The Ultimate unRAID Server? NEW MOTHERBOARD with BENCHMARKS!


TheMaster

Recommended Posts

I am new to the Lime Technology unRAID system, but became interested over the Thanksgiving timeframe and decided to build what I considered to be the ultimate unRAID system using top notch components at good prices that could be had during Black Friday.

 

I believe I have succeeded, and I went out on a limb to try a more modern motherboard (correct me if this was used here before, but I couldn't find a board with these features being used anywhere in the forums).

 

 

My Goals:

  • Build the Fastest, Reasonably Affordable unRAID NAS
  • Smaller Mid-Tower Form Factor for Home without Sacrificing Ventillation
  • 5 SATA/300 Hot Swap Drive Trays with Room for 8 Total Hot-Swap SATA Drives in Future
  • 4 SATA/300 Ports On Motherboard
  • Identify and Use The Fastest 1GB USB 2.0 Flash Available

 

 

New Motherboard:  ASUS P5LD2-VM R2.0

$91.18 at Buy.com with Google Checkout

 

Motherboard Capabilities:

  • 4 Onboard Serial ATA II (300MB/s) Ports - Recognized by unRAID
  • 4 PATA Drives Can be Added Onboard (If you want a mixed system)
  • Dual Channel DDR2 Memory Support (DDR2 667)
  • ICH7 I/O
  • Supports Intel 775 Pentium 4 through Core2Duo CPUs
  • Intel Pro Gigabit LAN
  • Intel 950 Integrated Graphics (DirectX 9)
  • PCI Express 16x Slot
  • HD Audio
  • More advanced Intel 945G Chipset
  • 8 USB 2.0 Ports
  • ASUS CrashFree BIOS2
  • Q-Fan automatic fan RPM control in BIOS (Quiet Operation)

 

 

Other Components:

 

COMPONENTPRICESTORE

Cooler Master Centurion 5 Tower with 380 Watt Cooler Master Power Supply$33.30SVC.com

Intel Pentium D 820 Dual Core 2.8GHz, 800MHz FSB$88.40Buy.com

1GB Dual Channel Corsair XMS2 PC5400 DDR2 675 SDRAM$104.00Newegg.com

Lexar Lightning High Speed (100x! 22MBps Read, 15Mbps Write) USB 2.0 Flash Drive$35.85Buy.com

ICY Dock 5 in 3 SATA III Bay (Fits 5 SATA III Drives in 3 Full Height Bays)$96.59Buy.com

Seagate Barracuda 400GB 16MB 7200 RPM SATA/300 HDD (Parity)$117.00Outpost.com

Maxtor 300GB 16MB 7200 RPM SATA/150 Drive 1$74.54Outpost.com

Maxtor 300GB 16MB 7200 RPM SATA/150 Drive 2$74.54Outpost.com

 

Benchmarks:

 

TEST                           My unRAID    Tom's Fastest NAS    Tom's Guide Fastest Model

Write - 64MB44.142.5DIY Home Built RAID 5

Write - 256MB49.332.2DIY Home Built NAS (not RAID 5)

Write - 1GB22.228.3Thecus High Performance NAS Server (N5200)

Read - 64MB53.146.1DIY Home Built NAS (not RAID 5)

Read - 256MB54.145.5DIY Home Built NAS (not RAID 5)

Read - 1GB12.122.0Buffalo Technology LinkStation Pro (LS-250GL)

Average Performance39.236.1

 

NOTE: Units are in MBytes/sec, Max Throughput on Gigabit LAN is 125 MBytes/sec

Tests Conducted from an AMD FX-60 Workstation (with Gigabit Adapter) through a Netgear Gigabit Switch

 

UPDATE:  Re-ran tests on 12/21 with crossover cable direct connection between unRAID and Workstation to eliminate effect of LAN traffic.

Workstation applications were shutdown with exception of normal windows background functions so benchmark was only additional application running.

Improvements were minor, but I updated the above scores to reflect the true performance of the NAS with these "noise" variables eliminated.

 

 

 

Please Post Your Results so we can Compare Configurations!

To Compare Apples to Apples Test Results, This is the Command Line I used (Same Used at Tom's Hardware Guide):

iozone -Rab unRAID_GB.xls -i 0 -i 1 -+u -f \\TOWER\disk1\filetest -q 64k -n 32M -g 1G -z

 

These Benchmarks were completed using IOzone Test Suite, available FREE from:

http://www.iozone.org/

 

Tom's Hardware NAS Charts are Viewable Here:

http://www.smallnetbuilder.com/nas-charts/

 

Pictures are now attached below.  I believe this is a very professional and attractive build with an amazing amount of features built in.  Please let me know what you think.

 

Please let me know if you found this helpful as it is one of my first posts here at Lime Technology / unRAID!

Link to comment

WOW! I am definitely interested in that motherboard.

 

Once I get my system stable I will run some benchmarks.

 

Hypyke - I look forward to seeing your benchmarks, along with anyone else in the forum who cares to give it a shot.  Use the command line I posted to assure you are taking the same measurements. 

 

Oh, and the benchmark results in the spreadsheet generated are in Kilobytes/Sec (multiply this by 0.0009765625 to get Mbytes/sec which is what's posted at Tom's Hardware Guide and on my own Benchmarks).

Link to comment

Great post,

 

It actually prompted me to join in all the fun.

Combining your specs and

http://lime-technology.com/forum/index.php?topic=379.0

 

I went with:

Cooler Master eXtreme Power RS-430-PCAR 430W ATX 12V V2.01  $25 @SVC

Cooler Master STC-T01-UBK Blue CM Stacker Full Tower Case $100-20= $80 @SVC

 

Motherboard:  ASUS P5LD2-VM R2.0 $91.18 at Buy.com with Google Checkout

ICY Dock 5 in 3 SATA III Bay (Fits 5 SATA III Drives in 3 Full Height Bays) $96.59 Buy.com

 

Intel Celeron D 352 Cedar Mill 3.2GHz  Retail  $59 @ Newegg

CORSAIR ValueSelect 512MB (2 x 256MB) 240-Pin DDR2 SDRAM DDR2 533 (PC2 4200) Dual Channel Kit System Memory Model VS512MBKIT533D2 - Retail  $56 @ newegg

Patriot 1GB Flash Drive $25 @Newegg 32MB/s Read 26MB/s write

 

$433 before tax/shipping

 

Good thing SVC has will call since shipping would have been $35.

 

Drives:

500GB Maxtor Maxline Pro 5yr warranty $149 ea @ newegg. Contemplated the Samsung but it only has a 3 yr warranty.

 

Comments:

 

I wasn't sure about the PSU, but it had good dual +12V rail specs.

There was a cheaper 1Gig OCZ flash usb with 25 Read/17 write at newegg.

It's hard to tell if the flash will have a unique ID.

 

Also, I wasn't sure if how critical the RAM size/speed and processor(I went with the 65nm CPU to save power) are to overall throughput.  I seem to recall that it was limited by PCI aggregate bandwidth. This MB has a native SATA southbridge though.

 

Unfortunately I won't be able to benchmark it because I don't have a GigE network.

 

Anyway it'll be fun to build it up over Christmas.

 

 

 

 

Link to comment

That's a nice build.

 

I didn't say so above, but for everyone reading this, most of the pricing I listed is still valid (except for the 2 hard drives and case that were discounted explicitly for Black Friday).

 

Also, I wasn't sure if how critical the RAM size/speed and processor(I went with the 65nm CPU to save power) are to overall throughput. I seem to recall that it was limited by PCI aggregate bandwidth. This MB has a native SATA southbridge though.

 

I don't really know how big a role the CPU plays - the only reason I went with the Dual Core Pentium D 2.8 was that, should the unRAID server not work out, the server would make a decent addition to my collection of PCs around the house as a gaming machine.  Same goes for the memory. 

 

Unfortunately I won't be able to benchmark it because I don't have a GigE network.

 

I did some benchmarking on the Gigabit Ethernet vs 100BT Ethernet and the performance difference is remarkable. 

 

The same unRAID server detailed above only achieved an average speed of 8.8 MBytes/Sec vs the 36.8 MBytes/Sec seen on the GBit LAN.

 

The unRAID performance was still good at 100MBit, because I tested a HNAS1 NAS Storage unit from Hawkwing Technology (these only have 100BT ports on them).  I thought it was a good buy at only $32 after rebate (you add your own hard drive up to 250GB so I purchased a Seagate Barracuda 250GB 16MB 7200.10 for $70, bringing the total cost to only $102):

 

http://www.hawkingtech.com/products/productlist.php?CatID=38&FamID=91&ProdID=242

 

Though it seems like a great deal, these dedicated NAS boxes transfer data at an average of only 2 MBytes/Sec, making the unRAID 4x faster even at 100MBit speeds!

 

NAS can be slow, so being able to run your NAS 4x faster at Gigabit speeds is a pretty big deal.  Consider this:  It takes around a minute to transfer a 1GB file to my unRAID server when connected on the Gigabit LAN.  It's more like 5 minutes at 100BT!  These are big performance differences - bigger than the difference people are going to see with component variants, I'm sure.

 

I have seen several good deals lately allowing you to buy a good quality gigabit switch for around $30.  This will probably be a worthwhile buy for you if you have a gigabit adapter on your workstation(s).

 

Here's one you can get for just $13.39 after rebate!  You need to add a $1.61 filler to get the price above $50 to use the $20 off $50 google checkout option:

http://www.buy.com/retail/product.asp?sku=10353206&loc=101&adid=17662&adid=17662

 

Also note that the compact 5 in 3 enclosure I purchased is more suitable for the smaller case I purchased.  Since you bought that monster stacker case, you may want to consider the drive trays that LimeTech uses on the servers he sells.  These can also be had for a great price at Buy.com right now with their google checkout deal.  Just search for the part numbers (it is hard to find the IcyDock stuff at Buy.com without the pat number first). 

 

Part number is MB123SK-1B and they cost $35.63 each at Newegg.com (hard to find them for less at any reputable store).

 

They are $30.15 at Buy.com and when you buy 2, you get $20 off bringing them to $60.30 - 20 = $40.30 or only $20.15 each.  They do charge shipping, but it isn't too terrible.  Just do several duplicates orders for 2 at a time if you need 4, 6 , etc.

 

I wish the admin could get the attachment function working and I would upload some images of my rig - I think seeing it may sway a few people from the monolithic Stacker case.  Especially if you think you may convert it to a desktop PC someday down the road.

 

Link to comment

Here are the details of my build:

 

CPU: AMD64 3200+

MB: Asrock 939SLI32-eSATA2

RAM: 1GB stick of Kingston valueram

GPU: cheap ATI card

HDD: 4x400GB Seagate (Only using 3 right now since I am trying the free version of unRAID)

Flash: 1GB Lexar Firefly

 

I don't know how to interpret the results from Iozone. So I am just posting the Writer and Reader Reports at 64K

 

Write:

 

64MB............41.41113281

256MB..........40.41992188

1GB..............20.29199219

 

Read:

 

64MB............63.55761719

256MB..........47.08105469

1GB..............9.193359375

 

I know the component lists for my server seem a littler weird since most people here go for Intel microatx boards, but I had initially put my system together to run a hardware RAID5. Which I was doing until I thought I would give unRAID a shot since I don't really use my server except for a few hours at night.

 

I am running into another problem which maybe someone could help answer. When my parity drive starts it makes this weird sound that none of my other drives make. And since all my drives are the same I find it a little odd. The sound is hard to describe because it is not a clicking sound more like a squeaky sound. Note that this only happens when it starts up. I don't get any errors when I run a parity check so I don't think that the drive is bad. The only source I could think of is my HDD enclosure SUPERMICRO CSE-M35T. I have already tried swapping out the PSU, but that did not solve the problem.

 

If someone can help me with the cause of my HDD sound please do.

 

Thanks.

Link to comment

 

I don't know how to interpret the results from Iozone. So I am just posting the Writer and Reader Reports at 64K

 

 

That's correct - the results used at Tom's Hardware Guide use the 64K record size.  He stated something about most transfers utilize a 64k record size.

 

Those results look great!  I did my benchmarks somewhat casually while there was other network traffic and some devices between my computer and the unRAID server.  When I get a chance, I will rerun them with the unraid server connected directly to the workstation using a crossover cable with nothing running on the workstation other than the benchmark.  This should be the best way to determine the true capability of the unRAID box in question.

 

Regarding your question about the squeaking sound of the HDD, I cannot say I have experienced that with any of my seagate drives.

 

Link to comment
  • 2 weeks later...

ok,

 

So I've had some time to play around and here is the information. As I have a "mixed" system I decided to run the tests against the SATA drive and one of the EIDE drives to see what kind of differences there would be between them.

 

My unRAID

CPU: Intel Pentium D 820

MB: Asus P5PE-VM

RAM: Crucial Ballistix PC3200 1GB 2X512MB DDR400 CL2-2-2-6

HDD1 - parity: Seagate Barracuda 7200.10 320GB SATA2

HDD2 - Disk1: Seagate Barracuda 7200.10 320GB SATA2

HDD3 - Disk2: Western Digital Cavier WD2500JB 7200 250GB EIDE Ultra ATA/100

HDD4 - Disk3: Western Digital Cavier WD2500JB 7200 250GB EIDE Ultra ATA/100

HDD5 - Disk4: Western Digital Cavier WD2500JB 7200 250GB EIDE Ultra ATA/100

HDD6 - Disk5: Western Digital Cavier WD2500JB 7200 250GB EIDE Ultra ATA/100

Flash: SanDisk U3 Cruzer Micro

IDE Controller: Promise Ultra 100 TX2 Flashed to latest firmware

 

The SATA drives are connected directly to the motherboard and all the EIDE drives are connected to the Promise controller.

 

TESTSATA2 on-boardEIDE on PCI controller

Write - 64MB45.0938.08

Write - 256MB34.5424.71

Write - 1GB20.8912.12

Read - 64MB44.0133.72

Read - 256MB46.5844.56*

Read - 1GB15.8210.37

Average Performance34.4927.26

* seems to be an anonomly

 

As expected, there is good jump between the EIDE drives and the SATA2 drives. I also think that the limiting factor in this test is my workstation which is just liitle bit slower than TheMaster's pimped-out workstation  ;)

 

So there you have it.

 

 

Link to comment

RADIatiON:

 

Thanks for the post.  It is great to see what the real difference is between the SATA and EIDE performance.  Now you have me wondering... those hard drives I purchased were mediocre, at best.  I wonder if there would be a measureable difference with higher performance SATA/300 drives for both the Parity and storage drives?  Too bad noone here as a pair of Raptor 150's hooked up to see what these transfer rates could top out at.  It would be an absolutely stupid choice for a NAS setup, but it would be interesting to know what the future will hold as 10,000 RPM hard drives become more common.

Link to comment

TheMaster,

 

I agree with you. It certainly would be interesting to see what kind of speeds you would get if you were using at least an SATA2 for one of your data drives.

 

I totally agree that it would be a bit of overkill for the high end 10K rpms, but at the same time, that might push this little unit into another target market... HDVideo processing etc. Although, I guess those are really better suited to local striped drives.

 

BTW, I'm interested in pics of your unRAID, have you thought of posting the pics to ImageShack or similar?

Link to comment

I think building an unRAID NAS and reading this thread I have some questions.

 

About the initial post, the cool_runner and the RADIatiON systems:

1) Why is faster to write 1GB than read it?

2) Why the big difference in performance in 256MB tests x 1GB tests, where is the bottleneck? RAM memory?

3) About (2), I ask if the performance in 1Gb test begins fast and remains dropping. So bigger transfers (>1GB) would be slower?

4) About (2), if I have two machines running the test at the same time the total data transfer will be 2x more or will split it (half to each machine?) (thinking where is the bottleneck...)

 

About performance, I think the parity must be the fastest drive (higher MB/s and slower ms/track), so it will not be a limiting factor. The higher RPM drives tend to have slower ms/track. It helps when the files are fragmented over the disk. Faster MB/s drives will not speed much through the network, but probably will give faster rebuild times (rebuild drive, recompute parity). And I would not mix master and slave PATA drives in the same cable because it will slow the rebuild times (never doing this with the parity drive!).

 

Link to comment

Thanks admin - the pictures are now up as everyone can see above.  I believe the server is compact, looks professional and still has great ventillation without sacrificing significant space for drives.  Note that with the onboard SATA 3 controllers, I can run up to 4 drives.  With the onboard PATA connectors, 8 (though I wouldn't do that).

 

My intention is to add a Promise 4 port SATA 3 card and a second ICYDOCK "3 into 2" drive cage which will give me 8 high speed ports and room for 8 drives - more than I will likely require.

 

SergioK:

 

You make some compelling arguments regarding drive speed.  I am not an expert in how the unRAID software utilizes the drives, buffers to memory or writes the parity, however, intuitively your statements regarding the need for fast seek times on parity ring true to me.  I am willing to reconfigure my array with some faster drives if we can determine what the optimal configuration would be.  If I do anything with the Parity drive, however, I want to do it once... I was hoping to wait for a fast 1TB drive to come onto the market and swap it into the Parity spot.

 

If I ever do this, however, I would probably rebuild the array entirely because fragmentation will undoubtably effect performance metrics.

 

A great website for storage device performance is:

 

www.storagereview.com

 

They have top drives by category and detailed drive performance results to do some back to back comparisons of various drives.

 

Perhaps limetech can give us some advice regarding the most critical features for parity and data disks to optimize performance?  While you're at it, I am also wondering about fragmentation.  This will ultimately be a factor on the data drives, but I don't see a way to defragment them in the software.  Any advice?

Link to comment

What kind of temperatures are the rest of you seeing on your Parity drive during a parity check operation?  I want to validate my statements about "good cooling".  Ideally one of you who are using a Seagate 7200.10 for your parity disk to compaer apples to apples (mine is the 400GB baracuda variety).

Link to comment

What kind of temperatures are the rest of you seeing on your Parity drive during a parity check operation?  I want to validate my statements about "good cooling".  Ideally one of you who are using a Seagate 7200.10 for your parity disk to compaer apples to apples (mine is the 400GB baracuda variety).

 

Um, that was kind of a thread hijack.  ;D  There is some info here - http://lime-technology.com/forum/index.php?topic=174.0

Link to comment

Can a man not hijack his own thread?  :) Turns out my peak drive temps on the Seagate are not the coolest around, but well within comfortable operating range - I peak at 38 deg on the parity drive (7200.10 Seagate Barracuda) during a parity check.  Administrator posted drive temp specs < 55C are generally within spec.  Cooling is more than adequte (but not world class) on this rig.

Link to comment

IOZone seems like a nice tool for testing the performance of your own particular system. I'm not sure if it makes for a good benchmark to test against anyone else's system though. The performance of the server looks to be directly related to the horsepower of the workstation running the test. While running it on my own system, I saw my HTPC's CPU utilization hit <95% for a good portion of the test. This was with all other applications shut down, just IOzone running. In reporting results here, it might be useful to post what your workstation's specifications are as well. My HTPC is a little outdated but it still manages to serve up my movies and run FFDshow just fine (albeit at ~80% CPU utilization during playback).

 

So with that said, here's my specification & results:

 

Media Server:

 

ASUS P5PE-VM

Pentium 4 (Prescott), 3 Ghz, 800Mhz FSB

Corsair DDR400, 1GB

Coolermaster Stacker

-4-in-3 x three modules

-Couple of case quiet fans

- Dual 500W redundant PS's

- 800VA APC UPS

Sandisk Cruzer Micro, 512MB

Seagate 500GB, 7200, 16MB, SATA2, 500GB (Parity)

" " (Disk 1)

Maxtor 300GB, 7200, 16MB, IDE (Disk 2)

Maxtor 250GB, 7200, 8MB, IDE  (Disk 3)

Maxtor 200GB, 7200, 8MB, IDE  (Disk 4)

Maxtor 250GB, 7200, 8MB, IDE  (Disk 5)

 

HTPC:

 

ASUS A7N8X

Athlon XP 2200 (1.8 Ghz)

Corsair PC3200, 512MB

Seagate 160GB, SATA

Edimax Gigabit PCI adapter

Pioneer 111D DVD Burner

Samsung SH-162D DVD reader ;)

Oddball midtower ATX case I stumbled across at Compusa awhile back. Placed on it's side, blends right into the rest of the home theater components.

500W Antec PS (quiet!)

MS XP MC Edition, Various MCE plugins and tweaks.

 

Write Performance:

                   IDE (Disk 5)                SATA-2(Disk 1)

32MB          25.5                             37.9 (oddity)

64MB          26.6                             24.4

128MB        23.7                             23.7

256MB        22.0                             22.8

512MB        21.5                             21.7

1GB            21.7                             20.3

 

Read Performance:

                  IDE (Disk 5)                SATA-2(Disk 1)

32MB          27.1                              27.2

64MB          27.4                              27.2

128MB        27.3                              27.5

256MB        27.4                              27.4

512MB        27.3                              27.7

1GB            14.4                              17.8

 

 

Notes/misc:

 

- Unraid Server version 3.0 (Final)

 

- I also built my server over the Black Friday sale. Total cost was $790 shipped. Build cost included the server itself, the two 500GB drives and the gigabit NIC for my HTPC. That total does not include the 12 drive version of Unraid or the UPS ($35 on Fleabay with a new battery!).

 

- HTPC & Server are connected via a dedicated crossover cable (cat 5E, ~30ft), no switches, static IP's.

 

- HDD temps (according to SMART) typically sit between 17C and 30C (peak on parity drive). I would attribute this to the cooler room where the server sits (~65 F) and the excellent case ventilation afforded by the Stacker case. When my drives used to reside in my HTPC they would run extremely hot. I experienced multiple drive failures over the course of 3 years or so. While the Stacker case is certainly a behemouth, it sits in a spare bedroom out of sight quietly churning away.   

 

- For "real world" file transfer tests, I use a little shareware app called Windows Commander. I've been using it for file management tasks since 1995 in various iterations. I typically see 14,000 - 18,000 kbytes/second when moving DVD vobs over to the server.   

 

- With a *proper* PC software solution foir HD-DVD and Blu-Ray just on the horizon, I'm a bit motivated to finally upgrade the ole' HTPC. The new HTPC I'm building here shortly will utilize an NVIDIA nForce 680i chipset based Mobo, 3 Ghz Core2 Duo, 2 GB ram, Foxconn GEFORCE 7950, 512MB, dual head DVI video card and Vista. Once this is stood up, I'll run IOzone again and post the results here. I'm guesing there may be a bit of variance.  ;)

 

 

Ryan

 

Link to comment

Sorry no IOZONE specs but I do have some power/temps:

 

Cooler Master eXtreme Power RS-430-PCAR 430W ATX 12V V2.01 

Cooler Master STC-T01-UBK Blue CM Stacker Full Tower Case

 

Motherboard:  ASUS P5LD2-VM R2.0 (default bios cfg)

 

Intel Celeron D 352 Cedar Mill 3.2GHz 

CORSAIR ValueSelect 512MB (2 x 256MB) 240-Pin DDR2 SDRAM DDR2 533

 

Patriot 1GB Flash Drive 32MB/s Read 26MB/s write

 

ICY Dock 5 in 3 SATA III Bay

Drives:

(2) 500GB Maxtor Maxline Pro

 

=

Power in Watts

 

Drives writing/parity Sync      107

Drives Idle                              96

Drives Spun down                  80

Suspended                              01

 

Temps (Ambient ~70F)

 

Idle   28C/29C degrees

Writing 31C/32C

 

Link to comment

I will purchase a power source with Active PFC (Power Factor Correction) because it will reduce the power bill:

 

"Power Factor represents the ratio of usable working power and the apparent power. Fox example power source equipped with Active PFC of 99% is able to use power up to 99% with only 1% loss. Most of passive PFC power supplies can only use about 70%, which surely increases electric utility bills."

(In http://metku.net/index.html?path=reviews/seasonicsupertornado350w/index_eng)

 

Someone know the real reduction in power (watts) it will be?

 

Also, I still wish to understand the smaller 1GB data transfers when reading.

 

Link to comment

SergioK:

 

Improving the power factor in a residential application is probably not going to change your power bill.  Though large industrial customers are billed by both Watts consumed and power factor, residential sites generally are billed only by Watts (actually Kilo-Watt-Hours) consumed.

 

Having said that, power factor correction doesn't change the amount of real power consumed by your device (a 500W supply will still consume approx 500W at full load).  It only changes the amount of current drawn to create that power.  So unless your utility company charges you for "Peak Demand" or "Power Factor Surcharge" (This is rarely the case for residential customers) your billing will not change.  A more important metric for a power supply is the efficiency.  A highly efficient power supply is able to transform your power without generating much heat (the heat is where an inefficient power supply is throwing away watts).

 

To put all of this in perspective, I believe most residential customers pay around 5-6 cents per KWH, so if you have a 500W power supply under full load you are spending 2.5-3 cents an hour to power it.  That's a maximum of $21.60 per month (if running continuously).

 

Hope this helps.

 

Here's a relatively accurate and more detailed explanation for anyone who cares to read more about power factor correction:

http://en.wikipedia.org/wiki/Power_Factor_Correction

 

Link to comment

 

> Hope this helps.

 

Yes this helped!

 

I will look for a good power source - the one from my main machine throw a good amount of hot by itself (warm air even with the case open)... Not only because the power bill, I think the server will be more dependable if all stays cold when not being used...

 

Link to comment

Anyone do any research on fast hard drives so we can test out the maximum performance of the unRAID server?

 

I found two very interesting technology developments today on the FAST hard drive front:

 

First (though the capacity would be of limited value in an unRAID system), this would be an INCREDIBLE boot drive for your workstation operating system:

 

SanDisk SSD UATA 5000

This is a 32GB solid state (i.e. flash memory) hard drive that claims 100x the performance of any hard drive:

http://www.sandisk.com/Oem/Default.aspx?CatID=1478

 

Note that they are comparing it to a 4200 RPM Notebook Hard Drive, but I'm still guessing this would still utterly crush a RAID 0 pair of Raptor 150's.

At a retail price of $600, however, it is also going to cost more than a RAID 0 Raptor setup with 1/10 the capacity. 

 

 

Second, the 1TB hard drive I have been waiting for has finally arrived:

 

Hitachi Deskstar 7K1000

http://www.hitachigst.com/portal/site/en/template.MAXIMIZE/menuitem.368c8bfe833dee8056fb11f0aac4f0a0/?javax.portlet.tpst=74ef8e8d695bcd876ccf7be1cf4362b4_ws_MX&javax.portlet.prp_74ef8e8d695bcd876ccf7be1cf4362b4_viewID=content&javax.portlet.prp_74ef8e8d695bcd876ccf7be1cf4362b4_docName=20070105_first_terabyte_hd.html&javax.portlet.prp_74ef8e8d695bcd876ccf7be1cf4362b4_folderPath=%2Fhgst%2Faboutus%2Fpress%2Finternal_news%2F&beanID=804390503&viewID=content&javax.portlet.begCacheTok=token&javax.portlet.endCacheTok=token

 

This would be a nice size for a parity drive on unRAID, but with 8.7ms seek time, it isn't the most amazing device ever made (The seek time of the 150GB Raptor is only 5.0ms, 4.7ms for the older 74GB.  Note, however, that this new drive has a 32MB buffer - which is something I haven't seen before.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  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.