This topic has been marked solved and closed to new posts due to inactivity. We hope you'll join the conversation by posting to an open topic or starting a new one.
I just got my ReadyNAS 316, installed two 6 TB WD Red drives, and (with the help of the awesome support techs) got everything up and running. Now I need to figure out which RAID level/type is best for my needs.
I have some "test" data on the current setup just to make sure everything was working nicely, but I can blow that away because it is stored on my laptop.
Here is what I want:
I would like complete mirrors. That is, I should currently have 6 TB of useable storage on drive A and drive B will be the mirror. I will only install drives in pairs, so my next set of 6 TB drives (C and D) will give 6 TB of additional storage on drive C and drive D will be THAT mirror... and so on.
The intent is to fill Drive A with photos and video. Wait until it's completely backed up/mirrored to drive B, remove drive B and store offsite, insert a new drive B and go on with life.
The same process would be done with drive C (and, consequently, drive D would be removed, stored offsite, and a new drive D would be inserted).
Here are my questions:
Is this feasible?
What kind of RAID would I use?
Can I just remove mirrored drives like this and insert new ones with the expectation (and result) that the RN 316 will see this new drive and think, "Hey, that should be the mirror for drive A (or C). Let me fill it up!"
Should I have separate volumes for each of the pairs? That is, should I call drive A "First stuff", drive C "Second stuff", and so on?
Is there a better way to do this while keeping the ability to pull drives and store offsite and have mirrors of everything in the RN 316?
I back up shares (to backup NAS, not external USB drives, but the idea is the same). I've had a NAS for some years, and currently have ~20 shares (a couple less). The main issue with backup is that you can't let a share get bigger than the backup drive. I've had to reorganize my backups occasionally but in general it hasn't been much of a problem. Organization on the NAS takes a little thought, but with volumes of the size you are planning you will need some folder structure anyway.
When you back up the full volume, of course you need a destination drive that can hold the full volume. Once you overflow your initlal 6 TB volume size, that won't work (since there are no 12 TB USB drives). So you'll end up share-by-share anyway.
The RAID level you describe would require 3 RAID-1 volumes (one volume created for each pair of drives).
Removing the mirror and storing it off site is not recommended.
An alternative I like better:
use XRAID single redundancy, install 2 disks and create single share on the volume. Then back that share up to an external USB drive until the backup USB drive becomes full.
Then add another 6 TB drive to the array (which will expand by 6 TB) and create a second share. Put all new files into the second share, and start backing that up to a second USB drive.
Repeat the above when the second USB drive fills.
Cost per TB is similar to what you are proposing, and the NAS storage is much more efficient (holding 30 TB of media instead of only 18).
Right now, the two drives are inserted and according to my admin page I have Raid 1 and X-RAID is green which means that I am using X-RAID with single redundancy, correct? I am assuming it is a single volume (called 'data') as I just did the generic setup. When you say "create single share on the volume" (and please don't smack your forehead when I ask this), what do you mean?
Yet one more question. I just don't get the X-RAID idea. If I have 5.5 TB of data on Drive A, 5.5 TB on Drive C, X-RAID only has 5.5 TB of "protection" on Drive B. If Drive A completely flatlines, can I get all the data back using Drive B? If Drive C completely flatlines, can I get all the data back using Drive B. If both Drive A and Drive C flatline, will I get all of my data back from Drive B?
With three or more disks X-RAID single-redundancy uses RAID-5. With three disks you could withstand the failure of any one disk.
There are of course other possible problemy you may run into including multiple disk failures. If the NAS will be used for primary storage then backups to USB disks, another NAS or some place else are vital.
Okay. So three of you have chimed in and I super appreciate it.
Summary: Do X-RAID. It will allow you to get back your data if any one drive fails. Don't use the RN 316 as a sole backup. Backup each disk to an external HD via USB. Store that external offsite.
...Backup each disk to an external HD via USB. Store that external offsite.
Though I think you have it, just be sure... You are backing up the data volume (which spans all disks with XRAID) to one or more USB drives. As was mentioned earlier, you can't back up individual drives, because RAID is assembling them into a single virtual drive.
Well, it's one of two choices for the external backups.
Create shares (which I am guessing are similar to folders) right now and backup each share. That is, have a "share" folder called Pictures 2015 that I will backup, and so on. This allows for smaller logical chunks to be backed up. Although I am guessing this is the messiest as far as organizing goes.
Do what you suggest and make a multi-drive backup once every-so-often. How challenging would this be?
I back up shares (to backup NAS, not external USB drives, but the idea is the same). I've had a NAS for some years, and currently have ~20 shares (a couple less). The main issue with backup is that you can't let a share get bigger than the backup drive. I've had to reorganize my backups occasionally but in general it hasn't been much of a problem. Organization on the NAS takes a little thought, but with volumes of the size you are planning you will need some folder structure anyway.
When you back up the full volume, of course you need a destination drive that can hold the full volume. Once you overflow your initlal 6 TB volume size, that won't work (since there are no 12 TB USB drives). So you'll end up share-by-share anyway.
