Server Fault is a question and answer site for system and network administrators. It only takes a minute to sign up.
Sign up to join this community
Our of RAID 0-5 and RAID 0+1 and RAID 1+0 which ones would be able to handle two disk failures at the same time?
I have read online that RAID 0+1 can not handle this
Also from my research it seems that RAID 1, 2, and 0+1 can handle this but I am really not sure.
This is very complex because many RAID levels can handle two disk failures but not ANY two disk failures. RAID 6 is the only I know of that can handle ANY two disk failures. RAID 10 (mirrors that are striped) can handle two disk failures as long as the disks are in different RAID 1 arrays. The same goes for RAID 50 (RAID 5 arrays that are striped). RAID 0+1 (stripes that are mirrored) can also handle two disk failures but only if the disk failures are in the same RAID 0 array (opposite of RAID 10 and 50). It all depends on the cost benefit analysis for your needs. And always remember that RAID is fault tolerance and is no substitute for backups and a plan for Disaster Recovery (DR) and Business Continuity (BC).
RAID-6 is the canonical answer, because it can handle exactly two disk failures. However, to cover your larger question, RAID1 can (maybe) handle two disk failures -- technically, an N disk RAID-1 array can handle N-1 failures, so whether a given RAID-1 will handle two failures depends on your configuration.
RAID-10 (or RAID-01, depending on your proclivities) can handle anywhere between M-1 and N/M failures (where N is the number of disks in the array, and M is the number of mirrors), depending on which disks happen to fail. If the M disks that fail are all the mirrors of the same data, then you're toast. On the other hand, if the N/M disks that fail are all mirrors of different data, you're OK. I'll leave the probabalistic analysis of what the chances of catastrophic failure for given values of N and M for you to do if you're so interested.
RAID 1+0 of (2xN) drives can handle up to N failed drives _as long as_ no 2 drives belonging to the same pair fail at the same time.
An illustration:
+------------------RAID 0----------------------+
| +---RAID 1---+ +---RAID 1---+ +---RAID 1---+ |
| | D0-1 D0-2 | | D1-1 D1-2 | | D2-1 D2-2 | |
| +------------+ +------------+ +------------+ |
+----------------------------------------------+
We have 2x3 drives. If drives D0-1, D1-2, and D2-1 fail at the same time, the whole array still survives. But if ever D0-1 & D0-2 (or any two drives in the same RAID 1 pair) fail at the same time, you'll lose the whole array.
RAID6
http://en.wikipedia.org/wiki/Standard_RAID_levels#RAID_6
According to the Storage Networking Industry Association (SNIA), the definition of RAID 6 is: >"Any form of RAID that can continue to execute read and write requests to all of a RAID >array's virtual disks in the presence of any two concurrent disk failures."
Any RAID level doesn't replace RAID Backup(locate it in separate building). For example, buggy software/driver writing randomly trash to disk, hackers/viruses (etc.) can trash your whole array at once .. For me, ENOUGH DATA SAFETY = backups + ((RAID1 or any redundant) + hotspare(s)) + RAID Monitoring.
