RAID 5 uses block-level striping, and the parity is spread across all drives. Any one drive can fail, at which point the RAID will run in a “degraded state” (slower than usual, and a very dangerous situation for the array) until a replacement drive is installed. It will still be available.
Many modern RAID controllers have the facility for hot-swapping and will automatically detect the failed / failing hard drive, drop it from the array, add the new, hot-swap drive and start rebuilding the array from the parity on the remaining disks.
However, by far the most RAID 5 recovery scenarious we encounter have been caused by data corruption, file system corruption or damage, or user error.
RAID rebuilds appear to be extremely common reasons for crashed RAID 5 systems; we believe that once one drive fails, if they were originally in the same batch, there is a better-than-usual chance another is going to fail shortly. The RAID rebuild failure is intensive, and places strain on the drives, making them more susceptible to failure and it is often during this process that we are contacted for data recovery.
Retrodata are able to recover RAID 5 arrays probably quicker than any other company. We have a perfect RAID recovery track record, maintaining a 100% successful RAID recovery service since June 2009. This covers RAID 0, RAID 1, RAID 3 – even a massive RAID 30 array in a storage area network consisting of 56 drives. Four of these drives had failed.
