Jump to content
Covecube Inc.
  • 0

CPU Usage Planning



I am building a new home server that will serve several purposes:

  • Windows Server 2012 Standard Host OS
  • (hopefully) StableBit's DrivePool and Scanner over all HDDs and SSDs
  • Hyper-V Virtual Machines on the RAID1 disks listed below.
    • 1 production server running a number of websites
    • Possibly 1 "Windows Server 2012 Essentials" Media Server, if I can convince myself to join all of my systems "to the domain".
      • It will share/access the host OS' DrivePool shares as its media.
    • 1 linux/windows machine running my, uh, SickBeard + Sabnzbd daemons.
      • NOTE: "nzb" translates to "seriously thrashes disks for huge file patches, unraring/recompressiong 10/20/30 GB at a time across 1000s of files, etc).


The hardware is able to handle all of that, considering I do it now with slower drives, slower ram, less ram, and an Q9450.


The new hardware is (I guess only the controllers matter with how I'll be using the Pools):

  • Intel Core i7 930 overclocked to 3.73 Ghz (boost to 4.1 Ghz) even with hyper-threading.
    • I've been rock stable with this board and CPU and overclock for 3 years as a developer machine.
  • 12 GB 1600 Mhz ram
  • Intel ICH10R with two separate RAID1 arrays
    • 2x RAID1 arrays
    • 1x "scratch" drive for nzb thrashing
    • 1x "backup" drive, for any and all backups (either Windows Server backup, or attached to the W2012E VM, etc).
  • 2x IBM M1015 controllers flashed to LSI firmware (16 ports + ability for an expander in the future)

Not too worried about the throughput of the RAID1 arrays with the VMs.  They aren't that busy anyhow.


Lastly, I do plan on about 20% of my DrivePool to be mirrored/folder redundant across about 5 or 6 disks by my estimates (sure, spread out to even more perhaps - see below).



So, my questions to the community/developers:

  1. How well does DrivePool and Scanner work with a CPU that is going to be taxed off and on?
  2. Should I focus on creating multiple DrivePools?  One for the bottomless blu-ray/movies shares, and one for all of my redundant folders?
  3. Should I create separate DrivePools for each extremely large folder duplication?  Think TV shows across about 40 different series, nearly all in large HD format.


The advantage to running my nzb utilities (which are very CPU and I/O intensive) in a VM is that I can throttle down the CPU usage and limit the threads it uses as virtual cores.  Again, it will be all limited to the scratch drive on the ICH10R controller (for now).

Link to comment
Share on other sites

3 answers to this question

Recommended Posts

  • 0

I run a similar setup to what you have (including Sickbeard and sabnzbd).

I've run both DrivePool and Scanner on a fairly low powered system (Intel Pentium D E5200, 2.5GHz, 4GBs DDR2), along with Subsonic (uses ffmpeg to transcode), and PHP/mySQL. I've never had an issue with this setup. Ever.

So it works pretty well with a well taxed CPU. But that Core i7 should be more than up to the tasks you've outlined.


As for multiple pools, only if you really want to. DrivePool supports per folder duplication, so you could disable duplication on  your TV folder, but enable it for movies, and music. Or however you want. 

Personally, I feel that adding additional pools just adds additional, unneeded complexity for me.


Also, I'm guessing that you want to run Hyper-V for development and testing.

Though, if you're running website's you could use the built in web server that is in Windows Server 2012 Essentials. May make things easier, I think. Especially if you install ARR and URL Rewrite for IIS. That way you could use the WSE's website to access(reverse proxy) Sickbeard and SabNZBd's webUI remotely (and securely).


And as for joining the domain.... if HyperV was also joined to the domain, it would make remote management simple. As well, as the "Folder Redirection" feature of WSE.




All in all, your server sounds very very similar to what I have set up.


Though, I run WSE on bare metal, and have transitioned it to Standard, so I can have the dashboard (and other "Essentials" features) as well as have Hyper-V (and initially WSUS).

Otherwise, it sounds like we have very much the same ideas about how to set up our servers. :)

Link to comment
Share on other sites

  • 0

Yep, very similar.  


I run a dedicated VM for my production websites for disaster recovery (I backup the VM to Windows Azure, and can start it at any time if my home server/network/power goes down).  Think portability as my reason.


W2012E requires, no, demands Active Directory setup on its box and any machines you install the client software too instantly get joined to the active directory without your consent (think work computers you work at home with, already on a domain).  It doesn't even let you run another DC (e.g. the HOST OS, thank you). Though, MS finally posted an MSDN article on how to edit the registry of the client machines to prevent the domain joining.  Most likely will stay away from this.


I only was thinking of W2012E for the media streaming to tinker with and maybe attempt to write my own transcoder, though there are several other options.  I recently ran across Subsonic myself, but it seems to be only for music - not transcoding MKVs on the fly (though that is a bit misnomer - transcoding MKV) but you listed ffmpeg which I've written encoders for in the past.  Odd.  Are you streaming movies with that?


Thanks for the feedback!

Link to comment
Share on other sites

  • 0

Yup, WS2012E requires AD. And the registry thing wasn't a "finally". It was posted before it hit RTM. In part because of the reasons you've listed, as well as others. 

Otherwise, the statement about not letting you run another DC is completely wrong. It must be the first DC on the network, and the primary DC. But you sure as heck can have other DCs on the network without issue. As long as they aren't WS2012E also. But this is how SBS has always worked (and make no mistake, WS2012 *is* SBS, even if MSFT wants to deny/hide that fact).

But other than those issues, domains are very nice, IMO. :)


As for subsonic. it originally started out as music only. Til somebody hacked it to play video. And then that got rolled into the official code. :)

But the music aspect is much more developed and works much better. The Video player relies on Flash, and i'm sure you can see why I say "more developed" for the music part....  There is a hack to use HTML5, but I've had issues with that remotely (I think my bandwidth is too low to reliably play it).  

But subsonic does use ffmpeg to transcode the video to flash (or webm if you use the html5 hack). 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Answer this question...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Create New...