Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In with Google

Poll

No poll attached to this discussion.

In order to better support our growing community we've set up a new more powerful forum.


The new forum is at: http://community.covecube.com


The new forum is running IP.Board and will be our primary forum from now on.


This forum is being retired, but will remain online indefinitely in order to preserve its contents. This forum is now read only.


Thank you,

Advice for a RAID newbie?

edited May 2013 in DrivePool
Hi guys. First of all, I'm a complete RAID newbie so please bare with me :)

I've recently bought a HP MicroServer N40L as a dual-purpose NAS/Media Player. It has a 120GB SSD as the main OS (Windows 8) & applications drive and four 3TB Western Digital Red HDDs for storage of my media and data (eg. tv shows, movies, music, photos, documents, etc).

Now I've been advised that because it's a NAS and will be storing a lot of data I should go with a RAID setup to ensure that if a HDD or two fails, the data wont be lost and can be rebuilt if you plug in a HDD(s) to replace the dead or dying ones. Is Drive Pool the solution for me?

I was also told about FlexRAID and have seen people say they're using FlexRAID with Drive Pool - does that mean I will have to use both? What does FlexRAID do that Drive Pool doesn't and vice-versa?

Thanks guys :)

Comments

  • edited May 2013 Member
    Sounds very similar to my setup :)
    I have a N36L with a 120GB SSD on the ODD connection and currently it's running with 1,5+1,5+3TB storage and is running Server 2012 doing NAS, seedbox, vm hosting etc.

    If you need disc redundancy is up to you; if you don't really care if a disc fails and you loose all the data on it then you don't need RAID :)
    But most likely you do and in that case what DrivePool does with it's duplication is similar to RAID 1 (mirroring) but more flexible.

    Not sure the benefit of using FlexRaid WITH DrivePool but you can run them side by side on the same machine if you wish...

    Why don't you try both and see which one works best for you? You can run them both side by side using 2 discs for each and once you've chosen move all data over to that and uninstall the other product.
    DrivePool is pretty much next, next, done while FlexRaid will probably take you a while longer to setup...
  • edited May 2013 Resident Guru
    Both offer data pooling. I find DrivePool is easier to setup/manage (and it's cheaper). FlexRAID also offers parity repair - but as a separate purchase to FlexRAID's data pooling. So some people are buying DrivePool for the pooling and then buying FlexRAID's parity repair to use on it (as it can be made to work with DrivePool).

    If you do a search of the DrivePool forums for "flexraid", there's several threads about setting it up with DrivePool, e.g. http://forum.covecube.com/discussion/899/stablebit-flexraid-questions-from-a-beginner/p1
  • Member
    Thanks guys. So really the only thing FlexRAID would be needed for if using DrivePool is parity repair? Is that any better/worse than Drive Pool's duplication or mirroring?

    Alternatively, are there any free options or alternatives to FlexRAID that will offer parity repair and work with Drive Pool? Thanks :)
  • Member
    Using parity instead of mirroring (duplication) gives you more usable disc space.
    In your case you would get 9TB of usable disc using a standard RAID5 parity setup. But only 6TB usable if you using RAID1 /mirroring /duplication.

    Of course with Drive pool you may choose to duplicate only part of your data if you don't need everything to be redundant. In my setup I don't have duplication enabled for "stuff" that can be easily re-downloaded from the internet for example.

    Dynamic discs or storage spaces which is built into your OS are "free" options for you. They just aren't so flexible when it comes to removing/replacing discs.
  • Covecube
    Dynamic disks are a bad idea. Avoid them when possible. They don't like moving from system to system...

    As for parity, yes, you save on disk space. But at what cost? Reliability, and up time. Sure, you save space... but if your files are mirrored, you can remove a bad drive, and *instantly* be back up and running within minutes. Not hours or days, or however long it takes to rebuild the array from the parity.

    Also, I've yet to have somebody adequately explain to me how parity isn't compression...
  • Resident Guru
    It's possible to design file-level parity systems so that you can continue using what's left of the array (at least, read-only) while the lost disk is rebuilt, but I don't know if anything off-the-shelf can do it yet. Saitoh183 would probably be the one to ask if FlexRAID has that ability.

    Parity isn't compression because the original data is not compressed. Simplistically, a "parity block" is calculated and stored on extra disk(s); if any part of the original disk(s) is lost, it is possible to rebuild it by reverse-calculating from the parity block and the remaining data, so long as the amount lost isn't bigger than the size of your parity block.

    As you say, the tradeoff is mirroring's extra disk space versus parity's CPU cost (since it has to keep recalculating the parity block as you write to the array) and lengthy rebuild time (reverse-recalculating lost data requires even more CPU time).
  • Member

    Dynamic disks are a bad idea. Avoid them when possible. They don't like moving from system to system...

    As for parity, yes, you save on disk space. But at what cost? Reliability, and up time. Sure, you save space... but if your files are mirrored, you can remove a bad drive, and *instantly* be back up and running within minutes. Not hours or days, or however long it takes to rebuild the array from the parity.
    Also, I've yet to have somebody adequately explain to me how parity isn't compression...
    Yep; I don't recommend them either. But they are "free" and that was the question :)
    Regarding reliability it depends on the implementation. For example double-parity (raid-6) provides better protection against disc failure than a 2-way mirror.
    But it's a much more complex task to replace a failed disc in a parity setup. And if you want to change to bigger discs you usually need to wipe the entire array and start from scratch.

    Parity isn't compression because the original data is not compressed. Simplistically, a "parity block" is calculated and stored on extra disk(s); if any part of the original disk(s) is lost, it is possible to rebuild it by reverse-calculating from the parity block and the remaining data, so long as the amount lost isn't bigger than the size of your parity block.

    As you say, the tradeoff is mirroring's extra disk space versus parity's CPU cost (since it has to keep recalculating the parity block as you write to the array) and lengthy rebuild time (reverse-recalculating lost data requires even more CPU time).


    Good explanation on parity :)

    Parity usually comes with a big performance hit on small random writes as well since many more disc IO requests need to be issued when compared to a mirrored or single disc setup.
Sign In or Register to comment.