Q. Why should stripped volumes not be used in Azure with GRS storage accounts?
A. A standard performance disk in Azure provides 500 IOPS (when connected to a standard tier VM) and can be up to 1023 GB in size. If you want a larger disk or require more IOPS per disk (another option for more IOPS would be using Premium Storage with a xS series VM) it is common to add multiple disks to a VM then merge them together within the guest OS using a technology such as Storage Spaces. If you are using a Geo-Redundant Storage (GRS) account then it is not recommended to aggregate separate disks into a single disk within the guest OS. The reason is the with GRS there are three copies of the data at the local region and another three copies at the paired Azure region 100's of miles away. The replication between the regions is asynchronous with each of the separate data disks that make up the aggregated disk inside the guest OS likely replicating at slightly different offsets. This means in the event of a region failover the disks in the remote region would be out of step with each other and the aggregated disk corrupt and unusable. Therefore if you need larger disks use an application level replication technology between regions or a volume level replication technology in the guest such as Storage Replica in Windows Server 2016.