Description field.

5.     For Chassis Package, choose 4.0(4d)C [OW(5] from the drop-down list.

6.     Click OK and then click OK again.

7.     Create the chassis firmware package.

A screenshot of a cell phoneDescription automatically generated

Figure 23.       

Creating a chassis firmware package

Create chassis maintenance policy

To create a chassis maintenance policy, complete the following steps:

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Policies > root > Chassis Maintenance Policies and right-click Create Chassis Maintenance Policy.

3.     In the Name field, enter SOBR-Maint.

4.     (Optional) Enter a description in the Description field.

5.     Click OK and then click OK again.

6.     Create chassis maintenance policy.

Create disk zoning policy for SwiftStack

Table 6 shows the disk zoning configuration details for SwiftStack.

Table 6.        Disk zoning configuration details

 

Server

Controller

Slot range

Veeam

Server-1

1

1 to 28

2

29 to 56

SwiftStack

Server-1

1

1 to 14

2

15 to 28

Server-2

1

29 to 42

2

43 to 56

 

To create a disk zoning policy, complete the following steps:

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Policies > root > Disk Zoning Policies and right-click Create Disk Zoning Policy.

3.     In the Name field, enter SwiftStack.

4.     (Optional) Enter a description in the Description field.

5.     Create disk zoning policy.

6.     Click Add.

7.     For Ownership, select Dedicated.

8.     For Server, choose 1.

9.     For Controller, choose 1.

10.  Add slot range 1-14 for the top node of the Cisco UCS S3260 Storage Server and click OK (Figure 24).

A screenshot of a cell phoneDescription automatically generated

Figure 24.       

Creating disk zoning

11.  For Server, choose 1.

12.  For Controller, choose 2.

13.  Add slot range 15-28 for the top node of the Cisco UCS S3260 Storage Server and click OK (Figure 25).

 A screenshot of a cell phoneDescription automatically generated

Figure 25.       

Adding slots to disk zoning policy for Server-1 and Controller-1

 

14.  Click Add.

15.  For Ownership, select Dedicated.

16.  For Server, choose 2.

17.  For Controller, choose 1.

18.  Add slot range 29-42 for the bottom node of the Cisco UCS S3260 Storage Server and click OK (Figure 26).

A screenshot of a cell phoneDescription automatically generated

Figure 26.       

Adding slots to disk zoning policy for Server-2 and Controller-1

19.  For Server, choose 2.

20.  For Controller, choose 2.

21.  Add slot range 43-56 for the bottom node of the Cisco UCS S3260 Storage Server and click OK (Figure 27).

A screenshot of a cell phoneDescription automatically generated

Figure 27.       

Adding slots to disk zoning policy for Server-2 and Controller-2

Create disk zoning policy for Veeam

To create a disk zoning policy, complete the following steps:

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Policies > root > Disk Zoning Policies and right-click Create Disk Zoning Policy.

3.     In the Name field, enter Veeam.

4.     (Optional) Enter a description in the Description field.

5.     Create disk zoning policy.

6.     Click Add.

7.     For Ownership, select Dedicated.

8.     For Server, choose 1.

9.     For Controller, choose 1.

10.  Add slot range 1-28 for the top node of the Cisco UCS S3260 Storage Server and click OK.

11.  Click Add.

12.  For Server, choose 1.

13.  For Controller, choose 2.

14.  Add slot range 29-56 for the top node of the Cisco UCS S3260 Storage Server and click OK.

Create a chassis profile template

First, create a chassis profile template for SwiftStack. Then create one for Veeam.

To create a chassis profile template, complete the following steps:

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Chassis Profile Templates and right-click Create Chassis Profile Template.

3.     In the Name field, enter SwiftStack.

4.     For Type, select Updating Template.

5.     (Optional) Enter a description in the Description field.

6.     Create the chassis profile template.

7.     Click Next.

8.     Under the Chassis Maintenance Policy button, select the chassis maintenance policy you previously created.

9.     Click Next.

10.  Click the + button and under Chassis Firmware Package select the chassis firmware package policy you previously created.

11.  Under Disk Zoning Policy, select the disk zoning policy you previously created for SwiftStack.

12.  Click Finish and then click OK again.

13.  Repeat the preceding steps to create a chassis profile template for Veeam.

Create chassis profiles from the template for SwiftStack

To create chassis profiles from the chassis profile template you previously created, complete the following steps (Figure 28):

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Chassis Profile Templates and select SwiftStack (the template you created previously).

3.     Right-click and choose Create Chassis Profiles from Template.

4.     In the Name field, enter SwiftStack.

5.     Leave the Name Suffix Starting Number field unchanged.

6.     Enter 3 in the Number of Instances field for all connected Cisco UCS S3260 Storage Servers.

7.     For Chassis Profile Template, choose SwiftStack.

8.     Click OK.

A screenshot of a cell phoneDescription automatically generated

Figure 28.       

Creating chassis profiles from a template

Create a chassis profile from a template for Veeam

To create a chassis profile for Veeam, complete the following steps (Figure 29):

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Chassis Profile Templates and select SwiftStack (the template you created previously).

3.     Right-click and choose Create Chassis Profile from Template.

4.     In the Name field, enter Veeam.

5.     For Chassis Profile Template, choose Veeam.

6.     Click OK.

 

A screenshot of a cell phoneDescription automatically generated

Creating a chassis profile for Veeam

Associate chassis profiles

To associate all previously created chassis profiles, complete the following steps:

1.     Select the Chassis tab in the left pane of the Cisco UCS Manager GUI.

2.     Choose Chassis > Chassis Profiles and select Veeam.

3.     Right-click Change Chassis Profile Association.

4.     Under Chassis Assignment, choose Select Existing Chassis from the drop-down list.

5.     Under Available Chassis, select ID 1.

6.     Click OK and then click OK again.

7.     Repeat the preceding steps for the SwiftStack chassis profiles, selecting IDs 2 through 4.

8.     In the pop-up window that appears at the top right, click Chassis Profiles and Acknowledge All Chassis Profiles.

9.     Click Apply.

10.  Click OK.

Creating storage profiles

A storage profile encapsulates the storage requirements for one or more service profiles. Logical unit numbers (LUNs) configured in a storage profile can be designated as boot LUNs or data LUNs and can be dedicated to a specific server. You can also specify a local LUN as a boot device.

Set disks for Cisco UCS S3260 M5 servers to Unconfigured-Good

To prepare the OS drives reserved from the S3260 M5 servers for storage profiles, make sure the disks are converted from JBOD to Unconfigured-Good. To convert the disks, complete the following steps:

1.     Select the Equipment tab in the left pane of the Cisco UCS Manager GUI.

2.     For S3260 M5 servers, choose Equipment > Chassis > Chassis1 > Servers > Server1 > Inventory > Storage > Disks.

3.     Select the two disks in slots 201 and 202 and right-click to change JBOD to Unconfigured-Good (Figure 30).

A screenshot of a cell phoneDescription automatically generated

Figure 30.       

Disks for OS boot\

4.     Repeat the preceding steps for the other S3260 M5 servers.

Create a storage profile for SwiftStack

To create the storage profile for SwiftStack, complete the following steps:

1.     Select Storage in the left pane of the Cisco UCS Manager GUI.

2.     Choose Storage > Storage Profiles and right-click Create Storage Profile.

3.     In the Name field, enter SwiftStack-SP.

4.     (Optional) Enter a description in the Description field.

5.     Click Add.

6.     In the Name field, enter OS-Boot.

7.     Configure the profile as follows:

a. Create a local LUN.

b. Set Size (GB) = 1.

c. Set Fractional Size (MB) = 0.

d. Select Auto Deploy.

e. Select Expand to Available.

8.     Click Create Disk Group Policy to create the RAID 1 LUN.

9.     In the Name field, enter os_boot_dg.

10.  (Optional) Enter a description in the Description field.

11.  For RAID Level, set RAID 1 Mirrored.[OW(6] 

12.  Select Disk Group Configuration (Manual).

13.  Click Add.

14.  For the slot number, enter 201.

15.  Click OK and then click Add.

16.  For the slot number, enter 202.

17.  Click OK and then click OK again.

18.  Under Select Disk Group Configuration, select the radio button for disk group policy you previously created for the boot disk.

19.  Select Disk Group Configuration and complete the steps in the wizard.

Create a storage profile for Veeam

Veeam requires two disk groups in the storage profile: one for the OS and the other for the Veeam repository. Follow the same steps as for creating a storage profile and disk group for SwiftStack. Add another disk group using the details in Table 7.

Table 7.        Disk group details

Number of disks

RAID group

Number of spans

Number of disks per span

Number of global hot spares

14

RAID 6

-

-

1

28

RAID 60

2

13

2

42

RAID 60

3

13

3

56

RAID 60

4

13

4

 

Here, for a 56-disk configuration of RAID 60, you will have 4 RAID spans, and each span will have 13 disks. The remaining 4 disks will be allocated for global hot spares.

Follow these steps to the create desk group:

1.     For Name, enter Veeam-Repo. For RAID Level, choose RAID 60, and select Manual Disk Group Configuration.

2.     Click Add, for Slot Number enter 1, for Role select Normal, and set Span ID to 0.

3.     Repeat the preceding steps for slots 2 through 13 with Span ID set to 0.

4.     For slot 14, set Role to Global Hot Spare and set Span ID to 0.

5.     Repeat the steps for slots 15 through 28 with Span ID set to 1.

6.     For slot 27, set Role to Dedicated Hot Spare and set Span ID to 1.

7.     For slot 28, set Role to Global Hot Spare and leave Span ID set as unspecified.

8.     Repeat the steps for slots 29 through 42 for slots 29 through 41 with Role set to Normal and Span ID set to 2. For slot 42, set Role to Global Hot Spare and set Span ID as unspecified.

9.     Repeat the steps for slots 43 through 56 for slots 43 through 57 with Role set to Normal and Span ID set to 3. For slot 56, set Role to Global Hot Spare and set Span ID to unspecified.

10.  When all 56 disks with 4 spans are configured, do the following:

a. Set Strip Size to 64 KB.

b. Set Access Policy to Read Write.

c. Set Read Policy to Read Ahead.

d. Set Write Cache Policy to Write Back Good Battery Backup Unit (BBU).

e. Set I/O Policy to Direct.

 

Creating service profiles

Before you create a service profile, you need to create service profile template.

Create a service profile template for the Cisco UCS S3260 Storage Server

To create a service profile template, complete these steps:

1.     Select Servers in the left pane of the Cisco UCS Manager GUI.

2.     Choose Servers > Service Profile Templates > root and right-click Create Service Profile Template.

3.     For Name, enter SwiftStack-Template.

4.     For Template Type, select Updating Template.

5.     In the UUID Assignment section, select the UUID pool that you created previously.

6.     On the Storage Profile Policy tab, select the SwiftStack-SP storage profile for the top node of the Cisco UCS S3260 Storage Server you created previously.

7.     Click Next.

8.     Keep the Dynamic vNIC Connection Policy field at the default setting.

9.     Set LAN connectivity to use the connectivity policy you created previously.

10.  From the LAN Connectivity drop-down list, choose SwiftStack (which you created previously) and click Next.

11.  Click Next to continue with SAN connectivity.

12.  For “How would you like to configure SAN connectivity?” select No vHBA.

13.  Click Next to continue with zoning.

14.  Click Next.  

15.  Choose Specify Manually form the drop-down list.

16.  In the PCI Order section, sort all the vNICs.

17.  Make sure the vNICs are listed in this order: Management > 1, Cluster > 2, Replication > 3, and Client > 4.

18.  Click Next to continue with vMedia policy.

19.  Click Next.

20.  For Boot Policy, choose the boot policy you created previously.

21.  Create the server boot order.

22.  Click Next.

23.  For Maintenance Policy, choose the maintenance policy you created previously.

24.  Click Next.

25.  Under Server Assignment, leave everything unchanged.

26.  Click Next.

Create a service profile template for Veeam

Follow the steps used to create a service profile template for SwiftStack.

1.     Make sure the storage profile for Veeam is selected. From the LAN Connectivity drop-down list, choose Veeam.

2.     Under vNIC/vHBA Placement, make sure that the vNICs are listed in this order: Management > 1, Storage > 2, and Client > 3.

Create service profiles from the templates

To create the service profiles, complete the following steps:

1.     Select Servers from the left pane of the Cisco UCS Manager GUI.

2.     Choose Servers > Service Profiles and right-click Create Service Profile from Template.

3.     In the Name Prefix field, enter Storage-Node1.

4.     For Service Profile Template, choose SwiftStack-Template (the template you created previously) for the top node of the Cisco UCS S3260 Storage Server.

5.     Click OK and then click OK again.

6.     Repeat these steps to create service profiles for the remaining S3260 M5 servers for the remaining five SwiftStack storage nodes.

7.     Follow the same steps to create a service profile for Veeam as well.

Associate service profiles with Cisco UCS S3260 M5 servers

To associate all the Storage-NodeX service profiles with the Cisco UCS S3260 M5 Storage Servers, complete the following steps:

1.     Select Servers from the left pane of the Cisco UCS Manager GUI.

2.     Choose Servers > Service Profiles and right-click Storage-Node1 (the service profile you created previously).

3.     Click Change Server Profile Association.

4.     From the Server Assignment drop-down list, choose Select Existing Server.

5.     Select the Available Servers button.

6.     From the chassis and slot lists, choose Chassis 2 and Slot 1 for Storage-Node1.

7.     Click OK.

8.     Repeat the preceding steps to create a service profile for Veeam and assign it to Chassis 1 and Server 1.

Create a service profile for the Cisco UCS C220 M5 server for SwiftStack controller nodes

SwiftStack controllers require a LUN to install the OS and management network. Thus, create a service profile with a single RAID 0 volume and a vNIC for management connectivity[OW(7] .

Installing the operating system

You need to install the operating system on the SwiftStack and Veeam nodes.

Install the OS on SwiftStack storage and controller nodes

Install RHEL 7.7 on the SwiftStack controller for Cisco UCS (Cisco UCS C220 M5 server) and storage nodes (Cisco UCS S3260 server).

Complete the following steps to install OS:

1.     Log in to the Cisco UCS Manager and select the Equipment tab in the left pane.

2.     Choose Equipment > Chassis 2 > Server > Server 1 and right-click KVM Console.

3.     Launch the KVM Console.

4.     On the Virtual Media tab, click Activate Virtual Devices.

5.     In the Cisco UCS KVM window, select the Virtual Media tab and click CD/DVD.

6.     Click Choose File, browse to the RHEL 7.7 installation ISO image, and click Map Drive (Figure 31).

A screenshot of a social media postDescription automatically generated

Figure 31.       

Attaching virtual media

7.     In the KVM window, choose Macros > Static Macros and click the Ctrl-Alt-Del button in the upper-left corner.

8.     Click OK and then click OK again to reboot the system.

9.     On the boot screen with the Cisco Logo, press F6 for the boot menu.

10.  When the boot menu appears, choose Cisco vKVM-Mapped vDVD1.24.

11.  Complete the installation.

Install the OS on the Veeam node.

Install Microsoft Windows 2019 on the Veeam node:

1.     To load the RAID controller driver for the S3260 M5, on the setup screen that asks “Where do you want to install Windows?” click “Load driver” (Figure 32).

A screenshot of a cell phoneDescription automatically generated

Figure 32.       

Loading the storage driver

2.     In the Virtual Media window, unmap the Windows installer ISO file and map the S3260 drivers ISO file. The drivers for the S3260 can be downloaded from www.cisco.com. Choose Support > Products > Servers - Unified Computing > Cisco UCS C-Series Rack-Mount UCS-Managed Server Software > Download Software > Unified Computing System (UCS) Drivers and select “ISO image of UCS-Rack related windows drivers only.”

3.     Install Windows on the RAID 1 configuration created from the two SSDs.

Deploying SwiftStack software

You need to install two components: the SwiftStack controller and claiming storage nodes. Then you need to configure the cluster. You will deploy the controllers on the Cisco UCS C220 M5 and the storage nodes on Cisco UCS C240 M5 servers.

Detailed information about SwiftStack software deployment can be found in the Cisco UCS S3260 Storage Server with SwiftStack Software-Defined Object Storage Validated Design and Deployment document at https://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/UCS_CVDs/ucs_s3260_m5_swiftstack.html.

Detailed information about deploying SwiftStack storage can be found at https://www.swiftstack.com/docs/install/index.html.

Install on-premises controller software

The high-level steps for installing on-premises controller software are as follows:

1.     Download the installer software from portal.swiftstack.com by navigating to Downloads section.

2.     Run the SwiftStack controller installer shell script and complete the installation process.

3.     Log in to the URL pointed to after the installation process is complete as the localadmin user and the default password.

4.     Enter the license key obtained from SwiftStack and enter your host name and new password to proceed. You can leave the other values at the default settings. Click Submit.

5.     After you log in again with the new password, the system will prompt you to add the nodes and create the cluster.

Claim storage nodes

Before running the installer software, make sure that Secure Sockets Layer (SSL) certificates are installed.

On the controller node, enter the following:

[root@swiftcontroller ~]# cd /opt/ss/etc/ 
[root@swiftcontroller etc]# ls -l ssman.crt
-rw-r--r-- 1 root 668 Jun  6 11:16 ssman.crt 

Copy this certificate to all server nodes:

scp ssman.crt root@swiftstack-node:/etc/pki/ca-trust/source/anchors/ 

Run update-ca-trust extract as the root user on the storage node:

[root@swiftstack-server28-2 .ssh]# update-ca-trust extract
[root@swiftstack-server28-2 .ssh]#

Run the Curl command on the storage node:

curl https://swiftcontroller.cisco.com:443/install | sudo bash

After processing this command, the system will print the claim URL.

Configure the SwiftStack controller for nodes

Follow these steps to configure the SwiftStack controller for the nodes:

1.     Claim the nodes. Run the Claim URL in a browser to claim the nodes.

2.     Click Create Cluster to create and configure the cluster (Figure 33).

A screenshot of a cell phoneDescription automatically generated

Figure 33.       

Creating a new cluster

3.     Click the Configure button and configure the basic settings as shown in Figure 34. Then submit the changes.

A screenshot of a cell phoneDescription automatically generated

Figure 34.     

Basic cluster configuration

4.     Click User and Accounts on the right side of the page to create Swift users. These are Swift Cluster accounts, and a minimum of one account is needed for the cluster (Figure 35).

A screenshot of a cell phoneDescription automatically generated

Figure 35.     

Cluster navigation tabs

 

5.     Click Nodes in the left pane and ingest the nodes.

6.     Confirm the networks when prompted by the system for outward-facing, cluster, and replication networks.

7.     Set up the node:

a. Select all the disks and format the drives.

b. Select the SSD drive. Click Add Policies and select Account and Container (Figure 36).

A screenshot of a cell phoneDescription automatically generated

Figure 36.     

Adding and removing policies

c. Select the HDDs (12 disks), click Add Policies, and select Standard-Replica.

8.     After the setup is complete, verify that the policies are displayed correctly on the node.

9.     In the left menu, enable the node.

This completes the setup of one node. Repeat the procedure for all the storage nodes.

Deploy the configuration

To deploy the configuration, choose Home > Clusters > Deploy and click Deploy Config to Swift Nodes (Figure 37).

A screenshot of a cell phoneDescription automatically generated

Figure 37.     

Deploying the configuration

Installing the Cisco HyperFlex cluster

The application virtual machines will run on the Cisco HyperFlex cluster.

This document does not cover the installation details the Cisco HyperFlex system. For details about the installation process, see [OW(8] https://www.cisco.com/c/en/us/support/hyperconverged-systems/hyperflex-hx-data-platform-software/products-installation-guides-list.html.

Installing Veeam

The high-level steps for installing Veeam are as follows:

1.     Download Veeam Backup and Replication Software v10 from Veeam’s Software Download page using you Veeam account login information.

2.     From the Download page, also download the license file.

3.     Install Veeam v10 on a dedicated Windows system.

For more details, see https://helpcenter.veeam.com/docs/backup/vsphere/install_vbr.html?ver=100.

Configuring SwiftStack and Veeam

Now configure SwiftStack and Veeam.

Configure SwiftStack

The SwiftStack controller is used to create a Veeam S3 user (for example, sobr), which is connected to the SwiftStack S3 target. The SwiftStack controller generates an S3 API key when the configuration is pushed to the cluster (Figure 38).

 

A screenshot of a cell phoneDescription automatically generated

Figure 38.     

Creating an S3 user for Veeam

A screenshot of a cell phoneDescription automatically generated

A dedicated S3 bucket (for example, veeam-data) is created for the Veeam user using the SwiftStack client. For the purposes of this solution, the bucket policy is defined as triple replica. However, 8-4 erasure coding can also be used for certain data-protection configurations.

Create the ReFS disk volume for the Veeam repository

ReFS volumes provide significantly faster performance for synthetic full backup creation and transformation. They also reduce storage requirements and improve reliability. Even more important, they improve the availability of backup storage by significantly reducing its load, resulting in improved backup and restore performance and enabling customers to do much more with virtual labs.

To create a disk volume for the Veeam repository, complete the following steps:

1.     Choose Server Manager > File and Storage Services.

2.     Navigate to Volumes > Disks and select the volume created earlier with the name Cisco UCS-S3260-DRAID (Figure 39).

3.     Create a new volume.

A screenshot of a social media postDescription automatically generated

Figure 39.       

Creating a Veeam repository

4.     Click Next until you reach the “Select file system settings” window (Figure 40).

a. Create a volume label.

b. Set the file system to ReFS.

c. Set an allocation unit size of 64 KB.

d. For the volume label, specify VeeamRepo.

5.     Confirm the file system settings and click Create.

A screenshot of a cell phoneDescription automatically generated

Figure 40.     

Configuring file system settings for the repository

Configure the Veeam primary repository

The Veeam primary repository is configured on the drive that you created previously.

1.     On the Veeam Backup and Replication Console, create a backup repository. Choose Backup Infrastructure > Backup Repository > Add Backup Repository (Figure 41).

A screenshot of a cell phoneDescription automatically generated 

Figure 41.     

Adding the backup repository

2.     In the Backup Repository wizard, select direct-attached storage options.

3.     Enter a name and description for the new repository. Click Next.

4.     Set “Path to folder” to the backup location you will use in the test setup as shown previously (Figure 42).

A screenshot of a social media postDescription automatically generated

Figure 42.     

Configuring the backup repository

5.     Click Advanced and select “Use per-VM backup files.” Click OK and complete the wizard (Figure 43).

A screenshot of a cell phoneDescription automatically generated

Figure 43.     

Selecting per–virtual machine backup files

Configure the Veeam object storage repository

Now configure the object storage repository:

1.     Under Backup Infrastructure in the navigation pane, right-click Backup Repository and choose Add Backup Repository.

2.     In the Backup Repository wizard, select Object storage.

3.     Enter a name and description for the object storage repository and click Next.

4.     For the service point, enter the S3 API URL from SwiftStack and specify the region (Figure 44).

A screenshot of a social media postDescription automatically generated

Figure 44.     

Configuring the Veeam object storage repository

5.     Click Add to add credentials. The access key is the user name of the SwiftStack user created previously, and the secret key is the S3 API key for the user (Figure 45).

A screenshot of a cell phoneDescription automatically generated

Figure 45.       

Adding credentials

6.     Identify the bucket (for example, veeam-data) created previously and also create a folder (for example, backup) for this repository.

7.     Complete the wizard. The SwiftStack client should show the three objects created under veeam-data bucket (Figure 46).

A screenshot of a video gameDescription automatically generated

Figure 46.     

The SwiftStack client showing veeam-data

Create the scale-out repository

Next create the scale-out repository:

1.     In the Backup Infrastructure view, click Scaleout Repositories and click the Add Scaleout Repository button on the ribbon.

2.     Enter a name and description for this new scaleout repository. Click Next to continue.

3.     On the Performance Tier screen, click the Add button to add a backup repository to the scaleout repository. Add the primary repository that was created for your backups. Click OK to add the repository (Figure 47).

A screenshot of a social media postDescription automatically generated

Figure 47.     

Creating the scale-out repository

4.     Click the Advanced button to view advanced settings for the performance tier. Deselect the “Use perVM backup files” option and click OK. Click Next to continue.

5.     On the Placement Policy screen, keep the default data locality policy. This policy allows all backup files (full and incremental) to be located on the same extent. Click Next to proceed.

6.     On the Capacity Tier screen, select “Extend scaleout backup repository capacity with object storage” and choose the object storage solution from the dropdown list. Also, select “Copy backups to object storage as soon as they are created” and “Move backups to object storage as they edge out of the operational restore window.”

7.      Click Apply to apply these changes (Figure 48).

A screenshot of a social media postDescription automatically generated

Figure 48.     

Configuring the capacity tier in the scale-out repository

8.     On the Summary screen, click Finish to create the scaleout repository for the test. When the process is complete, in the Inventory Pane click the scaleout backup repository that you just created. The main window will show all the extents in the SOBR.

Recommended best practices

This section presents some recommended best practices for Veeam and SwiftStack.

Veeam configuration best practices

Very little customization is needed to use SwiftStack for Veeam Cloud Tier, but a few best-practices settings will help you achieve optimal performance:

1.     In the Backup Infrastructure pane, click Backup Proxies. Right-click the created proxy and choose Properties to edit this proxy.

2.     On the Server screen, set the “Max concurrent tasks” to 4. Click Finish to apply these changes.

Veeam supports four block sizes for backup:

      Local target (large amount of data: 16 TB or more): 4096-KB data blocks

      Local target: 1024-KB data blocks

      LAN target: 512-KB data blocks

      WAN target: 256-KB data blocks

Veeam’s default block size for best performance is 1024 KB. However, for larger data amounts, you should use 4096-KB data blocks. The smaller the block size, the more calls need to be made to the object storage to upload the data, which will increase the cost to upload data to the object storage and affect overall backup performance.

SwiftStack configuration and parameter tuning best practices

Tuning several SwiftStack parameters can help optimize performance.

Storage policy

When creating a bucket (or container) in SwiftStack for use with Veeam, you need to decide which storage policy to apply to the bucket. Storage policies in SwiftStack define the physical servers and devices that are used to store data as well as the type of data-protection strategy used to protect the data. Policies can be configured to use either erasure coding or full replicas of objects.

SwiftStack supports both triple-replica and erasure-coding storage policies for Veeam Cloud Tier. Generally, erasure coding is better suited for larger-scale deployments (for example, more than five nodes per region) and containers that store relatively large objects (for example, several megabytes). Replicas typically are better suited for smaller deployments (for example, three nodes per region) and containers that store relatively small objects (for example, less than 1 MB). At scale, erasure-coding policies provide better physical storage utilization.

 

Figure 49 shows a typical SwiftStack deployment for protecting less than 300 TB of data using triple-replica storage policy.

A screenshot of a cell phoneDescription automatically generated

Figure 49.     

Sample storage design for less than 300 TB of data

Figure 50 shows a typical SwiftStack deployment for protecting more than 300 TB of data using erasure-coding storage policy.

A screenshot of a cell phoneDescription automatically generated

Figure 50.     

Sample storage design for more than 300 TB of data

Container sharding management

Over time, Veeam can store many millions of objects in SwiftStack. To optimize performance with hundreds of millions of objects in a single bucket, SwiftStack uses an intelligent container sharding feature. When the SwiftStack controller identifies a bucket (or container) that is a candidate for sharding, it presents a notice to the administrator on the Cluster > Monitor > Sharding > Manage Large Containers page. To maintain optimal performance, if you see this notice, you should take the recommended action to let SwiftStack perform the sharding function.

Tuning parameters

Using the SwiftStack controller, tune the parameters to the values as shown in Table 8 to achieve the best performance in offloading data from the performance tier to the capacity tier. After these parameters are changed from their defaults to the values shown in the table, you need to submit the changes and push them to the cluster using the SwiftStack controller.

Tuning parameters can be accessed through the SwiftStack controller page (choose Manage >Tuning).

Table 8.        SwiftStack tuning parameters

Setting

Parameter

Default value

Change to

Account settings

Account replicator concurrency

8

2

Account replicator databases per second

50

2

Container settings

 

Container replicator concurrency

8

4

Container replicator databases per second

50

10

Object settings

 

Object server replication concurrency

36

2

Object server replication concurrency per device

4

1

Object replicator concurrency

4

2

Object replicator workers

12

1

Object replicator rsync max connections per disk

4

2

Object reconstructor workers

12

1

Object auditor files per second

20

2

Object auditor zero byte files per second

50

2

Object updater concurrency

8

2

Object updater workers

4

1

Performance characterization

Testing was performed to characterize the performance of the SwiftStack, Veeam, and Cisco solution. This section reports the results.

Backup to performance tier

A new backup job was run with 10 test virtual machines of 300 GB each with 100 GB of random data.

The scale-out backup repository was used with the following Advanced settings on the Storage screen (Figures 51 1nd 52):

      Inline data deduplication is disabled.

      The block size is set to the local target (1-MB block size).

      The compression level is set to optimal.

Related image, diagram or screenshot

Figure 51.        

New backup job

Related image, diagram or screenshot

Figure 52.     

Advanced settings for new backup job

The default settings for the backup job were used for the Guest Processing and Schedule screens.

Related image, diagram or screenshot

Figure 53.        

Backup job summary

The backup job was completed in 31 minutes and 25 seconds (Figure 54).

Related image, diagram or screenshot

Figure 54.     

Backup job statistics

After the full backup job was completed, another active full backup job was run (Figure 55).

 

Related image, diagram or screenshot

Figure 55.        

Active full backup job

There are now two full backups: one in an active backup chain and one in an inactive backup chain. You can verify this by navigating to Backups > Disk. In the Backup Properties window, you can see that each of the 10 virtual machines has a full backup followed by another full backup as the latest restore point (Figure 56).

Related image, diagram or screenshot

Figure 56.        

Multiple restore points for virtual machines

Immediate copy to capacity tier

As soon as the backup job is completed, the SOBR Offload to Capacity Tier function starts automatically because immediate copy is enabled.

The immediate copy to capacity tier operation was completed in 33 minutes and 35 seconds (Figure 57).

Related image, diagram or screenshot

Figure 57.        

Statistics of Immediate copy to capacity tier operation

Conclusions from the backup and SOBR offload tests

      Veeam v10 provides significant architectural improvements to speed up backup and archive performance compared to the earlier v9.5.4 release.

      Veeam’s default block size of 1 MB provides optimal performance throughput.

      A larger block size of 4 MB will speed up performance even more, but this may negatively affect other factors such as the deduplication ratio.

      The immediate copy function in v10 provides near-term disaster recovery capabilities, with data available on the object storage target as soon as the first backup operation is completed. Thus, if the performance tier goes down, organizations have a disaster recovery copy immediately.

Restore a virtual machine from the capacity tier

The virtual machine restore function enables you to re-create the same virtual machine in the VMware ESX environment and provides a recovery option in the event that the original virtual machine is deleted.

Restoration from the capacity tier was performed in Veeam by restoring a virtual machine to the Vmware infrastructure. The restore operation can be performed by overwriting the original virtual machine or moving the data to a new virtual machine as shown in the following steps:

1.     Navigate to the Home view in the navigation pane, choose Backups > Disk, and look at all the virtual machines to find the one for the backup job. Right-click one of the virtual machines in the list and select “Restore entire VM” (Figure 58).

Related image, diagram or screenshot

Figure 58.     

Restoring a virtual machine

A wizard will appear for performing a full virtual machine restore operation. For this test, a specific restore point was selected to which to restore the virtual machine.

2.     On the Restore Mode screen, select the option “Restore to new location, or with different settings” (Figure 59).

Related image, diagram or screenshot

Figure 59.        

Selecting the restore mode

3.     On the Folder screen, select the name to change the name of the restored virtual machine to a new name (Figure 60).

 

Related image, diagram or screenshot

Figure 60.        

Naming the restored virtual machine

4.     When the restore operation is completed, you can view the statistics screen that is displayed (Figure 61).

Related image, diagram or screenshot

Figure 61.     

Restore summary

Delete data from disk

A test was run to delete the data for a backup job run previously (Figure 62).

Related image, diagram or screenshot

Figure 62.        

Deleting a backup

The delete operation was completed in 6 minutes and 17 seconds (Figure 63).

Related image, diagram or screenshot

Figure 63.        

Delete operation summary

Conclusions from the delete test

Veeam v10 uses a continuation token API for faster delete operations. In addition, SwiftStack enables asynchronous deletion in its middleware. The combined effect of both capabilities is to make delete operations run significantly faster than in the v9.5.4 release.

Conclusion

Storing, managing, and accessing enterprise data is critical in today’s always-on global economy. Veeam, SwiftStack, and Cisco deliver a data storage, backup, and recovery solution that can simplify IT infrastructure, provide active access to data, maintain data integrity, align with compliance requirements, and dramatically reduce unnecessary storage costs.

The Veeam Cloud Tier with SwiftStack and Cisco solution’s “grow-as-you-go” approach offers complete flexibility for Cisco UCS hardware models, the number of sites in the deployment, and policy controls for the data sets. The modular design scales by the disk, by the server, or by the site. System expansion never disrupts the data service. The customizable policy engine creates a fine-tuned service model for data storage, backup, and recovery. All policy provisioning is completed through a point-and-click GUI. Fast backup and recovery performance uses multiple disks and network interfaces to accelerate the transfer of data to and from the capacity tier. Throughput performance increases as storage servers are added. Data is durably protected by being stored in many locations, using all available system components.

 


 

About the Authors

Paniraja Koppa, Cisco Systems, Inc.

Paniraja Koppa is a Technical Marketing Engineer for UCS Solutions. He has more than 13 years of experience with a primary focus data center technologies such as Cisco UCS, Storage, Operating systems, Automation, Virtualization and Cloud on. In his current role at Cisco Systems, he works on applications, solution development, best practices, optimization, automation and performance tuning of software defined storage on Cisco UCS platforms. Prior to this, he has led QA efforts for 4 new virtual adapter card’s firmware and software features for Cisco UCS.  He also worked as technical consulting engineer in the Data Center Virtualization space.

Anup Pal, SwiftStack Inc.

Anup Pal is a technical member of SwiftStack’s Solution Engineering team. He has more than 20 years of experience with primary focus on data storage software technologies such as Operating System, File System,  Data Protection Software, Virtualization, Cloud and Software Defined Storage. In his current role at SwiftStack, he is responsible for creating Configuration Guided, Solution Briefs and Best Practice documents for customer use cases specific to Analytics, Data Protection and Video Surveillance space. Prior to joining SwiftStack, he was a member of Solution Engineering team at Violin Memory and also members of Technical Staff at Quantum Corporation, Pillar Data etc.

Acknowledgements

For their support and contribution to the design, validation, and creation of this Cisco Validated Design, the authors would like to thank:

      Anil Dhiman, Cisco Systems, Inc.

      Chris O'Brien, Cisco Systems, Inc.

      Jawwad Memon, Cisco Systems, Inc.

      Oliver Walsdorf, Cisco Systems, Inc.

      Eric Pounds, SwiftStack

      Chris Lo, SwiftStack

 


 [OW(1]Make sure you talk about a WP and not a “CVD” guide

 [OW(2]7.6 or 7.7? You use different RHEL versions in your documentation

 [OW(3]I would add a note to strongly recommend upgrading to 9.3(2) according to https://tools.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-20200205-nxos-cdp-rce#fs

 [OW(4]Why do SwiftStack Replication and Cluster network look like the same? Is it technically needed?

 [OW(5]You’re talking above about 4.0(4e)?!

 [OW(6]Does SwiftStack use platform default values for the RAID group?

 [OW(7]I would at least say that you can use the same service profile template from above but need to change a few things there.

 [OW(8]Link?

Learn more