Why NAS with RAID is Popular As Backup
NAS devices are becoming very popular as a shared backup device with convenient network access. For extra data safety, these often come with the option to use multiple disks in an array, bolstering the idea that your data is safe on your NAS.
The most common kind of disk array I see in NAS devices is RAID5, which utilises a minimum of three disks to store your data, being able to sustain the failure of one disk without data loss. As far as protection goes from physical disk problems, RAID5 is a good, basic solution.
When RAID and NAS Devices As Backup Isn’t Enough
However, there are problems with trusting a RAID5 alone with your data – or any single disk array, for that matter. The most common problem is that various logical issues such as virus infection, accidental deletion or data corruption can destroy your data in a single event, even if the disks in the array are still healthy. The second problem is more environmental in nature: power surge, fire, theft, flood, and so on. These events are not directly IT-related, but can absolutely wipe out your disk array and your data.
With these problems in mind, it may become clear why many IT support folk regularly preach, “RAID is not a backup!”. Indeed, I regularly receive multiple disk arrays from servers and NAS devices (Network Attached Storage) for recovery after one of the above events has occurred.
From Bad To Worse – The Silent Killer!
It gets worse. There is another problem that is less obvious, one that I view as something of a ‘silent killer’ of your data. It seems that many consumer-grade NAS devices with the classic 3 or 4 disk RAID5 configuration will continue working over the months or years without making a fuss, as long as the array hard drives pass a basic health check when powered up. The basic health test typically relies upon the built-in S.M.A.R.T system in each individual hard disk drive. As long as the array controller sees the disks reporting they’re OK and array synchronisation is maintained, there’s no problem to be logged.
The problem is that a hard drive can begin to deteriorate internally with growing areas with read problems, despite S.M.A.R.T reporting all is well. If array sync is maintained, the problem won’t even be noticed until older data is accessed for the first time in a while, at which point suddenly you may find the array goes offline, errors are suddenly noted, seemingly out of thin air! You may find then that one or more of the disks are actually approaching internal failure, and the problem has been silently growing, unchecked, for a year or two. It’s a surprisingly common scenario, in my experience.
What You Can Do About It
If you can’t afford to swap your disks out old-for-new in your array every so often (every 18-24 months isn’t a bad idea), at least consider investing in a single, large USB drive that can be connected to your router or NAS (if either have a USB port for secondary backups) and set an automated routine to keep your array backed up on the external drive. This is a minimum requirement to avoid needing data recovery! The better solution is to look into offsite data storage, but this increases your costs too, of course.
The problem is that HDD heads and media (platters) can deteriorate over time, and then one day you may wake up to find that your array is offline or in an error state even though your disks appear to be fine. Once the array arrives in my lab for data recovery, closer inspection typically shows that all the disks from the array have been gradually deteriorating with time and use and are now riddled with bad sectors. Recovery is usually possible with hardware imaging equipment, but even with Data Retriever’s lower pricing it still can be an expensive exercise given the many hours required to image failing drives and to then rebuild the array with emulation/recovery software.
Let me be clear that employing RAID and NAS devices as backup is a good idea. As for most digital storage scenarios, the advice is the same: keep multiple copies of your data in multiple physical locations (see this article for ideas). This is true redundancy, in comparison to the redundancy of RAID (‘redundant array…’) and it is the best way to avoid data loss.
Act now while everything is working – don’t put it off until you need my help!