Is RAID5 really such a bad idea for home use?

Soldato
Joined
28 Dec 2003
Posts
16,437
Ok so I'm in the process of building a new server to replace my clunky old box and am faced with that age old decision - RAID 1 or 5?

Now everything I'm reading lately is saying RAID5 is a bad idea and should be steered clear of, with the recommendation usually being RAID10 instead. I don't need the performance boost of striping so it'd just be basic RAID 1 mirroring in my case.

Now the logic seems to be that, if a drive fails in a RAID 5 array then, once the drive is replaced, the rebuild process puts a very high load on the remaining drives, which increases the probability of a second drive failing, at which point you're screwed.

Whilst I get that, surely the same applies to a basic mirror too? The entire contents of the surviving drive would still have to be read in order to rebuild the mirror so surely there's just as much chance of failure?

Cost is obviously a big consideration and, given that I need a decent chunk of space, it feels very galling having to halve the effective capacity with a mirror when I could get more bang-for-buck with a parity array.

If there are real, tangible benefits to the mirrored approach then I'd still consider it but I'm struggling to justify it right now.

Any thoughts?
 
For a home server, a single HDD as big as you can with a decent backup should be fine. Do you really need disk redundancy, will you be that put out if that disk were to fail?

Having built the home server thing and recently reshuffled, I tore down a 4 disk RAID 5 as it seemed to be using a lot of resource for the job it was doing. I'm now running a single disk which is backed up and have had no performance differences. to be honest I wish I hadn't have invested so much initially and just ran my storage in the cloud from the start.

Also:
https://forums.overclockers.co.uk/threads/raid-5-whoopsie.18867052/
 
Well, amongst other stuff, I have a fairly large media library which would be a total pain to lose.

As I'd need as much space again to back it up, I may as well have two drives in a mirror as I then don't need to faff about taking backups. (Critical data is backed up offsite anyway).

I do understand that redundancy and backups are not the same thing but, in my case, the former suits better.
 
For my home server use I go backups first, then redundancy, simply because I've potentially lost more data due to human error and software bugs than failing hardware (although I do monitor all drives so catch most issues early).

I used to use RAID 5 and have partial backups and still lost data due to the above. After changing to a mix of mostly single copy but a little duplication on critical data (RAID mirroring/software drive pooling etc) and backing up everything in various ways I've not lost any data in many years of running 2 to 3 home servers (CCTV, file/media, dev etc). I am also not using much more hardware than before and that's a key point when considering RAID 5 for a home server.
 
Well I've done some more snooping and this is how I understand it:

When a RAID 5 array rebuilds following the replacement of a single failed drive, the entire contents of every other drive needs to be read in order to reconstruct the lost parity data. As a result, the higher the number of drives in the array, the higher the chance that an unrecoverable error will be encountered somewhere, potentially causing the rebuild to fail.

With a basic RAID 1 mirror, whilst the entire contents of the remaining drive still need to be read, the simple fact that there's only one drive, as opposed to 2+ in RAID 5, mitigates the odds of a URE being encountered.

RAID 10 offers improved performance due to the striping but I can't see how it provides any better odds of successful rebuild compared to RAID 1.
Unless...
In theory a RAID 10 array could survive a twin drive failure if both drives were on either side of the mirror. In that case then I suppose that if, when rebuilding after a single drive failure, a second drive failed on the other side of the mirror, it could survive that?

Does this make sense?
 
Ultimately if you are relying on a RAID to rebuild successfully so as not to lose data then you should be looking at additional redundancy, if a drive fails due to the rebuilding process then it would have failed sooner rather than later anyhow unless the array is sitting there almost entirely idle and even then it would probably have given up when you next needed the data or soon after.
 
Yes I accept that I shouldn't be trusting any critical data to any kind of RAID array.

Anything I literally can't afford to lose is backed up separately and off-site. I'm more concerned with large media libraries which would be a pain to lose but it's not the end of the world if I do. In this specific case, redundant disks work better for me than taking regular backups.
 
I would never recommend RAID 5. If you’re going down that route then RAID 6 is the correct option. RAID on 3 or 4 disks isn’t really useful. RAID on 12 or 16 disks starts to make sense. The I stands for inexpensive. So they should be inexpensive disks.
 
The main problem with RAID 5 (and to a lesser degree RAID 1) is that you're tied to the RAID controller or RAID software. If you lose access to them (RAID controller dies and you can't get a replacement or it takes weeks/months; or something goes wrong with the software RAID) then you've lost all your data (or at the very least lost access to it for weeks while you sort out replacement hardware). For enterprise use there is vendor support, so those things aren't so much of an issue, and the ability to quickly replace a faulty drive (usually in <24h) without an outage is pretty valuable. For home use, the risk of losing access to your data because of the additional complexity of the RAID itself is just not worth it in my opinion. Much better to have synced copies where each copy is a completely stand-alone drive that needs no extra layer or software to read the data.
 
I have a 2 Bay ext Raid Enclosure and its set up as Raid 1 and i have tested it rebuilding etc and I am also sure in the past I pulled one drive and connected it directly to my Mobos SATA port and it seen the DATA and not sure why it wouldn't as its not Striped in this case the Raid Controller simply copies the files you send to it onto the other disk.

This is a backup of my Storage drive so basically the DATA is in 3 places being S (storage) Drive and X (ext) Drive Raid 1 so backed up twice and also C Drive is backed up to it via software to it every day and Windows File History also writes to it.
 
Statistically speaking with normal consumer drives and non recoverable read error rates as they are, any raid 5 array over 12tb in size is very likely to fail on a rebuild. Raid 6 or 10 does not suffer the same level of issue.
 
For media UnRAID and/or GDrive would be the default suggestions for home users, it’s actually cheaper for me to use GDrive than it is to keep a storage server running, that’s even allowing for UnRAID spinning down disks and only spinning them up to write cache data from a 2TB SSD.
 
Back
Top Bottom