UnRAID experiences

April 25, 2017 at 8:55 pm Leave a comment

Recently, NewEgg had a deal on an HP ML10 V2 server for about $170 after rebate.  It included an i3 processor at 3.5 Ghz, a 500 GB hard drive, and 8 GB of ECC ram.  I had a hard time passing up that good of a deal, so I didn’t.

After playing with VMware ESXi 6.5 on it for a bit, I decided to try UnRAID.  I was interested in using Docker on it, something I have dabbled with on my Synology.

Having used UnRaid for more than a week, I think I’m about ready to get rid of my NAS and use this instead.

The initial setup was easy.  I loaded the software on a USB drive, put several low capacity drives in it (largest being 1 TB) and created the array using the web interface.  It began the parity process and I started setting up shares and using it.

Let me explain a bit about how UnRAID works.  It’s not your traditional RAID array.  You basically put in whatever disks you want, select the largest one as the Parity drive, and start using your array (there are some WebGUI steps involved, but it’s very easy).  I understand that you can even take drives that already have data on them (in a format UnRAID uses), and that data is preserved, with the exception of the parity drive.  With UnRAID, you get the advantage of parity protection, so if a single disk dies, just replace it and it rebuilds.  If there’s a problem with more than one of your disks at once, you only lose data on the failed drives.  Your remaining working drives have all their data intact.

Another differences is the way shares work.  It has your traditional disk based shares, where you add a share for an individual disk, and write files to it the usual way, and it will create parity info on the parity drive should that disk fail, so your data is safe….  And it has what it calls “user” shares.  These shares span your disks.  So, you might have a media share, for example.  You copy a video over to it, which gets dropped on disk 1.  Later, you copy another video, and it gets dropped onto disk 2.  When you view directory listing of the share, though, you see a single view with all the files presented as if they were in a single structured set of folders, so you don’t have to know which disk a specific file is on…  UnRAID tracks that for you and presents it all as if it’s a single, large share.

Anyhow, over the next few days, I set up three Time Machine shares, along with a couple others and copied over the majority of the data from my NAS to it.  (I have not been storing nearly as much on my NAS recently, having cleaned off tons of media some time ago.)

The Docker container functionality is great.  You can load a docker container based on templates, so there’s not much to do but point and click, though you may have to type in a path or two, here or there.  Think of it sort of as Plug-ins or Apps – there’s a Plex container, MythTV, SageTV, and many, many more.

After the initial parity calculation was done, I moved my 4TB drive from my NAS over, replacing the parity drive in UnRAID.  It rebuilt the parity info after I adjusted the config in the WebGUI.  Then, I proceeded to swap another drive with a 3 TB drive, and let it rebuild that., and I’ve done that with yet another 3 TB drive.  At this point, only one of the original hard drives is in the array.

And, I actually want to remove that last 750GB drive from the array.  With traditional RAID, that’s pretty much a no-go.  With Synology’s hybrid RAID, or a Drobo’s approach to RAID, I think you have to stay with the same number of disks in the array, short of copying all the data off and recreating the array fresh with fewer disks.

With UnRAID, though, I’m now copying all the data from disk 3 to disk 2 using a simple rsync command.  Afterwards, according to what I’ve read, I can simply remove the disk, then create a new array with one less disk and it will recreate the parity information.

Why would I want to do this, you may wonder?  To add a cache drive.  UnRAID lets you add a cache drive (an SSD, or perhaps just a 10K RPM or 7200 RPM drive), and set up your shares to take advantage of the cache drive.  When data is written, it goes to the cache drive, and at 3:40 AM, data is moved off the cache drive to the other drives in the array, at which time parity info is calculated.

Now, if you run a business and keep critical data on UnRAID, you shouldn’t entrust the safety of your data to a single cache drive, as the parity info associated with the cache drive is only generated once per day, so there is the potential to lose whatever data has been written to the cache drive.  But if you are a home user, mainly using it for entertainment purposes, you can probably take the chance, for the performance improvement (especially with an SSD cache drive).

Although I’m still within my first 30 days of using UnRAID, it’s safe to say I’ll be buying it soon.


Entry filed under: Storage.

Palo Alto PA-220 The PA-220 Firewall is here!

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Trackback this post  |  Subscribe to the comments via RSS Feed


April 2017
« Mar   May »

Most Recent Posts

%d bloggers like this: