Which OS should I use for my NAS? Lots of space / Don't need high performance

M

Mysteriouskk

Weaksauce
Joined
Oct 21, 2006
Messages
122
I need a NAS OS that is easy to use and has good raid/zfs options so that I can restore a hard drive if it crashes. It is mainly for blu-ray/movie backups and I need it to sleep/power down hdds when not in use.

What should I use?
 
I am pretty sure unRAID will do that, but I've never used it. I've never used any of them, but am looking into the same thing.
 
I'm using unraid for a small server now and I like it, but for my new server I'll have 20+ drives and unraid won't be able to handle it.
 
NexentaStor all the way. The free version allows for presenting up to 16TB of data, beyond that you will need to purchase.

If you plan on having more than than I would recommend NexentaCore or Solaris Express 11 + Napp-It

I currently use Solaris Express 11 and Napp-it and it works great for me.
 
Definitely NOT any ZFS or other stripe-RAID system like NexentaStor or Solaris -- you cannot efficiently spin down drives if reading a single file requires all the drives in the stripe to be active.

The only choices that meet your original criteria that I would look at are unRAID and FlexRAID. You already eliminated unRAID in a subsequent post, so it looks like FlexRAID is my only suggestion.
 
john4200 said:
Definitely NOT any ZFS or other stripe-RAID system like NexentaStor or Solaris -- you cannot efficiently spin down drives if reading a single file requires all the drives in the stripe to be active.

The only choices that meet your original criteria that I would look at are unRAID and FlexRAID. You already eliminated unRAID in a subsequent post, so it looks like FlexRAID is my only suggestion.
Click to expand...

Flexraid is far away from the enterprise class data-security of a ZFS system.
- from mirrored system with two disks up to hundreds of terabyte.

Gea
 
john4200 said:
Definitely NOT any ZFS or other stripe-RAID system like NexentaStor or Solaris -- you cannot efficiently spin down drives if reading a single file requires all the drives in the stripe to be active.

The only choices that meet your original criteria that I would look at are unRAID and FlexRAID. You already eliminated unRAID in a subsequent post, so it looks like FlexRAID is my only suggestion.
Click to expand...

ZFS doesn't have hdd spindown? I don't mind if a set of drives spins up when I access a file. I just want all the hard drives to spindown when they are not in use.
 
Mysteriouskk said:
ZFS doesn't have hdd spindown? I don't mind if a set of drives spins up when I access a file. I just want all the hard drives to spindown when they are not in use.
Click to expand...

Of course, you can spin down your disks with a ZFS-OS.
On Nexenta, you have to disable the fault-management-service
(checks health periodically without access) or your disks won't sleep.

Gea
 
@_Gea: FlexRAID is of course not an enterprise class data security system. But it does have some advantages over ZFS. Mainly the power consumption issue. I had looked at ZFS a few times and although I like it and would probably have gone with a ZFS setup, the part I did not like was that all 50 drives had to be spinning all the time while I was accessing a file (whatever the size) from the storage pool. That's why I don't run a hardware RAID setup anymore either. The power wasted compared to data integrity ratio was just too high in my opinion, especially considering that (at least for my requirements) my server is more or less just a backup/easy access solution for all my blu-ray movies anyway. So now I use the hardware RAID card in JBOD mode and run FlexRAID instead. It provides me with some data protection while at the same time allows me to access files from the storage pool with only the drive that actually holds the data to be active. I can also easily expand the storage pool at any given time with any size HDD (or reduce it for that matter too).

ZFS is great for what it does, but for certain installations, FlexRAID certainly has it's place too!
 
treadstone said:
@_Gea: FlexRAID is of course not an enterprise class data security system. But it does have some advantages over ZFS. Mainly the power consumption issue. I had looked at ZFS a few times and although I like it and would probably have gone with a ZFS setup, the part I did not like was that all 50 drives had to be spinning all the time while I was accessing a file (whatever the size) from the storage pool. That's why I don't run a hardware RAID setup anymore either. The power wasted compared to data integrity ratio was just too high in my opinion, especially considering that (at least for my requirements) my server is more or less just a backup/easy access solution for all my blu-ray movies anyway. So now I use the hardware RAID card in JBOD mode and run FlexRAID instead. It provides me with some data protection while at the same time allows me to access files from the storage pool with only the drive that actually holds the data to be active. I can also easily expand the storage pool at any given time with any size HDD (or reduce it for that matter too).

ZFS is great for what it does, but for certain installations, FlexRAID certainly has it's place too!
Click to expand...

Treadstone, do you recommend me to use Flexraid instead of ZFS if I just want to use it for my Blu-Ray backups as well?
 
@Mysteriouskk: You really need to do your own homework and see what works best for you :)

Every solution has it's pro and cons and you need to weigh them against each other based on your own criteria to see what best suits your needs.

The following were some of the criteria that I had for my server: storage space, ease of access, power consumption, fault tolerance. You may have others as well. Look at what the different implementations will bring to the table and what the overall cost for each solution is and see who comes out on top. I put together the hardware and then tried a bunch of different software solutions including Ubuntu Server, WHS (v1 and v2), FreeBSD and settled on WS2008R2 with FlexRAID. It's not 100% exactly what I am looking for but it comes close enough and it works for me. I did some testing with FlexRAID before I committed to it by transferring something like 10TB to the server then pulled a random drive, inserted a new drive and recovered the data via FlexRAID... worked quite nicely. I also liked the fact that I could combine all my drives into a single volume 'view' that I can share across the network, so now I have a 96TB storage pool I can access all my blu-rays from (currently at 620+). My original hardware setup with 3 RAID 6 arrays was drawing an insane amount of power and I still had 3 different volumes to handle which wasn't what I was after.

Anyway, my suggestion to you would be to try different setups and see what you like best :)
 
I'm using 2.0 right now, but I'm still using the snapshot raid function (which is basically 1.4). I wouldn't use the 2.0 real time engine for a production environment just yet. It's still in development! But this new GUI Brahim put together including installation makes the setup and installation of FlexRAID SOOOO much easier :)
 
When you store data on a ZFS raid, all data is evenly spread out on all discs. This means if a disc dies, the other discs contain all data. Say you have 10 discs, then each disc has 10% of the file. This is the reason ZFS discs will not spin down when you access a file. Every ZFS disc has a small part of the file. When ZFS discs are not used, they can be spun down of course.

It sounds as if you are using FlexRaid or other solutions, then you can spin down some discs, and the other discs that contain the file will not spin down. If this is true, it is very bad from a Data Integirty view point. Say you have 10 discs, and then your file is spread out on 2 discs and the other 8 discs can spin down. Say one of the two discs crashes and then your entire file is lost, because there is not enough redundancy to repair the file from only one disc. This is a bad thing.

To advocate against ZFS because it spreads all data on every disc and can not spin down individual discs as "John4200" does - is not really clever. If you really prefer a solution that does not protect your discs, then why use raid in the first place? Then you can use individual discs and if one crashes, you have lost all that data. Very strange to advocate against ZFS "because ZFS can not spin down individual discs". ZFS gives the best data protection on the market, if you dont need that, I suggest you use ordinary hardware raid instead, or use individual discs. HEre is more info on the superior data protection that ZFS gives:
http://en.wikipedia.org/wiki/ZFS#Data_Integrity
 
@brutalizer: First, familiarize yourself with FlexRAID to understand how it works and what it brings to the table. From your message, you clearly don't know what it offers.

I like ZFS, ZFS is great and offers great redundancy, and loads of other features. In essence it is similar to a hardware RAID based system. The problem for a home based server is that when you want to access a single file from your storage pool, all drives (belonging to that pool) have to be active (spinning). Depending on your physical configuration of your server, the power consumption can be quite high. On the other hand, with FlexRAID, only the drive the file is stored on needs to be active (spinning) and hence the power consumption is fairly low (as low as you can get it depending on the hardware configuration). In my case, the difference for running a hardware based RAID (all drives active to access a single file) and FlexRAID (single drive active to access the same file) is is about 300W! This would also be true for a ZFS based setup on my server!

Now for redundancy, my current setup allows for one drive out of the pool to fail and I can still recover all the data. I am in the process to actually change that to 2 drives, so 2 drives could fail and I can still recover all my data. The interesting thing about FlexRAID is that if I happen to loose 3 drives, I would not be able to recover my data anymore, however I am only loosing the data that was stored on those 3 failed drives and ALL of the data stored on the remaining drives is still good and accessible!

So for example, say 6 out of my 50 drives fail, I won't be able to recover the data on the 6 drives, but the data on the remaining 44 drives is intact! And in case my server dies, I just take the drives out and hook them up to any other windows (NTSF) based machine or computer that can read NTFS and access my data... hassle free!

Correct me if I am wrong. but I don't think that you can do the same with ZFS?!?
 
You must log in or register to reply here.
Back
Top