Originally posted by: 0roo0roo
no, its a type of backup.
backup is redundancy period.
if my house burns down my dvd-r stack perishes.
if my dvd-r stack slowly goes corrupt because it was a bad batch of discs its toast.
does that mean its not backup?
of course not. you can honestly say that one shouldn't rely on one type of backup, thats a given. but to say its not backup is just playing games really.
Nope. You totally missed it.
RAID1, 3, 5, etc is not about having backup data, it's about having backup HARDWARE that does nothing to prevent the majority of data-loss scenarios and only assists with catastrophic drive failure isolated to one drive. For example, RAID1 using two drives ensures that the system does not go down due to a drive failure, not that your data is safe from ANYTHING other than complete spontaneous drive failure of a single drive ("spontaneous" is key because a PSU problem or lightning strike will take out both). As mentioned already, there are many more ways for data to be lost than drive failure. It can't even ensure uptime without a host of other redundancies (redundant power supplies, backup power sources, etc). Do a tally of scenarios where people from this forum have needed data restoration advice on this forum and you'll see that a large percentage are from accidental formats, corruption, and other hardware failures.
As RAID0 and RAID5 prove, drive count != backups. If a two drive RAID1 array equates to a data backup, then does a two-drive RAID0 array + offline optical disc backup equate to half a backup? RAID5 isn't even close to a backup, so let's just continue using RAID1 for this example that RAID is never a backup (unless you have two completely separate RAID arrays that are synchronized manually or on a schedule but not mirrored).
You can interpret "Backup drive" to mean two things:
1) A spare drive
2) A second drive used to routinely backup duplicate data
Just like in RAID5, a second, RAID1 drive is an online
spare drive that remains IN USE for 100% uptime. It is only ONE functional logical drive that remains vulnerable. In the event of a drive failure isolated to one drive, you still have one logical drive that remains without a backup. You had redundant hardware supporting/functioning as a single logical drive while that logical drive had no redundancy. Instead of having two pieces of hardware to support it, it now only has one. As we have demonstrated, hardware count doesn't equate to backups.
Would you say that the data duplicated in the memory of a second video card operating in SLI/Crossfire is a backup just because they are working together? Sure, those aren't data storage devices, but if the game crashes, you lose both... just like a drive corruption from the OS crashing would affect your so-called "backup" equally. If one card failed, you could take it out or disable multi-GPU to start gaming again, which similarly demonstrates that it is redundant hardware and not redundant data.
RAID1, 3, 5, etc keep a logical drive functional in a very specific scenario until it runs out of hardware redundancy. It's backup hardware, so don't call it a backup and don't ever treat it like it is.
I've had my fair-share or data-loss due to hard drive failure and I've had even more directly attributed to RAID arrays (more specifically, drives that weren't RAID-friendly). RAID introduces complexity and therefore even more can go wrong. Adding multiple drives for more capacity/performance introduces complexity and multiplies the chances of a drive failure, so redundant RAID arrays are only a step to prepare for that EXTRA vulnerability. The redundancy added by RAID is a step to minimize the extra vulnerabilies it introduced and don't do anything as far as providing a true backup.