Main Area and Open Discussion > Living Room
Building a home server. Please help, DC!
40hz:
there is no commercial grade NAS
-lotusrootstarch (September 01, 2011, 07:28 AM)
--- End quote ---
Hmm...
There's a huge market for something bigger than a file server that doesn't entail the complexity and expense of implementing a SAN solution. That's where NAS really shines. Several of my corporate clients already use NAS appliances. And several others plan on getting one.
So...perhaps somebody better inform NetApp, HP, Hitachi, and a host of other manufacturers, that there's no such thing as what they're selling? :P
-40hz (September 01, 2011, 03:56 PM)
--- End quote ---
True. I'm not trying to deny the business merits of NAS appliances, I was just saying that there are a bunch of factors that prevent NAS from playing a serious role. :)
To list a few:
1. The NIC becomes a serious bottleneck with increased capacity. How long will it take to back up a 10TB NAS with a SMB transfer speed of ~35MB? Very cumbersome to move large chunks of data for purposes like archiving/making backups.
Think about it even in a home usage scenario, a normal BD burning session at 8x from a NAS can easily max out its link.
2. Lack of fine-grained access management even with AD integration.
-lotusrootstarch (September 01, 2011, 05:29 PM)
--- End quote ---
I haven't found that to be the case in many of the "commercial grade" products. Most support CIFS, NFS, NCP, HTTP and FTP protocols. And virtually all have multiple 1GB NICS provisioned for failover, load-balancing, and teaming. So there's no dearth of usable bandwidth there. And many either currently have, or will soon have, FCT -so they're also SAN ready should you ultimately need to take you NAS box in that direction. Same goes for access granularity. The management software and OS is top-notch.
But this is commercial grade ($5k range and up) we're talking about right? I'm not talking about a SnapServer or something similar.
:)
lotusrootstarch:
The schism/stepping off point is (or seems to be) cost. For the cost of 10TB of BestBuy class NAS boxes one could easily just get a refurbished commercial server that will always have parts available, is designed to take 100+ times the beating you'll ever give it, and it has a proper true hardware RAID controller ... With a year warranty ... For roughly the same price.
-Stoic Joker (September 01, 2011, 06:00 PM)
--- End quote ---
The running cost, the noise, the heat, the storage space... all for just 10TB in a home storage context? Hardly worth it IMHO.
I'm going OT a bit to look at this from another perspective, do we really need the "redundancy and high availability" for this kind of usage? I used to deploy RAID 10, get my AV content carefully ripped, well organized, double-checked for corruption... only to have the majority of content replaced in less than a year due to newer releases which entail better content, higher definition, 2D->3D, etc.
Take Avatar (2009) for example (I assume most of you have watched it by now), I had over 150GB data turnover on this particular title alone.
1. Avatar BD 1080p - Ripped, organized, verified, gave away and removed from storage.
2. Avatar BD 3D 1080p - Ripped, organized, verified, gave away and removed from storage.
3. Avatar Extended Collector's Edition 1080p - Ripped, organized, verified, gave away and removed from storage.
4. Avatar Extended Collector's Edition 3D 1080p - Ripped, organized, verified, currently in use, hopefully it'll last another year or so before they re-master it to go beyond 1080p.
Same theory applies to audio as well, at least in the genres that I listen to (DTS Master HD soundtracks and classics).
Thus there's not much reason to spend so much money and effort to have HA storage in the home media context. :) The content is disposable and is likely to be phased out before you have a clue.
Stoic Joker:
The schism/stepping off point is (or seems to be) cost. For the cost of 10TB of BestBuy class NAS boxes one could easily just get a refurbished commercial server that will always have parts available, is designed to take 100+ times the beating you'll ever give it, and it has a proper true hardware RAID controller ... With a year warranty ... For roughly the same price.
-Stoic Joker (September 01, 2011, 06:00 PM)
--- End quote ---
The running cost, the noise, the heat, the storage space... all for just 10TB in a home storage context? Hardly worth it IMHO.-lotusrootstarch (September 01, 2011, 07:06 PM)
--- End quote ---
You can't possibly be serious. Servers went green to ya know?? running cost is dependent on load. no load. no cost. Sure it draws a bit at idle but without a monitor... less that any other desktop I'd wager.
Noise? Total bullshit. Anything from a PowerEdge 1800 or newer has fan speed controls. Sure under peak load they sound like a vacuum cleaner, but at idle - where they'll be spending all their time - they're whisper quiet.
Heat? Fans say there really ain't much to speak of (See above), especially if you spec it with its actual usage in mind. No 2nd Xeon, no extry 20GB of RAM, etc. etc.
Oh, and it up to the individual to decide if their data is worth keeping ... There are whole TV channels dedicated to movies from the friggin 20s. So obviously somebody had to be hanging on to that shit for some time now... Huh??
lotusrootstarch:
Noise? Total bullshit. Anything from a PowerEdge 1800 or newer has fan speed controls. Sure under peak load they sound like a vacuum cleaner, but at idle - where they'll be spending all their time - they're whisper quiet.
-Stoic Joker (September 01, 2011, 07:43 PM)
--- End quote ---
Well I cannot comment on the PowerEdge 1800 since it looks like it's barely entry level and you'd probably better off getting a beefy quality desktop for the price anyways. I was referring to normal servers such as DL385/R710/R810.
Oh, and it up to the individual to decide if their data is worth keeping ... There are whole TV channels dedicated to movies from the friggin 20s. So obviously somebody had to be hanging on to that shit for some time now... Huh??
-Stoic Joker (September 01, 2011, 07:43 PM)
--- End quote ---
Much less applicable for young generations but granted you've got a point. :)
steeladept:
Hm... (Let me scamper a bit further out on the limb...)
MTBF (Mean Time Between Failure) as I understand it is the earliest statistically likely point for a given device to fail.
Now (operating completely without a net...), the odds of a coin landing heads up are 50/50. Which is to say that statistically there is a 50% chance of it coming up heads (I do believe it's safe to interchange them in that fashion ...Yes?).
The fun starts when you look at the odds of a coin coming up heads if it's flipped (oh lets say...) 3 times ... Because it is still 50/50 due to each flip being a separate event with 2 possible outcomes.
So I have a bit of trouble getting my head around the idea that the MTBF of 3 devices, is lower than the MTBF of 1 device. When they all individually have the same odds (statistically 0 until age X) of failure at any one given point in time.
-Stoic Joker (August 05, 2011, 09:17 PM)
--- End quote ---
Only a month behind on this particular post but whateva....:D
Just thought I would try to explain where you went wrong here...You are right in your example, any given time, there is a 50/50 chance of it coming up heads. The idea of MTBF is not that it will come up heads any one time, but rather what is the chance of it coming up tails any one of the 3 times. It isn't how likely it is SOMETHING is going to fail at any given time, it is the likelyhood of ANYTHING NOT failing at a given time. Another example may make it clearer: It isn't the chance of the coin coming up heads or tails any 1 time, but rather the chance it will come up heads EVERY time. MTBF is the average number of flips that will statistically guarantee heads comes up every time. With a coin it is 50/50 (as is any even distribution), but with something designed NOT to fail, it has a lifespan that can be used as a reference to statistically determine how long the product should last before it fails.
Note that statistically means you are still 50/50 on your one device/component/whatever but that all devices/components/whatever will average out to that number. When you see an MTBF of 5 years, that doesn't mean it will last 5 years, what it means is the average life of any given sample of that product will work out to about 5 years. YMMV. The real value is determining between manufacturers and/or product lines. An MTBF of 3 years for one device and 5 years for the other means you are likely to have the 5 year one significantly longer.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version