highest performing storage when you have more than one type of storage available.
Capacity uses parity resiliency, and Performance uses mirror resiliency. In both cases,
you have two-disk redundancy. Details of the storage pool and tiers created can be
seen with the following PowerShell:
Get-StoragePool <name, e.g. S2D* if default> | FT FriendlyName, FaultDomainAwarenessDefault, OperationalStatus, HealthStatus
-autosize
Get-StorageTier | FT FriendlyName, ResiliencySettingName, MediaType, `
PhysicalDiskRedundancy -autosize[
Once enabled, create a volume. For example:
New-Volume -StoragePoolFriendlyName "
Capacity,Performance `
-StorageTierSizes
Storage Spaces Direct is available only in the Datacenter SKU of Windows Server 2016,
whereas regular Storage Spaces is still available in the Standard SKU.
Storage Replica
Although not a feature of Storage Spaces nor Hyper-V specific, Storage Replica is a
Windows Server storage feature that brings great value to Hyper-V environments.
Storage Spaces Direct provides resilient, highly available storage within a datacenter
where the nodes have fast, low-latency connectivity.
Disk or node failure, however, is only one type of disaster. Consider the many natural
disasters in recent years that have brought widespread flooding to many areas. Even
organizations with well-architected datacenters with redundant generators have
suffered multiday outages as a result of these natural disasters. For this reason, many
organizations implement disaster-recovery facilities and plans that are used when an
entire facility is compromised. If you need to fail your workloads over to another
datacenter that is at least 50 miles away (likely more) to ensure that it’s not impacted
by the same disaster that impacted your primary datacenter, you need a way to have
your workloads and data replicated.
Numerous replication solutions can operate at different levels, such as application
replication (SQL Server AlwaysOn, Active Directory multimaster replication),
hypervisor replication (Hyper-V Replica), in-OS replication (Azure Site Recovery for
VMware or physical systems), and storage replication (traditionally provided by SAN
vendors or third parties). Typically, the best solution is the one that runs at the
application level; because the application is aware of the replication, the failover and
can take appropriate steps. Many times, however, this is not available or practical. This
is a Hyper-V book, but at times Hyper-V Replica may not meet your requirements
either, because it is asynchronous and the architecture calls for a stretched cluster
with automatic failover, or perhaps the 30-second minimum interval is too long.