Website under construction

Scenario 1: One or more sectors on a drive has failed 
In this scenario, Storage Spaces will reallocate the extent that is affected by the failing sectors. The 
destination drive for the reallocation could be another drive in the same node or another drive in 
another node that does not already have a copy of the extent. So, if the three copies of the extent are 

---

59 
CHAPTER 2 | Software-defined datacenter 
on node A, B, and C, and the extent on node A is affected by a sector failure, the new copy can be 
generated on a different drive in node A or any drive in Node D. Drives in node B and C cannot be 
used, because these two nodes already have a copy of the extent. 
Scenario 2: A drive has failed 
In this scenario, Storage Spaces retires the physical drive from the storage pool when it discovers the 
drive has failed. After the physical drive has been retired, each virtual drive starts its repair process. 
Because the physical drive has been retired, the virtual drives generate a new copy of the extents that 
were on the retired physical drive. The new copies follow the same logic as in scenario 1. 
Scenario 3: A drive is missing 
In this scenario, Storage Spaces will do one of two things: 

If only the physical drive is missing, Storage Spaces will retire the disk. 

If the storage node or the physical enclosure to which the physical drive is attached is also 
missing, Storage Spaces will not retire the physical drive. 
The reason Storage Spaces won’t retire the physical drive in this second case is that during a storage 
node restart or temporary maintenance of a storage node, all the drives and physical enclosures 
associated with that node will be reported missing. Automatically retiring all of those drives and 
enclosures would potentially result in a massive amount of repair activity because you would need to 
rebuild all of the extents on those drives elsewhere in the storage system. This could easily be multiple 
terabytes of data. If the drives and enclosures are really missing and will not come back to the storage 
system, the administrator will need to retire the missing physical drives and start the repair process.