I never trust a single hard drive with anything. All of my important business systems use RAID, mirroring, and sometimes even replication to protect against common hard drive failures. Every hard drive is going to fail. Even solid pass it exam state drives are not immune to failure. In fact, I am not yet convinced that SSDs are any more reliable than traditional spinning hard drives. None of this keeps me awake at night. 250-351 Any hard drive can fail and I don’t lose a bit of data. Even with this protection, I still backup all important data because there are other failure scenarios that RAID, and mirroring, and replication don’t help with.
We had a database server with high-end HP RAID system attached. There were two physical arrays, one for the database, and one for the logs. Each array was RAID AVG (AKA RAID 6), which allows up to two hard drives to fail simultaneously without losing any data. In addition the arrays had hot spare drives
A30-327 that are activated if any drive fails. We were also using the highest quality enterprise SCSI drives. Sounds like we Parajumper Jakke e had it covered, right?
That database server went down when the RAID controller failed. We had no access to any of the data, even though every drive was still healthy. We couldn’t afford to let the server stay offline for a day while a new controller was shipped in. In fact, no controller was going to arrive for a few days because of a snow storm. Fortunately, we had the data backed up and it was easily restored to an alternate database server.
The loss of a RAID controller is an obvious problem that can occur and common RAID redundancy is going to be of little help. There are many other scenarios where a good back is the only solution. RAID, mirroring and replication definitely reduces the stress associated with a disk failure in a critical system, but it is by no means a substitute for effective and reliable backups..