I have been searching on how to backup several TB of data, and have repeatedly seen the statement "RAID!=backup". How can this be? Does this refer to someone who has two drives on a RAID1 array and that's it? Surely no one would use the same set of HDDs as their everyday use stuff as well as their back up...
In my case, I have several computers, all multi TB, and I had long since been considering building a dedicated backup PC with a large array. My biggest concern is the array rebuild time. And perhaps disc failures while retrieving data from backup. Please help me understand though, is this not viable backup?
The data would exist on it's native PC and the backup PC.
If the data was lost on the native PC, it would be accessed from the backup PC.
For this data to be lost, the backup PC would also have to have multiple disc failures at the same time.
I realize this is possible, but if that is a poor backup strategy, then what is better?
To give you some numbers to play with, we're talking about various PCs (3-4) totaling maybe 4TB of space. Hypothetical backup PC would have maybe 6TB of HDD on RAID. Budget is $2000. Preferably less that $1500. I've hear people mention tape, but that is lunacy given how long it takes to backup that much data to tape, the costs involved in that hardware (I''m an individual not a corporation) and I'm dealing with media and convenience not financial data and marketplace survival. Many people say just back things up to external HDD, but how is that any different than backing up to separate PC? A hrad drive is a hard drive regaurdless if it's in an enclosure or a tower. And until 10TB hard drives can be had for a hundred bucks or two RAID will have to be used.
RAID is not backup because it is not resistant to Viruses and/or Human error. Deleting a file on your RAID1 array results in the loss of the data, while deleting a file while having it backed up on a separate drive does not.
If you can live with this, then RAID will "act like backup" in your situation. RAID1 and RAID5 allow for the failure of 1 disk, and RAID6 allows for the failure of 2.
There are a lots of reasons why RAID alone makes for a very poor backup, including:
- Any problems that affect files on your drive (file corruption, accidental deletion, etc.) are immediately and instantly replicated on all copies in the RAID set.
- Any environmental issues (power hits that damage drives, knocking over the system, etc.) affect all of the drives.
- External risks such as theft, fire, etc. apply to all of the drives.
If you data is critical and you want to ensure that it survives almost all possible risks, then it needs to include the following strategies:
- the backed-up data should be OFFLINE. This eliminates any threat from viruses, accidental deletion, etc.
- there should be AT LEAST TWO COPIES of the backed-up data that are alternated. This goes a long way toward eliminating failure of the backup media as a concern, and eliminates the problem of loosing everything when a system failure interrupts the new backup you were writing over top of your old backup.
- at least one copy of the backup should be stored OFFSITE. This eliminates environmental risks to the data.
- you should keep MULTIPLE GENERATIONS of backup. This protects against loosing a file that's been corrupted or accidentally deleted but not discovered for a period of time.