With 6 drives in RAID 10 one will get 3 drives effective storage and 3 redundant drives, so in a best case scenario up to 3 drives can fail without loosing any data. But in a worst case scenario (now correct me if I'm wrong) even two failed disks could make the whole array fail (if those drives are mirrored onto each other, like in this array).

Is there a way around this possible failing state? I am liberal regarding software RAID and non-standard RAID setups, so maybe there is some way? Or it could just be me having bad luck thinking. If so - please explain to me how things really work. :)

  • It's a bit broad a question since there are many ways to build raids, though you are correct in your assumption. Consider reading: en.m.wikipedia.org/wiki/RAID – Mattias Åslund Dec 12 '13 at 18:38
  • Are you trying to avoid two-drive failures bringing down the whole array? If so, RAID6 is what you are looking for. Or are you looking to survive failure of any three drives? – ChrisInEdmonton Dec 12 '13 at 18:41
  • RAID 6 would be an option, but it would be really nice to have the speed of RAID 10, why I wanted to do some research on this first. But RAID 6 (or RAID Z2 even) maybe will be my best alternative? – lindhe Dec 12 '13 at 18:45
  • ChrisInEdmonton: No, I will settle at any two drives. I think I will go for RAID 6 (or more likely RAID Z2, since I don't want to invest in a RAID card). RAID Z2 have greater fault tolerance than 10, and since this will be on a NAS on a Gigabit network I won't be able to utilize of all that extra speed from RAID 10. – lindhe Dec 12 '13 at 21:19
  • If you are doing Linux software RAID10, which does a lot of weird things above and beyond a standard RAID10, you can setup n mirrors so long as your number of disks is at least n+1. By that I mean you could have N copies of every chunk instead of just the standard 2 copies. The more copies you have the less storage space you will have available. – Zoredache Dec 13 '13 at 0:50

You're over-thinking this. Regardless of RAID setup, they are not infallible. They are there to increase the odds that a disk failure (or perhaps even two) won't cause the system to crash, but can never totally eliminate the fact that drives fail.

Think of it this way: "What if my RAID Controller fails?" -- You're down, regardless of how many drives or what RAID configuration you chose.

RAID exists (primarily) to help keep up-time (and perhaps help with disk access speeds), Backup is there to prevent you from losing data. The two are NOT the same thing, and you should have both(unless uptime isn't a big deal, then just make backups).

  • That is true. But this is the system that I do my backups to, therefore I concentrate primarily on making the data more "secure" rather than backing up the data. One could think of redundancy as creating disk with much lesser probability of failing, which would be to prefer as a backup solution compared to a drive with greater risk of failure. – lindhe Dec 12 '13 at 21:12
  • @Lindhea "One could think of redundancy as creating disk with much lesser probability of failing", One SHOULD think of it like that. :) Once shouldn't think that there is a magic RAID level that will protect against any drive failure regardless of the number of failed drives or the arrangement they are in. :) Keep in mind to if you are backing up to a backup server, then you have TWO copies of it (the backup and the original). – Ƭᴇcʜιᴇ007 Dec 12 '13 at 22:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.