Microsoft Storage Spaces automatically rebuilds VDs by using free
disk space and does not require assigning hot-spares. If a physical
disk fails, the failed physical disk is regenerated from free storage
disk space without user’s intervention. The rebuild begins immediately
after a disk failure is detected, limiting the exposure to additional
physical disk failure while the VD is running in a degraded state.
Dell recommends that you enable automatic rebuilds.
Run the following PowerShell command to enable automatic rebuilds
(configured for each pool).
Set-StoragePool –FriendlyName <poolName> -RetireMissingPhysicalDisks Always
-
NOTE: Before performing
maintenance operations, and when storage pools are configured to automatic
rebuild, prior to turning off a storage enclosure, you must change
the
RetireMissingPhysicalDisks attribute of all affected
storage pools to
Never. When the network communication with
the drive becomes unavailable because the storage enclosure was turned
off for maintenance, changing the attributes prevents the VDs from
immediately initiating rebuilds. Ensure that you change
RetireMissingPhysicalDisks back to the original value after the maintenance operations completes.
When a storage pool is configured for automatic rebuilds,
it is important to maintain sufficient free disk space in each pool
to enable VDs to automatically rebuild in the event of a physical
disk failure. Sufficient free disk space is equal to the capacity
of one HDD plus an additional 8 GB (for storage spaces overhead) multiplied
by the number of disk failures to automatically recover from the storage
pool’s HDD tier, and the capacity of one SSD plus an additional 8
GB (for storage spaces overhead) multiplied by the number of disk
failures to automatically recover from the SSD tier. Equations for
sizing the HDD and SSD tiers are specified in the Storage Tiers section
in this document.
There are two types
of rebuild methods — parallel and sequential. During a parallel rebuild
process, the data required to rebuild is obtained from multiple physical
disks in the pool. Parallel rebuilds are very fast and reduce the
time a VD is in a degraded state. However, because of multiple physical
disks sourcing the repair data, there is some impact to storage I/O
performance during rebuilds. A sequential rebuild only sources repair
data from one physical disk in the pool at a time. This means that
there is less impact to storage I/O performance during rebuilds. However,
the rebuild process takes longer.
Dell recommends parallel rebuilds because it provides the fastest
method to ensure all VDs return to their full resiliency. To optimize
VD repairs the Microsoft November 2014 KB rollup (KB3000850) is required
on all storage and compute nodes. Also, you must perform the tasks
listed in the following URL:
https://technet.microsoft.com/en-us/library/dn858079.aspx
Run one of the following PowerShell commands
for configuring the rebuild method (configured for each pool).
Set-StoragePool –FriendlyName <poolName> –RepairPolicy Parallel
Set-StoragePool –FriendlyName <poolName> –RepairPolicy Sequential