Do I need RAID 1 if I have a backup?

Question

I have a main PC with 2 x 750 GB RAID 1 (full), another PC with 500 GB (almost full) and a NAS with 2 x 1.5 TB (ca. 1 TB used, partially with backup of the most important data from the PCs). I have also a 4-bay HP microserver (running OpenMediaVault), with 2 x 4 TB RAID 1 storage (currently almost empty).
My plan is to use the PCs and the NAS as a primary storage and the microserver as a backup. I will also set up an additional off-site backup of the microserver (crashplan to a friend's house).

Now, the NAS cannot be expanded (it handles max. 2 x 2TB), so my plan is to dismantle the RAID to double the capacity. My thinking is that since I have a backup, the extra redundancy is not needed - am I right, or am I missing something? (I want to keep the RAID in the PC, because that is where I dump pictures from my camera - so until syncing to the microserver is finished, they are without bakup.)
Or perhaps there is some different, better, way to use the hardware I have?

Answer 1

RAID does not protect you against the same things that a backup does.

In a work desktop computer, for example, you may well want to use RAID1 and would almost certainly want to back up your data. If a hard drive fails, RAID1 means that you can continue to do your work. On the other hand, if you accidentally delete an important file, RAID won't help you at all. Then, it's time to go to backup. Also consider what happens if your work computer is infected with malware. Would the malware be able to trash the backups?

In your case, all of your data is backed up. And you seem willing to pay the cost in time if your NAS completely fails. That is, you'd be willing to go a few days until you could rebuild the NAS and restore the data. The RAID isn't buying you much on your NAS, therefore. But be very aware, by dropping RAID1 from your NAS, you are explicitly saying you are willing to lose all the data in the event of a hard drive crash, and are willing to pay the cost in time to rebuild it in the event of a failure. This seems reasonable in your case.

If I was in your case, I'd be tempted to benchmark and consider using the microserver as your primary storage, with the NAS (without RAID1) as the backup. But this is a pretty small change. +1 to OpenMediaVault, by the way; that's what I use.

Answer 2

The two serve different purposes. For mission-critical applications, you want to use both.

RAID is designed to protect against disk failures and can prevent service interruption, but cannot protect against data loss due to other causes.

• The point of RAID 1 (or any redundant disk array other than RAID 0) is to keep the system up and running even while a disk(s) has failed—the failed disk(s) can be replaced and the array rebuilt without taking down the system. This is necessary in mission-critical applications where any interruption is potentially very disruptive, and backups do take time to restore. In addition, because backups are usually taken at regular intervals, restoring from a backup generally involves loss of data since the last backup was taken; RAID can prevent this from happening due to a disk failure.

• It is important to remember that RAID by itself is not a backup—it merely prevents data loss and interruption of system operation due to a disk failure, and cannot protect against data loss due to any other cause. Data deletion or corruption due to human or software error or malicious activity will simply propagate across the drives in the array.

Backups protect against data loss from a broad range of events but restoring generally involves downtime.

• Backups provide protection against a variety of different events that can render the computer unusable or lead to data loss, ranging from failure of the RAID controller or too many disks in a RAID array, to malware and accidental data deletion, to physical destruction of the computer hardware itself.

• However, most backup solutions cannot restore the system without interrupting service. Most also cannot provide continuous protection; as mentioned above, any data since the last backup can be lost in a data loss event. (There are continuous data protection solutions that can protect against this, but as they need to monitor all file changes and back up the changed data in real time, they consume considerably more resources than ordinary backup solutions.)

Answer 3

Do I need RAID 1 if I have a backup?

No, if you have off-site backups then you do not need it.

If you cannot afford downtime then some form of RAID is still recommended though, but for most people this is not an issue.

My most used example for this is a fileserver at work, serving hundreds of people. Multiple backups are off-site and there is little risk of data loss. But you still do not want that server to go down in the middle of the day. For those kinds of things RAID can be great. When a disk fails the server keeps running. Then after office hours the failed disk is replaced.

• No loss of precious (unscheduled) down time
• No need to download massive backups over slow lines.

At home this is usually less of a problem, and with SATA drives it is less likely that a disk will fail and get dropped gracefully. (Error recovery for SAS drives is part of the SAS protocol).

For home usage though: You already have backups. You are ahead of 90% of the users. Skip RAID unless you need the extra speed, or unless you want to play with it to gain experience.

And for your personal case: How fast is the off-site backup? How important is that data? Depending on that just drop the mirror on the NAS and create two separate volumes (not a stripe, 2x JBOD).

Answer 4

raid 1, and well nearly any raid level isn't about backup, its about availability. Backups mean that if your computer is eaten by a very hungry bear, you can get another computer (and a picnic basket to distract the next bear), restore and get on with your work.

Raid 1 means if you lose one disk, you can replace it with another disk (with minimal downtime, since you already have the data on hand) and rebuilt the raid array. While 'better' raid levels do give better security, none of them will help you if your data is corrupt, deleted, or destroyed. On a system with hotswap, this means you can go buy a drive, pop out the dead drive, install a new drive... and not have any downtime at all. With no hotswap, its a matter of shutting down and swapping in the new drive.

With backups, you would need to copy things over, which would take time.

If downtime is not an issue, breaking up the raid makes sense if you have up to date backups near enough to do a restore,

That said, the smart way to do this feels like backing up first then breaking the raid. Breaking the raid may be risky, and this is the sort of situation where you do want backups.