For a new EC2 file server, you want RAID fault-tolerance for 1 TB of data. Which option should you choose?
1. RAID5 with two EBS volumes of not less than 1 TB each.
2. RAID0 with two EBS volumes of not less than 500 GB each.
3. RAID1 with two EBS volumes of not less than 500 GB each.
4. RAID1 with two EBS volumes of not less than 1 TB each.
5. RAID0 with two EBS volumes of not less than 1 TB each.
Answer 4 is correct.
RAID0 offers no drive fault-tolerance. RAID1, also known as mirroring, requires 2x the required volume space. RAID5 requires 3 volumes at a minimum. Further information:
Q: Is redundancy provided by RAID1 not intrinsic to EBS itself and therefore only RAID0 makes sense?
2 Answers
Option 4 makes sense it offers fault tolerance for the whole 1TB dataset instead of option 2 which wouldn’t make sense to utilize 500GB EBS volumes to provide redundancy to 1TB of data (it would not fit). RAID0 would not make sense "There is no redundancy in this level. The failure of any one disk in a volume will absolutely result in the loss of all data on the volume (unless the failed disk is fixed and the blocks on it are recovered)." Link below.
https://raid.wiki.kernel.org/index.php/What_is_RAID_and_why_should_you_want_it%3F
There is certainly a fair amount of redundancy built into EBS. It is still possible though to completely lose an EBS volume.
From the EBS features page: "Amazon EBS volumes are designed for an annual failure rate (AFR) of between 0.1% – 0.2%, where failure refers to a complete or partial loss of the volume, depending on the size and performance of the volume."
Agree. And if you use RAID1 you can minimise those numbers even more, i.e. to 0.0001% – 0.0004%
Why would you need Raid1 if EBS volumes are already being replicated within an Availability Zone (AZ)?
Because as Werner says frequently "Everything fails all the time."