FTT vs URE

[Randy_Evans]

Has anyone worked out the risk of FTT=1 versus FTT=2?

With today's disk capacities it has become evident that RAID 5 has a high likelihood of being unable to rebuild after a drive failure because an unrecoverable read error (URE) occurs on the remaining drives. RAID 6 has become best practice for vSphere storage where there are many eggs in one basket. Although RAID 5 tolerates the loss of a single drive, a URE during rebuild can result in data loss. The ability of RAID 6 to tolerate two drive failures significantly reduces the risk of data loss.

A simplistic analogy could compare RAID 5 and RAID 6 with FTT=1 and FTT=2.

Both RAID 5 and FTT=1 can tolerate only one failure. Is FTT=1 at high risk of data loss due to a URE during rebuild?

Both RAID 6 and FTT=2 can tolerate two failures. Does FTT=2 significantly reduce the risk of data loss?

For production systems, is FTT=2 the best practice for keep the risk of data loss low?

Looking forward to your thoughts. thanks.

[depping]

I have had this question a couple of times now and it is difficult to answer to be honest. FTT=1 is what we expect most customers to use to be honest, in most of my discussions with customers they seem to indicate that the majority of workloads will use FTT=1. I can see however that for more mission critical systems FTT=2 would be desired, but there is a cost associated with it and that is the extra component you end up storing.