Understanding the UnRaid Raid

[chris_netsmart]

I hope that you can help as I am trying to get my head around this

 

I am looking at purchasing the UNRAID software for my home NAS but I am not 100% sure on how the raid works.

As I understand it that I can have any six HD in my Raid as long as the largest is no bigger then my Parity Hard drive, and I can add a Hard drive to my Raid at any time and I will see the total total space as free space

 

But what I am not sure is how the Raid works in this case, as I under raid with the concept of stripping the data across all the drivers, but reading about the UnRaid it sounds like the UnRaid will find any space disc space and save the data, and not strip the data.

Do I have this right?

[trurl]

4 minutes ago, chris_netsmart said:

Do I have this right?

Yes

 

Unraid doesn't stripe data. Each file is contained completely on a single data disk. Each data disk is an independent filesystem and can be read by itself on any linux.

 

Since it isn't striped, each file must be read from a single disk, so can only be read at the speed of a single disk. Also, any write updates parity in realtime. Writing to a data disk in the parity array is actually somewhat slower than writing to a single disk due to the parity update.

 

But, because of how Unraid parity works, disks of different sizes can be used, and any disk failure doesn't impact data on other disks. In order to rebuild a failed or missing disk, parity plus all other disks must be read to provide the calculated data for the rebuild.

 

Here is a wiki with a good overview:

https://wiki.unraid.net/UnRAID_6/Overview

 

And here is a topic discussing how parity is updated and options you can set to affect that:

https://forums.unraid.net/topic/50397-turbo-write/

 

 

[Squid]

9 minutes ago, chris_netsmart said:

but reading about the UnRaid it sounds like the UnRaid will find any space disc space and save the data, and not strip the data.

Correct.  It will place the files on certain drives according to various rules which you can set.

 

The non-striping has both pros and cons.

 

Cons: Read Speed is limited to the speed of the individual drive

 

Pros: Far, far easier to expand the array, and drive sizes do not have to match so long as Parity(s) are the largest  If you ever have more drives fail simultaneously than your # of parity drives, you will only lose a portion of the files, not the entire array like you would in a traditional RAID setup.  (This is the one thing that I could never wrap my head around.  Why would anyone choose a system where worst case scenario is losing everything vs with unRaid worst case scenario is losing some)

[Delarius]

Another advantage that may not impact some users is that I can shutdown the server, remove any data drive - and mount on any machine that understands XFS (I mostly use Linux anyways, so this isn't an issue for me.) If I mount read only then I won't have to parity check when I put the drive back and turn unRaid back on.

For me unRaid provides a simple way to organize my files, offers some protection from drive failures and most importantly for me, can be managed using linux commands that I'm already entirely familiar with.

 

[trurl]

3 hours ago, Delarius said:

can be managed using linux commands that I'm already entirely familiar with.

But note that normal usage doesn't require any linux command knowledge, and management can and should be done in the webUI.

[chris_netsmart]

thanks for you reply.

ask I said before I understand normal RAID and with comes with Raid ReSync Prioity and Raid Scribbing does Unraid come with this or something similar

 

[JorgeB]

Not resync priority, parity check is the equivalent to scrub on mdraid.