One of the critical infrastructure components to any virtualization environment is storage. After all, no matter how much compute, memory, or network resources you have configured or available, without storage, you won’t have anywhere to store virtual machines, and ultimately, your data.
Microsoft’s Hyper-V hypervisor is a robust platform for running virtual infrastructure and provides a multitude of options for provisioning and connecting to storage infrastructure. As each version of Hyper-V has been introduced, there have been powerful new storage capabilities and options that have been introduced as well that improves the flexibility and options available to administrators when provisioning storage for Hyper-V.
In this post, we will take a look at connecting Hyper-V hosts to storage infrastructure. This will include a look at the various types of storage that can be connected to Hyper-V infrastructure.
Why Data is Important
Before taking a look at the various ways to connect Hyper-V hosts to storage infrastructure, let’s take a step back to being with and think about why data is important.
When you think about the most important reason that any IT infrastructure exists, it is to serve out or allow access to data of some sort. After all, customers, end-users, and others do not traverse the Internet or any business systems simply for the ride. There is always data at play somewhere in the chain of network communications or server resources that is either being accessed, sent, consumed, or saved. Data is the heart of IT infrastructure.
Data has been described as the “modern gold” of the Internet and future revenue. Businesses today are capturing massive amounts of data at scales that were unimaginable just a few years back. Due to the upsurge in machine-learning technology and capabilities, organizations can analyze and make useful extrapolations from the data that is collected, furthering business interests.
Without diving too much further down the rabbit hole of data collection and machine-learning, the bottom line with data is that it is crucially important. This is what is driving today’s businesses that are making use of technology.
Virtualization administrators must understand the nuances of how data is stored, accessed, best-provisioned, and backed up. Generally speaking, the other components of the virtualization infrastructure including compute and network serve the purpose of facilitating data access in some form or fashion. The storage infrastructure is the aspect of virtualized environments that is directly responsible for the data housed in the system. The hypervisor is responsible for brokering the access to and from the guests to the storage infrastructure.
Without further ado, let’s take a look at the different options for connecting Hyper-V hosts to the storage infrastructure and the advantages and disadvantages of each method and technology.
Connecting Hyper-V Hosts to Storage Infrastructure
There are various technologies that are driving the storage front when it comes to the Hyper-V hypervisor. Some of these have come about in recent versions. This is certainly the case on the front of software-defined storage.
Let’s take a look at the following ways of connecting Hyper-V hosts to storage infrastructure as well as some of the ancillary storage technologies that extend the capabilities of Hyper-V storage.
We will briefly look at the following:
- Direct Attached Storage
- Shared Storage
- Cluster Shared Volumes
- Storage Spaces Direct
Direct Attached Storage
Direct Attached Storage or DAS is probably the least desirable storage option in the case of Hyper-V as it means that the storage is not shared between hosts and you have a single point of failure.
***Note*** Storage spaces direct use direct-attached storage, however, it is “shared” between the other hosts in a Windows Server Failover Cluster running Hyper-V using a software-defined approach.
We will take a look at Storage Spaces Direct shortly.
Direct-attached storage can exist inside the server chassis or in an enclosure of some sort that is directly attached to the Hyper-V server host. Additionally, storage inside of DAS can be configured as Just a Bunch of Disks or JBOD where the drives are simply presented to the Hyper-V host directly. Or, the may be configured in a RAID array of some sort where parity information is calculated between the drives and data is able to be preserved in the event of a drive failure, based on this parity information.
Moving up into the realm of shared storage opens up a lot of new possibilities when it comes to the high-availability of your data and the mobility of your virtual workloads. To create a Windows Failover Cluster that runs Hyper-V, the Windows Failover Cluster needs to have shared storage for the cluster to work together in hosting production workloads. Basically, all of the advanced features and functionality of the Hyper-V hypervisor requires shared storage.
As mentioned, high-availability is one of the reasons you run a Windows Failover Cluster to begin with. The Windows Failover Cluster with shared storage running Hyper-V virtual machines is able to restart virtual machines on a healthy host if the host the VM was running on fails. Shared storage also unlocks the possibilities of performing maintenance without any downtime. VMs can be migrated from one host to another to perform patching or other host maintenance that requires taking the host down for a period of time.
The traditional approach to shared storage generally involves provisioning storage on a Storage Area Network (SAN) and attaching the storage to the Hyper-V hosts by using storage protocols such as iSCSI or NFS. The storage is provisioned on the storage appliance, mounted to the Hyper-V hosts, and multiple connections or multipathing is configured from all hosts to allow simultaneous access to the storage.
In order for Hyper-V hosts to all have simultaneous access to the storage volumes, a new type of shared volume was introduced in Windows Server 2008 R2 called Cluster Shared Volumes or CSVs.
Cluster Shared Volumes
Previous to Windows Server 2008 R2, simultaneous reads and writes from all Hyper-V hosts in the cluster for a particular volume was not possible. However, CSVs are a special kind of shared volume that can be formatted as either NTFS or ReFS (more detail on this later) that allows simultaneous read and write operations from all nodes in the Windows Failover Cluster.
CSV is a general-purpose, clustered file system that is layered above NTFS or ReFS that allows for clustered virtual machines and scale-out file shares to store application data. A CSV is able to get around the locks on the file system by utilizing metadata that is updated by a coordinator node in the Windows Failover Cluster that owns the LUN.
Storage Spaces Direct
Storage Spaces Direct is perhaps the most exciting development in Hyper-V storage in recent versions.
With Windows Server 2016, Storage Spaces Direct or S2D was introduced as Microsoft’s introduction into true software-defined storage running on commodity hardware.
As a direct competitor to VMware’s vSAN technology, S2D allows customers to run a software-defined storage architecture that provides tremendous flexibility in provisioning, management, and scalability.
Using Storage Spaces Direct, customers can provision shared storage across nodes in a Windows Failover Cluster running commodity hardware. This is configured using direct-attached storage in all the cluster nodes. Using a cache tier and a capacity tier, the software-defined storage is able to provision these tiers of storage to contribute to the overall software-defined storage location.
With Windows Server 2019, Microsoft has greatly improved the capabilities of Storage Spaces Direct and has introduced many new and exciting features. These include the ability to have ReFS and deduplication together at last, simplified architecture, “True two-node” cluster architecture utilizing a USB drive, and better visibility to drive health and performance.
ReFS or “Resilient File System” is another storage-related technology that is continuing to extend the features and capabilities of the file system when used in conjunction with Hyper-V. The resilient file system is a self-healing file system that is highly resilient to corruption. It also provides tremendous performance benefits with the “block cloning” capabilities that it contains. As the ReFS file system matures, it will no doubt be the most heavily adopted file system, especially when used with Windows Failover Clustering and Hyper-V.
Connecting Hyper-V Hosts to storage infrastructure is a crucial step in the overall virtual platform configuration. Data is the most important resource in IT infrastructure in general. With Microsoft Hyper-V, there are various capabilities and ways of configuring the storage infrastructure to fit the specific business needs as they arise. The many options that Hyper-V offers for storage provisioning, including DAS, shared storage, and Storage Spaces Direct, allows organizations to have tremendous flexibility in how their storage infrastructure is architected. With such a wide range of options, businesses can decide if they want to take a more traditional approach to provision shared storage or look to using software-defined technology like Storage Spaces Direct to provision storage for production workloads.Like what you read? Rate us