Simple, Fast and Power Efficient NAS...?

[stevech]

Since the consumer grade NASes are quite competitive to homebrew prices (at best you save $100 and approach their performance), it isn't until you want much faster and/or much bigger that the payoff is really realized.

Very sage advice.

 

[loyukfai]

Sorry. No, I'm just me - keying off of the thread's title. Offering unwanted help, apparently.

I used to build my own PC as well, but now I tell people to buy one off-the-shelf, to me the effort just doesn't worth it anymore. I welcome and treasure insightful and informative discussions, but merely reiterating the same rhetoric without adding new information and insights eventually wears me down, perhaps.

FreeNAS at it's simplest is almost turn key, you shouldn't have a problem. It is ABC.

I agree with you, if you want to save money, have the inclination, and the time, DIY is the way to go. I generally recommend it primarily to folks who want higher capacity than what is offered by the consumer grade NAses.

Since the consumer grade NASes are quite competitive to homebrew prices (at best you save $100 and approach their performance), it isn't until you want much faster and/or much bigger that the payoff is really realized.

I can imagine, in most cases, if someone asks me about NAS, I will probably tell him/her to buy an off-the-shelf model and call it a day.

I expect this to eventually be a 4/6-bay setup and able to saturate a Gigabit Ethernet connection (hdparm -tT on the array says the transfer speed of the RAID-1 array is >90MB/s). It seems that pre-built solutions capable of doing these would cost >USD400...?

Cheers.

P.S. Sorry, I didn't mean to be rude or mean, and it's my fault that I didn't state my needs and conditions better earlier in the thread. But merely saying the same thing again and again based on some presumptions can hardly lead the discussion advance.

 

[GregN]

I have found even with the most conservative of builds, unless you have most of the components already, or are willing to go used, cost is around four to five hundred bucks. I find it kind of weird, it is like 500 is some watershed number.

Maybe you can do better, if you plan on growth, and don't need the bells and whistles of a fancy GUI, I'd go for it - it is much more interesting than the black box offered by consumer NASes.

Saturating a gig net channel can be a challenge, one usually requiring multiple clients or high performance gear on both ends of the wire. I have a 4gig fibre connection, and unless I'm running test suites I never cap it.

But the best way to approach that kind of performance is a DIY sol'n.

 

[loyukfai]

Again, it's my fault not to explain in clearer terms my needs and conditions earlier. BTW, this setup, sans drive, cost <150USD.

I should emphasis that the current lack of need of a NAS chassis contributed to my decision to go the DIY route. A 4-bay NAS chassis easily cost 100USD these days.

OTOH, a used ATX chassis probably cost <20USD.

If floor space is a significant constraint, then a pre-built solution will have more attractiveness.

Besides, the ability to use standard parts is also a contributing factor - When something breaks down, it's easier to have it replaced than some proprietary parts.

Cheers.

 

[loyukfai]

Did a quick test with iperf and simple Windows copying (one 4GB file) using 2 laptops via direct GigE connection, the network throughput is about 700-750 Mbit/s and file transfer is >60 MB/s.

Wondering if it's limited by the laptop's drive or the server.

It's not well controlled so take it with a grain of salt.

Cheers.

Edit: I meant that I did the tests with 2 different laptops. The tests was not done simultaneously.

 

[loyukfai]

Turns out it's webmin's monitoring script that keeps the drive from staying in a spinned down state.

Cheers.