Best Practices Dell EMC SC Series: Microsoft Hyper-V Best Practices Abstract This document provides best practices for configuring Microsoft® Hyper-V® to perform optimally with Dell EMC™ SC Series storage.
Revisions Revisions Date Description June 2009 Initial release August 2009 Updated for Windows Server 2008 R2 September 2011 Updated for Windows Server 2008 R2 SP1 and SCOS 5.5 October 2012 Updated for Windows Server 2012, Enterprise Manager 6.2, and SCOS 6.2 December 2012 Added support for virtual Fibre Channel with SCOS 6.3 and Enterprise Manager 6.
Table of contents Table of contents Revisions.............................................................................................................................................................................2 Acknowledgements .............................................................................................................................................................2 Table of contents .................................................................................................
Table of contents 2.17 SC Series data reduction and Hyper-V ............................................................................................................26 3 SC Series snapshots and Hyper-V ............................................................................................................................27 3.1 SC Series Replay Manager support for Hyper-V .............................................................................................27 3.
Executive summary Executive summary Dell EMC™ SC Series storage provides a powerful and complete set of storage integrations, and management and monitoring tools for Microsoft® Windows Server® environments. This document provides best practice guidance for deploying and optimizing the Windows Server Hyper-V® role with SC Series arrays. The documentation at Dell.
Introduction 1 Introduction Microsoft Windows Server Hyper-V and SC Series storage are feature-rich solutions that together present a diverse range of configuration options to solve key business objectives such as storage capacity, workload optimization, performance, and resiliency. 1.
Introduction Install the Hyper-V role with the Add Roles and Features Wizard Initially offered with Windows Server 2008, Hyper-V has matured with each release to include many new features and enhancements. It has evolved to become a mature, robust, proven virtualization platform. In simplest terms, it is a layer of software that presents physical host server hardware resources in an optimized and virtualized manner to guest virtual machines (VMs) and their workloads.
Introduction 1.4 Best practices overview Best practices are derived from the collective wisdom and experience of developers and end users over time, and this knowledge is built into the design of next-generation products. With mature technologies such as Hyper-V and Dell EMC storage arrays, default configurations typically incorporate best practices.
Optimize Hyper-V for SC Series 2 Optimize Hyper-V for SC Series SC Series storage is an excellent choice for external storage for stand-alone or clustered Windows Servers including servers configured with the Hyper-V role. Core SC Series features such as thin provisioning, Data Progression, data reduction, snapshots (Replays), and replication work seamlessly in the background regardless of the platform or OS.
Optimize Hyper-V for SC Series Aside from performance problems, one of the key indications that integration services are outdated or not present on a Windows VM is the presence of unknown devices in Device Manager for the VM as shown in Figure 3. Unknown devices listed for a guest VM indicates missing or outdated integration services For versions of Hyper-V prior to 2016, use Hyper-V Manager to connect to a VM.
Optimize Hyper-V for SC Series To verify the version of integration services for a VM, click the Summary tab in Failover Cluster Manager. Verify integration services version with Failover Cluster Manager Verification can also be performed using PowerShell, as shown in the following example: PS C:\Windows\system32> get-VM | Select-Object name, integrationservicesversion Name IntegrationServicesVersion ---------------------------MG-VM12a 6.3.9600.18080 MG-VM12b 6.3.9600.18080 MG-VM12c 6.3.9600.
Optimize Hyper-V for SC Series require the boot disk to use a virtual IDE disk controller. Generation 2 guests instead use virtual SCSI controllers for all disks. Virtual IDE is not a supported option with generation 2 VMs. Specify a guest as generation 1 or generation 2 2.2.1 Convert VMs to a newer generation The warning message in the wizard in Figure 6 indicates that the VM generation cannot be changed once a VM has been created. However, third-party tools are available to convert VMs.
Optimize Hyper-V for SC Series 2.3 Virtual hard disks A virtual hard disk (VHD) is a set of data blocks that is stored as a regular Windows file on the host system. VHD files end with a .vhd, .vhdx or .vhds extension depending on the type of VHD. All VHD formats are supported with SC Series storage. Virtual hard disk file (vhdx) on a cluster shared volume 2.3.
Optimize Hyper-V for SC Series 2.3.2 Virtual hard disk type In addition to the format, a VHD can be designated as fixed, dynamically expanding, or differencing. Since SC Series storage leverages thin provisioning, only the data that is written to a virtual hard disk, regardless of the disk type, will consume space on the storage array. As a result, determining the best disk type is more a function of workload requirements rather than how it will impact storage utilization on SC Series storage.
Optimize Hyper-V for SC Series SC Series storage supports any of the VHD types. From the perspective of the host server, there are some best practice performance and management considerations to keep in mind when choosing the right kind of VHD type for your environment. • Fixed-size VHDs: - - • Dynamically expanding VHDs: - • Recommended for workloads with a high level of disk activity, such as SQL Server®, Microsoft Exchange, or OS page or swap files.
Optimize Hyper-V for SC Series • • Longer chains of differencing virtual hard disks are more likely to negatively impact read performance. It is therefore a best practice to keep native Hyper-V based checkpoints to a minimum if they are used. Administrators can leverage array-based SC Series snapshots to replicate data to other SC Series arrays for archive or recovery of Hyper-V guest VMs and workloads to avoid the use of native HyperV checkpoints. - 2.3.
Optimize Hyper-V for SC Series 2.3.4 Overprovisioning with dynamic virtual hard disks When using dynamic virtual hard disks and thin provisioning, there is an inherent risk of either the host volume or a storage pool on the SC Series array running out of space. See Figure 11 for an example. If the dynamic disk used by VM2 on the host volume expanded far enough, it would fill up the host volume and negatively impact VM1 and VM2.
Optimize Hyper-V for SC Series Present SC Series volumes directly to guest VMs: - In-guest iSCSI Virtual Fibre Channel (vFC) with Windows Server 2012 Hyper-V and newer > • Present SC Series volumes indirectly to guest VMs as pass-through disks in Hyper-V - 2.5 SC Series Replay Manager offers limited support for protecting guest VMs using vFC. Use of pass-through disks is a legacy configuration introduced with Windows Server 2008 HyperV.
Optimize Hyper-V for SC Series 2.6 MPIO best practices The Windows Server operating system and Hyper-V (2008 and newer) natively support MPIO by the built in Device Specific Module (DSM) provided by Microsoft. SC Series Storage supports the use of this DSM when implementing MPIO. It is very important to adjust MPIO timeout settings for your Hyper-V environment (both hosts and VMs) to avoid service outages when performing routine SAN maintenance. This includes hosts that are configured to use single path.
Optimize Hyper-V for SC Series Use VM Settings in Hyper-V Manager to add a virtual Fibre Channel adapter to a guest VM 2.8 Guest VMs and direct attached storage Although SC Series arrays support in-guest iSCSI and virtual FC disks mapped to guest VMs, direct-attached storage for guest VMs is generally not recommended as a best practice unless there is a specific use case that requires it.
Optimize Hyper-V for SC Series Limitations when using direct-attached storage for guest VMs include: • • • • • • Checkpoints: The ability to perform native Hyper-V checkpoints is lost. However, the ability to leverage SC Series snapshots is unaffected. Simplicity: More complicated and therefore requires more management overhead. This is especially true for vFC. Mobility: VM Mobility is reduced due to creating a physical hardware layer dependency.
Optimize Hyper-V for SC Series feature enhancements offered with newer releases of Hyper-V (generation 2 guest VMs, VHDX format, and shared VHDs in Windows Server 2016 Hyper-V and newer). Use cases for pass-through disks are like those for direct-attached storage (see section 2.8). Although SC Series Replay Manager supports protecting pass-through disks, use of in-guest iSCSI is recommended, since in-guest iSCSI is required to perform a recovery.
Optimize Hyper-V for SC Series Map storage to a cluster object to ensure consistent LUN numbers on all nodes 2.11 SC Series LUN limits for larger Hyper-V clusters For large Hyper-V clusters with many nodes, there are two LUN limits to be aware of: The functional LUN limit: Although the SC Series storage supports up to 254 LUNs per cluster server object, resources on Hyper-V server nodes might be consumed before the physical limit of 254 LUNs reached.
Optimize Hyper-V for SC Series 2.12 Volume design considerations for SC Series One of the design considerations for which there are often no clear answers is how many guest VMs to place on an SC Series volume, including cluster shared volumes (CSV). In most cases, a many-to-one strategy is a good starting point, and adjust for specific uses cases. Some advantages for a many-to-one strategy include the following: • • • Simplicity: Fewer SC Series volumes to create and administer (avoid volume sprawl).
Optimize Hyper-V for SC Series 2.14 Disable automount To prevent a Hyper-V host server from automatically assigning drive letters to newly mapped volumes, disable the automount feature, which is enabled by default. Having automount disabled is also beneficial when recovering a volume using an SC Series snapshot. Disable the automount feature To verify or change a server automount configuration, open a command prompt window and run Diskpart.
Optimize Hyper-V for SC Series 2.16 Placement of Active Directory domain controllers It is a best practice to avoid placing all the Microsoft Active Directory® (AD) domain controllers for a domain on the same Hyper-V cluster (as VMs) if the cluster service requires AD authentication to start. If the cluster goes off line (along with all the domain controller VMs), the cluster service be will not be able start.
SC Series snapshots and Hyper-V 3 SC Series snapshots and Hyper-V SC Series snapshots are space-efficient, meaning they consist of pointers to frozen data blocks, and therefore consume no additional space unless, for example, a View Volume is created from a snapshot, mapped to a host and new data is written. SC Series snapshots can be taken of volumes mapped as LUNs to a Hyper-V environment regardless of content.
SC Series snapshots and Hyper-V 3.2.1 Recover a guest VM on a standalone Hyper-V host In this scenario, the virtual hard disk and configuration files for a VM reside on a data volume that is mapped to a Hyper-V host. If a VM virtual hard disk and configuration files reside on separate host data volumes, then it is a best practice to configure a consistency group for these volumes on the SC Series array so that crash-consistent snapshots occur at the same exact time.
SC Series snapshots and Hyper-V 3.2.2 Recover a guest VM on a cluster shared volume The process of using SC Series snapshots to recover guest VMs that reside on a cluster shared volume (CSV) is like the process of recovering a guest VM to standalone host, as detailed in the preceding section. However, recovering a VM from a CSV may require changing the disk signature first. Windows servers assign each volume a unique disk ID (or signature).
SC Series snapshots and Hyper-V 7. From the Diskpart command prompt, type rescan and press Enter. 8. Use Disk Management on the host server to bring the View Volume online. 9. Return to the Diskpart command prompt window and type list disk and press Enter. 10. The new disk (Disk 3 in this example) should now be listed. Usually, the bottom disk will be the one just added. 11. Type select disk # (where # represents the number of the new disk, in this example, disk 3) and then press Enter. 12.
SC Series snapshots and Hyper-V 15. Type uniqueid disk again and press Enter to verify the new ID. 16. Now that the View Volume has a new signature, it can be unmapped from the standalone host server and mapped to the cluster without causing a disk ID conflict. 17. Recover the guest VM. 3.
SC Series snapshots and Hyper-V 3.5 Leverage SC Series to create gold images With SC Series storage, an administrator can create gold images to accelerate and simplify the process of deploying new servers.
SC Series snapshots and Hyper-V 4. Power on the server and run Sysprep, choosing the Generalize, Out-of-box Experience, and Shutdown options. 5. Once the server is powered down (which will ensure that it is in a consistent state), manually create another SC Series snapshot of the volume and set it to never expire. Assign it a descriptive name that clearly identifies it as a gold source. 6. Using this snapshot as the gold source, create a View Volume and map it to the desired host server. 7.
SC Series snapshots and Hyper-V 5. Create a new VM. Make a copy of the gold VHD and place it in the desired location to serve as the boot volume for the VM. 6. Rename the VHD to reflect the name of the VM or its purpose, and then attach the VHD to the new VM as the boot volume. 7. Power on the VM and customize the VM as needed. Note: Although this is not as space efficient, it is a very quick and easy way to provision new VMs.
SC Series snapshots and Hyper-V In the example below, array SC18 is comprised of two controllers: SN 716 and SN 717. Two new server hosts named TSSRV200 and TSSRV201 have been provisioned from a boot-from-SAN gold image source volume that is owned by controller 717. As a result, the View Volumes created for hosts TSSRV200 and TSSRV201 are also owned by controller 717.
SC Series snapshots and Hyper-V As a best practice, add the controller SN to the name of the gold image volume. Add the controller serial number to the volume name As new servers are deployed from View Volumes, alternate between gold image source volumes so that each controller head stays balanced with roughly the same number of volumes. 3.
Data Progression and Hyper-V 4 Data Progression and Hyper-V Data Progression is a core SC Series feature. While this feature is most commonly described in terms of how it can optimize SC Series arrays that are comprised of multiple disk tiers, it will also optimize data placement on SC Series arrays comprised of a single disk tier. With Data Progression, data is automatically and intelligently placed in the optimal storage tier and RAID level based on usage and performance metrics.
Data Progression and Hyper-V 4.1.1 Data Progression with archival data In this example, a Hyper-V VM workload creates a large amount of archival data, such as image files or video. This data is stored on separate virtual hard disks on one or more dedicated CSVs, and once this data has been written to disk, it is infrequently accessed. • • • Option 1: Leverage the Recommended (All Tiers) storage profile.
Data Progression and Hyper-V SC Series arrays do provide some protections against this scenario. For example, when replicating a volume from one SC Series array to another, by default the replicated data is placed into the lowest tier. However, if performing a file level copy in Hyper-V at the host or VM level, new writes will occur in the highest tier allowed by the storage profile that is applied to the target volume, which will usually be the Recommended (All Tiers) profile.
Disk space recovery with Hyper-V 5 Disk space recovery with Hyper-V In Microsoft Windows, deleting a file just removes the file pointer, not the actual data itself. The operating system will report this deleted space as free to be overwritten by new data. However, there are situations where the SC Series will not return the deleted space to the SC Series array page pool to be used elsewhere. This can result in inefficient use of the SAN storage capacity.
Disk space recovery with Hyper-V • Space recovery from a pass-through disk is supported only if the following conditions are met: - • 41 Server Agent is installed on the guest VM. The pass-through disk is presented to the guest as a virtual SCSI device (virtual IDE will not support disk space recovery). The pass-through disk is a basic disk formatted as an NTFS volume.
Boot-from-SAN for Hyper-V 6 Boot-from-SAN for Hyper-V SC Series storage supports boot-from-SAN when hosts are configured with FC or iSCSI cards that also support boot-from SAN. In Microsoft environments, boot-from-SAN is supported with standalone and clustered Hyper-V hosts and nodes. Some of the pros and cons of booting from local disk or from a SAN are detailed below. Boot-from-SAN advantages: • • • SC Series snapshots of boot-from-SAN volumes provide for quick recovery.
PowerShell integration 7 PowerShell integration SC Series storage has incorporated Windows PowerShell for many years and now supports a large library of cmdlets. The SC Series PowerShell SDK command set provides administrators with the ability to perform many SC Series tasks from the command line and create automated scripts. To learn more about PowerShell integration with SC Series storage including many examples, see the Dell Storage PowerShell SDK Cookbook. 7.
PowerShell integration leveraging PowerShell, many of these manual steps could be automated making it easier to meet recovery time objectives (RTO). 7.3 Best practices for PowerShell GUI interfaces for the most part keep administrators safe by guiding their steps with wizards that provide warning messages that help protect against inadvertently executing commands that are destructive. The trade-off is that GUI functionality is sometimes limited.
Business continuity with Hyper-V and SC Series 8 Business continuity with Hyper-V and SC Series A good business continuity strategy will always incorporate disaster recovery and disaster avoidance planning. At a high level, a disaster recovery plan is a process where a company ensures they can recover as quickly as possible from data loss or from an interruption or failure that prevents access to data.
Business continuity with Hyper-V and SC Series 8.2 Disaster recovery and disaster avoidance Disaster recovery usually means reacting to an event that causes down time that takes place unexpectedly with little or no warning. These events can be categorized as follows: • • • Events that cause data loss such as malware infection, corruption, accidental deletion, sabotage, or hardware failure of disks or disk arrays.
Business continuity with Hyper-V and SC Series workload, are gracefully paused before snapshots are taken. This is especially important when protecting a transactional workload such as a database, to help ensure recovery without errors or data corruption. For more information about Replay Manager for Hyper-V see the Dell EMC SC Series Replay Manager and Microsoft Hyper-V best practices guide and the Dell EMC SC Series Replay Manager 7 for Hyper-V demo video.
Technical support and resources A Technical support and resources Dell.com/support is focused on meeting customer needs with proven services and support. Storage technical documents and videos provide expertise that helps to ensure customer success on Dell EMC storage platforms. A.