RAID on large drives for small files. It makes sense?

Question

RAID

Good afternoon!
I ask to clarify the moment:
task: fast work with tempDB sql2016.
we have: ESXi 6.5, 256Gb ddr-4, 200GB DB, 8x650GB SAS3.0 15k RPM, controller with raid 0-50 support.

Question:
If the file size is smaller than the disk size in the array, will the read/write use the entire array bandwidth, or just the bandwidth of one disk?

Does it make sense to lose the amount of available space by building raid 0 or 5 under a virtual disk for tempDB??

Answer 1

Do you have a productive SQL server?

Answer 2

File size does not affect bandwidth in any way. You can write to a small file so that no band is enough, or you can never use a large file.
Make raid1 for tempDB and forget this question.
Why raid1? - you don't want to stop all work when 1 disk fails, and mssql uses tempdb quite actively and will not work without it. In addition, the reading speed will increase.
raid 5 is now out of fashion, because. with modern disk volumes, raid5 recovery takes too long, and this will affect the speed of mssql.

Answer 3

thanks for the points of view!
I decided the following: test the current config, then a new one with tempDB placed on one spindle, on raid0 from a pair, then from three disks, on one SSD.
choose according to the results.
Once again, thank you all