Safety of RAID Z1 and URE (HDD reliability)

[ku-gew]

It is said often that RAID Z1 are no more reliable because of the URE rates given by manufacturers (1E-14 URE/bit read is common).
Typical source of this statement:
http://www.zdnet.com/blog/storage/why-r ... n-2009/162

It appears that value has been lowered for marketing purposes, to sell higher class drives:
http://www.high-rely.com/hr_66/blog/why ... -2009-not/

Therefore, RAID Z1 is more reliable than we thought and it's not unreasonable for a home-level NAS4free to use RAID Z1 if disk prices are SO important (but are they? HDDs are cheap and RAID 1 has better performances where it matters).

Just for info.

[b0ssman]

the problem with a 1 drive redundancy is that if one drive fails and you are rebuilding the following can happen.
- because of the strain of rebuilding, another drive can fail see most recent example viewtopic.php?f=66&t=6403
- any corruption that has gone unnoticed can not be repaired anymore because there is no data to reconstruct it from.

[ku-gew]

I know the issues of 1 drive redundancy, it's just a matter of probabilities and I wanted to point out that the URE probabilities are much lower than expected. Of course mechanical issues are still there.
Interesting with calculations:
http://superuser.com/questions/516949/f ... d-rebuild/

I chose RAID10 for my main server with disks bought from different batches and at months of distance (the ones bought early have been put to use immediately: the wear is different within each pair).