WHS Performance

[mtber]

Hi,

Just built a WHS system and I’m looking for some feedback on performance. Some IOZone testing results are attached.

The IOZone command was: iozone -Rab Test2.wks -i 0 -i 1 -+u -f s:\test\test.tmp -r 64k -n 64k -g 1g -z

My network and systems are outlined below.

How do these performance metrics look? About what I should expect, or are there opportunities for improvement?

The primary uses for the WHS system are to store Photoshop Elements 7 files and catalogs for shared use between the four PCs, store iTunes files and catalogs, and for backups.

All NICs in the systems at their default settings.

Thanks for any input.

John

Network

Cable modem to
Linksys WRT150N to
D-Link DGS-2208 10/100/1000Mbps 8-Port Desktop Green Ethernet Switch to
PCs below


WHS
AMD Sempron 140 Sargas 2.7GHz Socket AM3 45W Single-Core Processor
GIGABYTE GA-MA785GM-US2H AM3/AM2+/AM2 AMD 785G HDMI Micro ATX AMD Motherboard
Patriot Extreme Performance 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Seagate Barracuda 7200.12 ST31000528AS 1TB 7200 RPM SATA 3.0Gb/s 3.5" Internal Hard Drive
2 x Seagate Barracuda 7200.10 ST3500630AS 500GB 7200 RPM SATA 3.0Gb/s Hard Drive


Dorene - XP

AMD Athlon 64 X2 4600+ Windsor 2.4GHz Socket AM2 Processor
Foxconn FV-N84SM2DT GeForce 8400GS 256MB 64-bit GDDR2 PCI Express x16 Video Card
GIGABYTE GA-M68SM-S2 AM2 NVIDIA GeForce 7025 Micro ATX AMD Motherboard
Transcend JETRAM 4GB (4 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model JM2GDDR2-8K


Kelsey – Vista 64

ASUS M2N-SLI Deluxe AM2 NVIDIA nForce 570 SLI MCP ATX AMD Motherboard
MSI NX8600GT-T2D256E OC GeForce 8600GT 256MB 128-bit GDDR3 PCI Express x16 SLI Supported Video Card
OCZ SLI-Ready 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Crucial Ballistix 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Seagate Barracuda 7200160GB 7200 RPM SATA 3.0Gb/s Hard Drive

Eric – Vista Home

AMD Phenom II X3 720 2.8GHz Socket AM3 95W Triple-Core Black Processor
GIGABYTE GA-MA785G-UD3H AM3/AM2+/AM2 AMD 785G HDMI ATX AMD Motherboard
CORSAIR XMS2 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 1066 (PC2 8500)
Western Digital Caviar Blue WD5000AAKS 500GB 7200 RPM SATA 3.0Gb/s 3.5" WHS
nVidia GPS 250

John – XP Media Center

Dell Inspiron 9300
2.00 gigahertz Intel Pentium M
2 GB RAM
100 MB 7200 RPM HD

[thiggins]

Sorry, but the thumbnails are too small to read.

[mtber]

Sorry about that.

I've replaced the figures with PDF files, hopefully they will be legible.

Thanks,

John

[thiggins]

Thanks. Write results are dominated by write cache effects, which is responsible for all results > 125 MB/s. You really need to set the iozone -g switch to a value higher than the amount of memory in the NAS under test. Since you have 2 GB in the WHS machine, you need to set it to at least 4G.

The read speeds from the more powerful systems are reasonable @ 60 - 70 MB/s or so.
To get higher than that, you'll probably need to put a RAID 0 array of two SATA drives on each client and use Vista SP1 or better (or Win 7).

[mtber]

Thanks Tom.

The read feedback is reassuring. I'll have to get to some time to repeat the writes at the higher size.

From a general principles standpoint, would there be significant/worthwhile benefit in bumping up the WHS RAM from 2 gb to 4GB? I can pick-up another 2GB for $33, just wondering if it will/should make any difference.

Thanks,

John

[00Roush]

My guess is that adding more memory will not really change your speeds at all. It really depends on how large of files you are transferring and how much data you are moving in and out of the server at a given time. More memory should allow a busy server to keep more data cached and allow maximum disk throughput. So if you are transferring large files from\to multiple clients it could improve speeds a bit.

For reference a while back I upgraded my server from 1GB of RAM to 2GB of RAM and overall I have not seen any faster transfer speeds than before. Transfer speeds are still on average 80-100 MB/sec for large files. I have noticed that remote desktop works faster though.

00Roush

[mtber]

Thanks 00Rousch, I'll stick with my 2 GB Ram for now.


I've done some real world file copy tests and am very disappointed relative to the IOZone results. (haven't had a chance to run it at up to 2GB size yet).

It looks like I am getting real world transfer speeds of 8-25 MB/s unless I copy from disck to disk (Kelsey to eSata) on the same system where I get ~80 MB/s.

Results from some tests copying a 1 GB folder with 351 files (ranging from small up to ~ 40 MB, with most arounf 6-8mb) are below.



From To Sec MB/Sec Networking CPU

John Kelsey 113 8.8 15-20% 30-45%

John WHS 132 7.6 15-20% 30-45%

Kelsey eSATA 12 83.3

Kelsey WHS 47 21.3 ~35%

WHS Kelsey 38 26.3

WHS Dorene 75 13.3 ~12% 5-10%



It doesn't look like the newtwork or CPU usuage is saturated and I get good speeds when going disk to disk on a single system, so doesn't that indicate that it isn't disk speed limited?

I've also posted screen captures of a much larger transfer from the WHS to the eSata drive on the Kelsey Vista 64 system where the sustained rate was ~6 MB/S.

Am I expecting too much or is something seriosuly out of whack?

(sorry about the small attachement size, I don't understand why they display so small after uploading)

Thanks,

John

[benogil]

You might try some of the tips from here -
http://forum.wegotserved.com/index.php?/topic/7316-whs-vista-gigabit-slow-transfers/

Ben

[wathman]

Your network should be capable of higher transfer speeds, one question I have right off is what hard drive do you have installed in Kelsey? did a little searching, and the only 160 GB SATA II hard drives from Seagate I could find are the 7200.7 and 7200.9. Both of these have pretty bad performance in comparison to the newer drives you have installed in the other systems.

That would only partly explain the poor performance for Kelsey, but your other systems are having slower speeds as well so it may be that your WHS system's processor isn't keeping up with the load, it is only a low power sempron running WHS as an OS.

Do you have any data from resource usage during transfers from the WHS side?

The rest of the WHS build looks very solid so I'm not too concerned there. I'm also assuming you took care of the basics like ensuring all your cables are cat 5e or above.. if your network is anything like mine was, I cobbled it together over the years using accumulated extra ethernet cables, only just recently switched out everything to at least cat 5e.

[mtber]

The Seagate 160 in the Kelsey PC is older and probably not optimal (I pirated the two 500 GB drives now in the WHS from her system which I had been using as the "home server"). As you note, I don't think that is the overall problem though.

I connected to the WHS via Remote Desktop and copied the same test files from the WHS to the John system, 8.9 MB/s. The CPU usage on the WHS's Sempron was ~50%. Are there other parameters that would be useful to monitor?

I also did a copy test between John and Dorene, and had similar results, 8.5 MB/sec.

This is all making me think it is a network issue. I have Cat 6 running between all of the PCs and the D-Link DGS-2208 switch.

John


From To Sec MB/Sec Networking CPU

John Kelsey 113 8.8 15-20% 30-45%

John WHS 132 7.6 15-20% 30-45%

Kelsey eSATA 12 83.3

Kelsey WHS 47 21.3 ~35%

WHS Kelsey 38 26.3

WHS Dorene 75 13.3 ~12% 5-10%

WHS (RDC) John 8.9 ~50% Semperon usage
John Dorene 8.5

[mtber]

I also did an xcopy from John to WHS and the speed was similar to my File Explorer copies, 8.8 MB/S.

I looked through the link referenced above and didn't see anything that jumped out at me as being the same as what I'm seeing. I think most of teh problem reported there were slow transfer one way or under one set of conditions, were my set-up seems to be slow all around.

John

[wathman]

Have you tried any file transfers between your client machines, as in from Kelsey to John? If that transfers at a better rate, you'll know your infrastructure and switch are fine, and can focus on the WHS hardware and configuration.

Also, I'm just trying to think of stuff that might have been overlooked. Do you have all the latest service packs on your windows machines?

Something else to look at would be the event logs in your WHS Admin tools control panel. Maybe you'll find some clues as to what's causing the slowdowns there.

[Jay_S]

SNB has this how-to for checking your network infrastructure: Measuring Network Performance - Jperf (http://www.smallnetbuilder.com/lanwan/lanwan-howto/30408-measuring-network-performance-jperf). Jperf is a very handy tool. This will at least confirm that your cabling, switch, etc., is performing to spec.

[mtber]

Any transfers invloving an XP machine (desktop or WHS) is very slow. Vista (32) to Vista (64) transfers are significantly better

UPS just delivered a cross-over Cat 6 cable from NewEgg. I get a 13 MB/s cross over transfer between Kelsey (Vista 64) to John (XP), compared to 8.8 MB/s through my normal network set-up.

All of my transfer tests are summarized below (n-1 for each). I've also re-attached my original IOZone test results. (I'll try to check out JPErf when I get a chance).

I'm up to date on all service packs.

Do the totality of these results indicate that there are confirguration issues between the various machines and not a fundamental network issue? (IOZone results & poor cross over cable transfer speeds)

Thanks,

John

All transfer tests 351 files totaling 1 GB

Eric (Vista 32) to John (XP) 16.9 MB/S
Eric (Vista 32) to John (XP) 27.0 MB/S
John (XP) to Kelsey (Vista 64) 8.8 MB/S
John (XP) to WHS 7.6 MB/S
John (XP) to Dorene (XP) 8.5 MB/S
John (XP) to Eric (Vista 32) 7.6 MB/S
John (XP) to Kelsey 9Vista 64) - cross over 13.0 MB/S
John Xcopy to WHS 8.8 MB/S
Kelsey (Vista 64) to Kelsey (Vista 64) eSATA Drive 83.3 MB/S
Kelsey (Vista 64) to WHS 21.3 MB/S
Kelsey (Vista 64) to Eric (Vista 32) 37.0 MB/S
WHS to Dorene (XP) 13.3 MB/S
WHS to Kelsey (Vista 64) 26.3 MB/S
WHS (RDC) to John 8.9 MB/S

[Jay_S]

Kelsey (Vista 64) to Eric (Vista 32) 37.0 MB/S
This is with your crossover - yes? I didn't see Vista->Vista charts in your earlier posts, but it looks like these 2 PC's are your quickest.

1) Do you see Vista->Vista speeds of 37 MB/s through your network? Or just with the x-over cable?
2) What are the model#'s or P/N's of the HDD's in these PC's?

Read Tim's "How to Build a Really Fast NAS" series - especially Part 6: "The Vista Difference (http://www.smallnetbuilder.com/nas/nas-howto/30679-how-to-build-a-really-fast-nas-part-6-the-vista-sp1-difference)". You'll note a few things over the series:

1) XP's file copy engine sucks.
2) XP's SMB implementation sucks.
3) The mechanical, non-cache, speed of our HDD's is more limiting than we think.

Jperf (or iperf) will tell you how fast your network CAN be. Your drives, OS, protocols and applications will take it down from there.

[mtber]

The cross over was John (XP) to Kelsey (Vista 64) at 13.0 MB/S, compared to the same machines through the network at 8.8 MB/s.

Now all of these numbers are n's of 1, so the absolute values might change a bit if I was averaging three or more runs, but the ternds seem clear.

The only "decent" transfer I get is between Eric (Vista 32) and Kelsey (Vista 64), which are also my higest performing hardware.

Am I off base to expect better than <13 MB/s transfers between my XP machines and my WHS or low 20's from my Vista machines? In none of my testing have I seen anywhere near maxed out Network bandwidth or CPU load.

My systems are:

WHS
AMD Sempron 140 Sargas 2.7GHz Socket AM3 45W Single-Core Processor
GIGABYTE GA-MA785GM-US2H AM3/AM2+/AM2 AMD 785G HDMI Micro ATX AMD Motherboard
Patriot Extreme Performance 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Seagate Barracuda 7200.12 ST31000528AS 1TB 7200 RPM SATA 3.0Gb/s 3.5" Internal Hard Drive
2 x Seagate Barracuda 7200.10 ST3500630AS 500GB 7200 RPM SATA 3.0Gb/s Hard Drive


Dorene - XP

AMD Athlon 64 X2 4600+ Windsor 2.4GHz Socket AM2 Processor
Foxconn FV-N84SM2DT GeForce 8400GS 256MB 64-bit GDDR2 PCI Express x16 Video Card
GIGABYTE GA-M68SM-S2 AM2 NVIDIA GeForce 7025 Micro ATX AMD Motherboard
Transcend JETRAM 4GB (4 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model JM2GDDR2-8K


Kelsey – Vista 64 Home

ASUS M2N-SLI Deluxe AM2 NVIDIA nForce 570 SLI MCP ATX AMD Motherboard
MSI NX8600GT-T2D256E OC GeForce 8600GT 256MB 128-bit GDDR3 PCI Express x16 SLI Supported Video Card
OCZ SLI-Ready 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Crucial Ballistix 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)
Seagate Barracuda St3160812AS 7200160GB 7200 RPM SATA 3.0Gb/s Hard Drive

Eric – Vista 32 Home

AMD Phenom II X3 720 2.8GHz Socket AM3 95W Triple-Core Black Processor
GIGABYTE GA-MA785G-UD3H AM3/AM2+/AM2 AMD 785G HDMI ATX AMD Motherboard
CORSAIR XMS2 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 1066 (PC2 8500)
Western Digital Caviar Blue WD5000AAKS 500GB 7200 RPM SATA 3.0Gb/s 3.5" WHS
nVidia GPS 250

John – XP Media Center

Dell Inspiron 9300
2.00 gigahertz Intel Pentium M
2 GB RAM
Seagate Momentus 7200.1 ST910021A 100GB 7200 RPM 8MB Cache HD

[Jay_S]

Am I off base to expect better than <13 MB/s transfers between my XP machines and my WHS or low 20's from my Vista machines? In none of my testing have I seen anywhere near maxed out Network bandwidth or CPU load.
No, I'd expect 40's using XP as the client, and better using Vista/Win7.

Your WHS is plenty capable. The GIGABYTE GA-MA785GM-US2H's NIC is the Realtek 8111C, which is gigabit on the PCIe bus. So that's good. The 7200.12 in your server and the WD caviar blue in Eric are your fastest drives. The 7200.9 is slooow (I own one - 5-year warranty back then though!).

Part of the problem is the complexity of your project - you're trying to "solve" all these clients at once. Let's start with 2 and get them up to speed - I pick WHS and Eric (also has Realtek 8111C NIC).

Make sure both 8111C NICs have current drivers.

Re-cable them so they connect through your DGS-2208 switch. You *should not* ever need a crossover cable with gigabit ethernet (it auto negotiates regardless of cable type). Check windows NIC properties - connecting at 1000 Mb/s?

Don't worry about jumbos right now - in case you have them enabled in any of the NICs.

Can you run desktop applications on WHS? I'm thinking specifically of jperf (iperf with a Java GUI). We need to confirm that your NICs, cables, and switch are OK. With TCP window sizes of 64k you should see 900Mb/s. If not, something is wrong with a NIC, cable or your switch. Let's start here.

[mtber]

I should be able to get JPerf running on the WHS, will take me some time to get a chance to do it though.

All of my systems are running through the DGS-2208 switch with straight thorugh Cat 6. (I just used the cross-over to temporarily connect John & Eric for testing.)

Jumbos are disabled

All Nics say 1GB connections.

So, focusing on WHS & Eric, the current driver data is below.

I'm not sure why Sys Info reports a different NIC than the spec sheet (RTLE8023xp for WHS versus 8111C).

If I were to update drivers, would I use the ones from the following RealTek page?

http://www.realtek.com/downloads/downloadsView.aspx?Langid=1&PNid=13&PFid=5&Level=5&Conn=4&DownTypeID=3&GetDown=false

Version 6.227 for Eric (Vista 32) and 5.736 for the WHS?

WHS

NIC- Realtek 8111C
Sys Info reported -RTLE8023xp
Driver Date - 4/23/2009
Driver version - 5.724.423.2009


Eric

NIC-Realtek 8111C
Sys Info reported-RTL8169
Driver Date - 5/4/2009
Driver version - 6.222.504.2009

The full driver configuration for the WHS and Eric is:

WHS

NIC Realtek 8111C
Sys Info reported RTLE8023xp
Driver Date 4/23/2009
Driver version 5.724.423.2009

VLAn Tagging Disabled
Auto Disable PCIe Disabled
Auto Disable PHY Disabled
CheckSum Offload Rx & Tx Enabled
Flow Control enabled
Green Ethernet Disabled
Jumbo Frame Disabled
Large Send Offload enabled
Network Address not present
Shutdown Wake On lan enabled
Speed & Duplex auto negotiation
Wake on LAn Capabilities Pattern match and magic packet
WOL & Shutdown Link Speed 10 Mbps First

Eric

NIC Realtek 8111C
Sys Info reported RTL8169
Driver Date 5/4/2009
Driver version 6.222.504.2009

Auto Disable Gigabit disabled
Interupt moderation enabled
Auto Disable PCIe disabled
Auto Disable PHY disabled
TCP Checksum offload (IPv4) Rx&Tx enabled
TCP Checksum offload (IPv6) Rx&Tx enabled
UDP Checksum offload (IPv4) Rx&Tx enabled
UDP Checksum offload (IPv6) Rx&Tx enabled
IPv4 checksum offload Rx&Tx enabled
Flow Control enabled
Green Ethernet disabled
Jumbo Frame Disabled
Large Send Offload (IPv4) enabled
Large Send Offload v2 (IPv4) disabled
Large Send Offload v2 (IPv6) enabled
Priority & Vlan enabled
Receive buffers 512
Receive side scaling enabled
Transmit buffers 128
Network Address not present
Shutdown Wake On lan disabled
Speed & Duplex auto
Wake on LAn Capabilities none
WOL & Shutdown Link Speed 10 Mbps First


Thanks,

John

[Jay_S]

I'm pretty sure you're supposed to use WinXP / Server 2003 Driver (5.736 as you noted) for WHS. More here: How to update Realtek NIC drivers during Bare Metal Restore (http://social.microsoft.com/Forums/en-US/whsfaq/thread/b14e3cf2-7b02-4569-8c26-c75b7a58d6e7). The MS forums are filled with people having issues with Reatek NICs on WHS.

You can find more via google - search this string (without quotes): "8111C +site:forum.wegotserved.com".

I'm not familiar with WHS. Can you log on locally? Like, using a KB/mouse & monitor attached to the WHS? Just asking because, if driver installations or re-installations kill your network connection, you'll have no other way to admin the WHS.

You can always throw money at the problem! A Intel PCIe NIC is only $30 (http://www.newegg.com/Product/Product.aspx?Item=N82E16833106033), and Intel NICs are supported by nearly everything.

[wathman]

Something else I came across while troubleshooting my desktop's I/O performance: I wasn't getting expected performance out of my drives so I tried some tweaks. Turns out that power management can unpredictably affect performance if drivers aren't optimal. See if changing your power scheme to Performance and see if that changes your results any.

[wathman]

You can always throw money at the problem! A Intel PCIe NIC is only $30 (http://www.newegg.com/Product/Product.aspx?Item=N82E16833106033), and Intel NICs are supported by nearly everything.

Yeah, I second this idea. I ended up doing this with a previous build and I was trying to get Untangle Server up as a firewall. Was having issues with my built in NIC, so I got impatient and bought a Dynex gigabit PCIe adapter from the local Best Buy. Worked perfectly out of the box.

[mtber]

I can hook up a monitor and keyboard to the WHS if I get stuck.

I tried some transfers with my current power scheme on Eric and in Performance mode. There wasn't much of a difference between the two. It was interesting that successive runs got faster then hit a plateau.

Performance Power Mode 20, 25, 31, 30 MB/S
Typical Power Saving Mode 18, 22, 25, 28 MB/S

I’ll check out the Realtek drivers link.

I don’t really have the cash to get new NICs for all of my systems at this time, so I hope that something can be sorted out.

I got JPerf running and the results for Eric as client to WHS as server are below with the default parameters.

Anything I should set differently to trouble shoot?

bin/iperf.exe -c 192.168.1.101 -P 1 -i 1 -p 5001 -f M -t 10
------------------------------------------------------------
Client connecting to 192.168.1.101, TCP port 5001
TCP window size: 0.01 MByte (default)
------------------------------------------------------------
[128] local 192.168.1.105 port 49393 connected with 192.168.1.101 port 5001
[ ID] Interval Transfer Bandwidth
[128] 0.0- 1.0 sec 34.7 MBytes 34.7 MBytes/sec
[128] 1.0- 2.0 sec 31.7 MBytes 31.7 MBytes/sec
[128] 2.0- 3.0 sec 32.0 MBytes 32.0 MBytes/sec
[128] 3.0- 4.0 sec 32.0 MBytes 32.0 MBytes/sec
[128] 4.0- 5.0 sec 34.0 MBytes 34.0 MBytes/sec
[128] 5.0- 6.0 sec 33.5 MBytes 33.5 MBytes/sec
[128] 6.0- 7.0 sec 33.9 MBytes 33.9 MBytes/sec
[128] 7.0- 8.0 sec 30.6 MBytes 30.6 MBytes/sec
[128] 8.0- 9.0 sec 32.4 MBytes 32.4 MBytes/sec
[128] 9.0-10.0 sec 32.9 MBytes 32.9 MBytes/sec
[128] 0.0-10.0 sec 328 MBytes 32.8 MBytes/sec
Done.

[00Roush]

I usually use this command line for Iperf as it gives results in Mbps... iperf -w 64k -c 192.168.0.2. As Jay_S said, if everything is working correctly you should be in the 900 Mbps range.

Your network setting looked fine except for the different driver versions. I always go directly to the manufacturer for network drivers. They have the most up to date drivers and in many cases extra drivers for different OSes. I would think that both the Vista and WHS (Win XP/Server 2003 just like Jay_S mentioned) drivers should have nearly the same list of options. Makes me wonder if something might be a little off there.

So far I have not read anything in this thread about what exact version of Vista (on Eric) or WHS you are running. Could you let us know which service packs you are running?

00Roush

[00Roush]

Almost forgot... if possible try turning off I IPv6 in Vista and WHS. Should be able to do this in the properties for a particular network connection.

00Roush

[mtber]

The Os's and versions are summarized below.

WHS Windows Home Server ver 5.2.3790 Service Pack 2 Build 3790
Eric Vista™ Home Premium (32 bit) ver 6.0.6002 Service Pack 2 Build 6002
John XP Media Center 2004 ver 5.1.2600 Service Pack 3 Build 2600
Dorene XP Media Center 2004 ver 5.1.2600 Service Pack 3 Build 2600
Kelsey Vista™ Home Premium (64 bit) ver 6.0.6002 Service Pack 2 Build 6002

I've also attached an Excel file summarizing the NIC configurations on each machine.

I did find some mismatches between duplex and a couple of other settings as noted in the Excel file. Have harminized them as noted, but again no really noticable differences in JPerf results.

I haven't had a chance to update the NIC drivers on Eric and the WHS yet.

I have been looking for drivers for the other machines (Realtek 8211BL & Marvell PHY/nVidia Network controllers) and it's driving me nuts. I can't seem to find anything for them . (I thought that I had one from D-Link's site for the D-link DGE-660TD PCMCIA card in John, but when I try to insall it I get a message that it is the same as alredy installed - even though the version numbers are different)

I turned off IPv6 on Eric, didn't seem to make much difference. Wasn't installed on WHS or either of the XP machines.

I understand the wisdom of focusing on one problem at a time (Eric to WHS), but am a little baffled as to why it seems to be very sub-par performance no matter which machine combos the transfer is tested between. Maybe I'll try another direct connect cross over cable test between two of the systems - if no significant difference, then that eliminates network switch and cables and points to NIC to NIC issue (or software/driver) right?

Maybe I'll have to break down and get a couple of PCI-E cards to try as suggested - just hate to keep throwing money at the problem. (I used to have all of our shared files on the Kelsey Vista 64 machine and built the WHS in hopes of better access performance.)

Thanks for all of the help,

John

[Jay_S]

John,

I just noticed, in your jperf results it looks like you forgot to set TCP Window @ 64k.

I was sorta joking about throwing money at more nics. I don't think you should need to do that. I have the same 8111C in three computers, including my unRAID server. Unfortunately, unRAID doesn't ship with iperf and I don't know how to add it. But jperf between my gaming PC and HTPC looks like this:

http://img188.imageshack.us/img188/7569/clipboard01ur.th.jpg (http://img188.imageshack.us/img188/7569/clipboard01ur.jpg/)

A bit under 800MB/s. The only thing I tweaked was TCP Window Size - to 64k from whatever XP defaults to (I think 8k). I'm losing a little speed right now because I'm transcoding video over my network ;)

On one side is a Gigabyte GA-G31M-ES2L. Realtek 8111C with driver version: 5.720.327.2009 (3/27/2009). XP Pro x86.

The other is a Gigabyte GA-MA74GM-S2, 8111C, driver version 5.708.1030.2008 (10/30/2008). XP Pro x86.

I wired my house with cat6, through a patch panel, through a TRENDNet TEG-S80g unmanaged gigabit switch.

Re-try your jperf test with TCP WIndow set to 64k. I bet you get better results.

Try jperf/iperf between some other machines. Any machine with a gigabit NIC on the PCI express bus should do 800+ MB/s, assuming everying's working properly. All of your machines except for John appear to have their NICs attached to the PCIe bus or directly to their chipsets. This should all be good enough. Try between any of them except for John.

[mtber]

Awesome,

Here's the JPerf output (Kbits/sec) with TCP Window Size set to 64 K.

So, does this mean that I need to set my Vista and XP systems to 64 K TCP Window size? (or determine optimal if I get really adventeruous)?

(and update the drivers that I can find)

Thanks,

John


bin/iperf.exe -c 192.168.1.101 -P 1 -i 1 -p 5001 -w 64.0K -f k -t 10
------------------------------------------------------------
Client connecting to 192.168.1.101, TCP port 5001
TCP window size: 64.0 KByte
------------------------------------------------------------
[128] local 192.168.1.105 port 49936 connected with 192.168.1.101 port 5001
[ ID] Interval Transfer Bandwidth
[128] 0.0- 1.0 sec 97664 KBytes 800063 Kbits/sec
[128] 1.0- 2.0 sec 105656 KBytes 865534 Kbits/sec
[128] 2.0- 3.0 sec 111296 KBytes 911737 Kbits/sec
[128] 3.0- 4.0 sec 109440 KBytes 896532 Kbits/sec
[128] 4.0- 5.0 sec 101240 KBytes 829358 Kbits/sec
[128] 5.0- 6.0 sec 99240 KBytes 812974 Kbits/sec
[128] 6.0- 7.0 sec 110872 KBytes 908263 Kbits/sec
[128] 7.0- 8.0 sec 110896 KBytes 908460 Kbits/sec
[128] 8.0- 9.0 sec 110280 KBytes 903414 Kbits/sec
[128] 9.0-10.0 sec 110320 KBytes 903741 Kbits/sec
[128] 0.0-10.0 sec 1066912 KBytes 873927 Kbits/sec
Done.

[00Roush]

Awesome,

Here's the JPerf output (Kbits/sec) with TCP Window Size set to 64 K.

So, does this mean that I need to set my Vista and XP systems to 64 K TCP Window size? (or determine optimal if I get really adventeruous)?

(and update the drivers that I can find)

Thanks,

John


bin/iperf.exe -c 192.168.1.101 -P 1 -i 1 -p 5001 -w 64.0K -f k -t 10
------------------------------------------------------------
Client connecting to 192.168.1.101, TCP port 5001
TCP window size: 64.0 KByte
------------------------------------------------------------
[128] local 192.168.1.105 port 49936 connected with 192.168.1.101 port 5001
[ ID] Interval Transfer Bandwidth
[128] 0.0- 1.0 sec 97664 KBytes 800063 Kbits/sec
[128] 1.0- 2.0 sec 105656 KBytes 865534 Kbits/sec
[128] 2.0- 3.0 sec 111296 KBytes 911737 Kbits/sec
[128] 3.0- 4.0 sec 109440 KBytes 896532 Kbits/sec
[128] 4.0- 5.0 sec 101240 KBytes 829358 Kbits/sec
[128] 5.0- 6.0 sec 99240 KBytes 812974 Kbits/sec
[128] 6.0- 7.0 sec 110872 KBytes 908263 Kbits/sec
[128] 7.0- 8.0 sec 110896 KBytes 908460 Kbits/sec
[128] 8.0- 9.0 sec 110280 KBytes 903414 Kbits/sec
[128] 9.0-10.0 sec 110320 KBytes 903741 Kbits/sec
[128] 0.0-10.0 sec 1066912 KBytes 873927 Kbits/sec
Done.

Looks like things are working just fine on the network side, but I would still update your drivers.

Well you can't really set Windows to use a particular window size. You can set the maximum but generally from Windows 2000 and up the window size is auto tuned. In my experience Windows will use a window size of 60k-64k on a small local LAN.

So now you have to move onto other things that could cause your problems. It is quiet odd though that all of you machines are having slow file transfers. Have you tried switching your test to using a single large file of around 512 MB-1GB?

00Roush

[mtber]

Updated the drivers on Eric (Vista 32) and WHS. No significant difference in Windows Explorer file copy speed.

Tried a single file 1GB copy versus same files as 351 inviduals in a folder. Single 1GB file transfered at 50 MB/Sec versus 26 for the collection.

Obviously nothing is simple ... The reason I asked about setting the TCP Window Size to 64 kbits is that when I Googled setting teh Window size I saw lots of web pages about problems with Auto Tuning and instructions for XP registery editing to set the window to 64 K. () So, not a good idea to jump to any of those measures (at least yet) in my case?

I don't have any data to support it, but I suspect that my slow transfer speeds problem existed before I added the WHS. I wonder (but don't know how to investigate) if there is some underlying conflict problem that causes re-sends as the copy progresses. I had noticed that on very large transfers using the Vista 64 system (Kelsey) that they would start off fast and in short order slow to a crawl (<10 MB/s, like down to 2-3).

Thanks,

John

[Jay_S]

Awesome,
Indeed! This confirms that your NICs, cables and switch are OK. And defines the upper limit of what's possible. Now you have to test everything resting on top of that.

You'll need FAST drives on both ends. Once file sizes exceed available cache (drive cache, system RAM, CPU cache, etc.) of either system, your transfer speed will not exceed the speed of that systems HDD. For example, your WHS has 2GB of ram, Eric has 4GB. If you transfer a file larger than 2GB from Eric to WHS, you exceed WHS's cache and will be limited by the physical ability of the WHS's HDD to write data to the platters - regardless of Eric's performance.

You can use iozone to test local disk speed. As Tim suggests in post #4, make sure you test file sizes larger than the available system ram. If you have 2GB of ram, test up to 4GB file sizes. Your original iozone command:
iozone -Rab Test2.wks -i 0 -i 1 -+u -f s:\test\test.tmp -r 64k -n 64k -g 1g -z
...tests from 64k to 1GB. Bump that up to something higher. I started at 256m for a run this morning:
http://img27.imageshack.us/img27/3007/720012iozone.th.jpg (http://img27.imageshack.us/img27/3007/720012iozone.jpg/)
So, roughly 73-77 Mb/s non-cached writes, 115 MB/s non-cached reads. This is the fastest drive in my house.

Run iozone locally on Eric's WD Blue drive (which should be the 2nd fastest drive you own, next to the 7200.12 in WHS). Let's see where that goes.

[Jay_S]

I had noticed that on very large transfers using the Vista 64 system (Kelsey) that they would start off fast and in short order slow to a crawl (<10 MB/s, like down to 2-3).
Kelsey has the Seagate 7200.9. I also own this drive, and they're slow. See Anandtech's HD Tach results (http://www.anandtech.com/storage/showdoc.aspx?i=2577&p=8): at best 65MB/s sequential reads. At worst around 35 MB/s. And I've found HD Tach to be pretty optimistic! Either way, assuming it's really capable of 65MB/s, once you factor in SMB peculiarities, file copy engine performance, network protocol overhead, etc., I wouldn't be surprised if you saw half of that in real life.

This doesn't explain why your performance drops off like that. I'm just saying you're not starting with a quick drive.

[mtber]

If I can get these issues sorted out, I may need to put a better drive in Kelsey since I pirated the better 500 GB drives out of her system to build the WHS.

Ok, a couple of plots (full sacle and expanded) from two IOZone runs from Eric to WHS today (with read/write data from 18-Sep for comparison) are attached.


Thanks,

John

[00Roush]

If I can get these issues sorted out, I may need to put a better drive in Kelsey since I pirated the better 500 GB drives out of her system to build the WHS.

Ok, a couple of plots (full sacle and expanded) from two IOZone runs from Eric to WHS today (with read/write data from 18-Sep for comparison) are attached.


Thanks,

John

How full is the drive in Eric?

00Roush

[mtber]

used 159 GB, free 305

defrag is scheduled on a weekly basis

[mtber]

WHS Disk Usage

The primary is 23% full and the two 500 GB drives are each 36% used.

[00Roush]

With your current disk usage levels I still think you should be able to see higher speeds between those computers for the 1 GB file you tested earlier. From looking up the information about your drives I would guess that the disks would be capable of reading and writing at around 60-70 MB/sec in there current state. Add a bit of overhead for the file transfer and in my opinion you should be able to see 55-65 MB/sec. This is just a guess though and it also assumes that no other software or service is trying to use the disks. Anti virus/spyware software, file duplication, and system restore can all interfere with file transfers. So if possible disable all of these when testing. I suppose your 50 MB/sec for the 1 GB file could be right but I still think it is too low given your hardware.

I think Jay_S wanted you to test your C drive in Eric with Iozone and see what kinds of speeds you get for just local reads and writes. A command line something like this iozone -Rab c:\iozone\results.xls -i 0 -i 1 -+u -f c:\test.txt -y 64k -q 64k -n 256m -g 8G -z. It would be great if you could test the drives on your WHS machine as well but I don't know if you can use iozone on WHS. Basically this gives a good idea of what your actual read and write speeds for the drives you are using in their current state. If file copies over the network are within 5-10 MB/sec of the speeds seen locally chances are you won't see much better unless you swap to faster drives.

Do you have Cool n Quiet enabled on your server? I noticed in your screen shot from post #7 that you were at 40% of maximum frequency so I assume at least that machine is using Cool n Quiet. If so you might consider disabling it in your BIOS while testing. I just recently did some testing at home and found that when enabled on the server or client transfer speeds went down. When enabled on the server it was more noticeable as it dropped read and write speeds down 10-20 MB/sec. Not sure if it is just my setup or what but definitely something to consider.

I say you take your WHS and Eric machines and slim down all of the software to just the bare essentials running. Is any particular software on all of your PC's that might be causing a slow network. It could be something that looks harmless. Virus... Spyware? Figured I would throw that out there. Also make sure you are not playing any media while you are trying to transfer files to/from a Vista PC. Try using just IP addresses to connect to your WHS machine.

If you get a chance could you get a screen shot of the Resource Monitor during a 1GB file transfer writing from Eric to WHS or reading from WHS to Eric?

00Roush

[mtber]

Thanks. I'll try those suggestions as I gat a chance to tinker this week. I would like to create a table of measured HD performance and target throughput for all of my systems. Then I could see how close each is so that I'm not wasting everyone's time on chasing diminishing returns. As a get a chance to run the single disk IOZone tests, I'll add the info to the table.

I did run HD Tune on all of the drives last night - I'm trying to figure out how to poast in table format so it would be easy to compare.


Would using the average be a good approximation of max rate to expect to be able to approach? (knocking off some for other inefficiencies, etc?)

Thanks,

John

[mtber]

WHS primary drive

[Jay_S]

Hunting bottlenecks can be really frustrating. One might have a nice raid controller capable of 800MB/s local read/write performance, but that doesn't matter if you're bottlenecked by a gigabit LAN. One might have a gigabit LAN, which doesn't matter if the PC on the other end is incapable 100MB/s I/O. And so on.

After you've done a fair bit of bottleneck hunting, you'll settle on a methodology that works for you. My methodology has been developed around free software and lots of thinking, mainly because I have more time & energy than I have money! I start by testing low-level stuff first (the network infrastructure: NICs, cables, switches, etc.), using jperf/iperf. Then moving up to HDDs, using IOZone locally. I've found IOZone to provide a more useful picture of HDD performance than any of the HD Tune/Tach/ATTO etc., mainly because the user can custom tailor the test to better simulate their specific application. Then I test network HDD performance using IOZone. And finally I do some timed network file copy tests.

This gives 4 data points:
iperf results = the upper limit. Period. You can't exceed the "wire speed" of your network.
local IOZone results = the upper read/write limits to your client PC's HDD.
remote IOZone results = the upper read/write limits for networked file transfers.
network file copy tests = the actual read/write limits for networked file transfers.

You'll discover that network I/O is much slower than local I/O. This isn't surprising, but the reasons are not obvious. And there are lots of reasons, unfortunately! And most do not lend themselves to direct testing, so lots of research, discussion and reasoning is needed. Two reasons I am most comfortable discussing are:


Network transfers will always be slower than local transfers because of the way our networks work. Each layer of the network protocol stack chops up data and encapsulates it (http://www.tcpipguide.com/free/t_IPDatagramEncapsulation.htm) before sending it to the next network layer. Data is chopped and encapsulated several times before being sent out on the wire. The meta-data that encapsulates the actual data is "wasted space" on the wire - or, as it's typically called, overhead. This is why tricks like large TCP windows, jumbo frames, etc., sometimes work - they permit sending a higher ratio of data to meta-data (more data per header). Plus, TCP is a "reliable" protocol (it requires acknolwedgements), which adds another slow-down. Applications have their own overhead as well. Often, FTP file transfers are faster than Windows Explorer drag-n-drop file transfers because of what Explorer adds on top of all the overhead present in TCP be default. FTP is less convenient, though.
As you discovered in post #14 (http://forums.smallnetbuilder.com/showpost.php?p=13816&postcount=14), Vista copies faster than XP. Vista SP1 brought a lot of improvements to the windows file copy engine. I previously posted a link to Tim's The Vista Difference (http://www.smallnetbuilder.com/nas/nas-howto/30679-how-to-build-a-really-fast-nas-part-6-the-vista-sp1-difference), in which he posted a link to Mark Russinovich's Inside Vista SP1 File Copy Improvements (http://blogs.technet.com/markrussinovich/archive/2008/02/04/2826167.aspx) that discusses this at length - well worth the read. 00Roush has written about this (http://forums.smallnetbuilder.com/showthread.php?t=1802) as well in his sticky.

[mtber]

Jay_S & 00Roush, thanks for all of your help.

I'll get local IOZone tests run in the next couple of days.

My objective in all of this is just make sure my set-up is not unnecessarily hamstrung. I don't need to tweak the maximum performance out of it, just want to make sure my stock Camero isn't running like a Chevette.

After I get data on all four of Jay_S's points, I guess there will be a pretty good picture of where things stand.

I have taken a look at the XP versus Vista file copy links. I wonder what the story will be with Windows 7? I plan to get the Home Premium Family Pack and move the Eric, Kelsey, and Dorene systems to Win 7 in the next few months.

Thanks again,

John

[mtber]

yep, I probably have Cool & Quite enabled. Will have to check into it.


ok, ran the local IOZone tests on all of my systems last night. Results are attached. The first is a plot of Read results, the 2nd is Write results and the third is a summary of the machines and the full IOZone results for each.

What do they suggest?

Are the horrendous write results for the two IDE drives (John & Dorene) representative for IDE or is soemthing else up? (I had forgotten that an IDE was in the Dorene system, I guess that I need to upgrade her? John is an aging Dell Inspiron 9300 notebook, he will not be getting any upgrades.)

Thanks,

John

[Jay_S]

Now we're getting somewhere!

I'm pretty sure these IOZone tests measure sequential reads and writes. So, similar to a single large file being written/read. Results will be quite different for multiple smaller files.

Your WHS looks plenty fast: 100MB/s reads & writes.

Let's focus on Eric, as your fastest client (ignoring kelsey's eSata drive for the moment). The 4 & 8GB results suggest that Eric's HDD is physically only capable of: 51-52 MB/s reads and 45-50 MB/s writes. That's pretty poor by today's standards, as demonstrated by WHS's Seagate 7200.12 results.

If you're transferring a file from Eric to WHS, Eric's 51-52 MB/s read speed is your bottleneck. Factoring in network and application protocol overhead, the 50MB/s you reported here (http://forums.smallnetbuilder.com/showpost.php?p=13900&postcount=29) sounds pretty good, actually!

What do you see transferring to/from Kelsey's eSata drive to WHS? Kelsey is running Vista (SP1? check on that, as SP1 improved the file copy engine), and the eSata HDD is capable of 80MB/s reads - so I'd expect transfers from Kelsey's eSata drive to WHS to be better then Eric.

[wathman]

You had a question about how windows 7 performance stacks up, and since I've been able to do a little real world testing with the final version of Win 7 Ultimate 64-bit, (Was lucky enough to be a Win 7 Launch Party Host) My fastest transfers are between my core i7 860 (with the win 7 install) as client, and a C2D E6850 running Vista Ultimate 64-bit. Anything going between those two machines transfers at speeds between 100-107 MB/sec. Both machines are using RAID 0 arrays so I have a disk performance advantage on my server, but eventually I will migrate it to a RAID 5 array when I can invest a bit more in extra drives.

Basically, it should be about the same as Vista SP 2 performance, maybe a bit better on hardware limited systems as it is supposed to be more friendly with resource usage under win 7.

[mtber]

So the Eric system is performing pretty much as good as it will get with the current hard drive? Wow, there is a huge difference between the single 1 GB file transfer versus the 351 files totally 1GB (the latter being more typical usage for my systems right now) – 50 versus mid 20’s MB/S.

Kelsey is running Vista 64 - 6.0.6002 Service Pack 2 Build 6002

I don't have manual transfer data between Kelsey's eSata and WHS, I'll generate that tonight.

However, a massive copy (couple hundred gigs of thousands of 2-8 MB digital pics) from WHS to Kelsey's eSata is what got me concerned to begin with. That transfer started off fast (in the area of 80 MB/S), but within a minute or so dropped off to around 3 MB/S and took all night long. But,


That was with a different 500 GB eSATA drive, a Seagate Barracuda 7200.10 ST3500630AS versus the WDC WD5001AALS-00L3B Caviar black on there now (I have two external 500 GB eSata drives that I back up all of the family photos, etc to and always keep one at work in case the house burns down or something)
The Barracuda 7200.10 is older and not in the same league as the Caviar Black
I realize that is a much different copy than my little 1GB 351 file test set.

Here is the manual file copy data (MB/S) that I do have for Kelsey.

Kelsey (Vista 64) to WHS - 21
WHS to Kelsey (Vista 64) – 26

Kelsey (Vista 64) to Eric (Vista 32) – 37

Kelsey (Vista 64) to Kelsey (Vista 64) eSATA Drive - 83

John (XP) to Kelsey (Vista 64) – 9
John (XP) to Kelsey (Vista 64) – via cross over cable – 13

So, to round out the data set, I need to generate the large manual single file copy data to compare to the IOZone Local Test data I just generated.

Another question: The performance that I’m most concerned about is accessing digital photos stored on the WHS for reviewing and editing by either PhotoShop Elements (all catalogs, etc stored on WHS) or Lightroom (catalog local, images on WHS). Is the file read/write performance a reasonable surrogate for telling if the particular system is performing as well as it can with it’s hardware/OS (and maybe identifying what could be upgraded to improve performance)?

Thanks,

John

[mtber]

Nevermind that last question. On reflection, it's primarily the read performance off of the WHS that's going to impact performance when moving through a Photoshop Elements Catalog and pulling up a file to edit. Same for Lightroom, with a little more emphasis on the local machine since the catalog and thumbnails are local.

I think.

[mtber]

wathman, thanks for the Win 7 feedback.

Here's the manual Win Explorer copy data for Kelsey eSata (WDC WD5001AALS-00L3B Caviar black ) and the WHS


Single 1 GB file

Kelsey (Vista 64) eSATA to WHS 56 MB/S - maintained speed throughout copy
WHS to Kelsey (Vista 64) eSATA 30 MB/S - initiated from Kelsey - started >60 MB/s, finished ~30 MB/sec

1 GB folder of 351 files


Kelsey (Vista 64) eSATA to WHS 28 MB/S - steady speed throughout copy
WHS to Kelsey (Vista 64) eSATA 14 MB/S - steady speed throughout copy

[wathman]

seems like you're getting better performance numbers, but I think your older hard drive is doing more harm than good since it's contribution of virtual memory to the system is not exactly speedy. Since that system has 4GB of system RAM, it might be interesting to see what your transfer speeds do if you turn off virtual memory.

Also, I noticed that you are using different brands of DDR2 RAM, this in itself is not a problem, though I'm guessing one set may be better than the other set. In cases like that, the motherboard should automatically clock all your RAM to the slower of the two, but it may have done something strange. Might be worth at least verifying your RAM timings are set correctly on Kelsey.

[benogil]

This is a a response from another forum, with a parallel scenario. May be worth a try.
I've had similar problems with my WHS / W7 RTM setup and these seem to have helped, though I'm still early in testing.

http://support.microsoft.com/kb/948066

Try disabling this feature by changing the following registry setting to 0xffffffff (DWORD):
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\Network Throttling Index

You can also change
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\SystemR esponsiven ess
to 10 from 20.

You will need to reboot after making this change.

Other things to try (though it looks like you've adjusted adapter settings):
Network loss due to the PC network card:
- Right click on My Network Places and choose properties.
- Double click on the Local Area Connection icon to bring up the status page.
- Check that the Speed is reported as 100Mbps or 1Gbps.
- Click Properties. Check the brand/model of the network interface:
-- nForce based interface: Click Configure and switch to the Advanced tab. Disable checksum offloading features.
-- Intel based interface: Click Configure and switch to the Advanced tab. Test with alternative Interrupt Mitigation settings. onfigure are switch to the Advanced tab. Increase the Receive Buffer size (if this option is present). Test with the Interrupt Mitigation/Moderation setting both On and Off (if this option is present).

Ben

[00Roush]

wathman, thanks for the Win 7 feedback.

Here's the manual Win Explorer copy data for Kelsey eSata (WDC WD5001AALS-00L3B Caviar black ) and the WHS


Single 1 GB file

Kelsey (Vista 64) eSATA to WHS 56 MB/S - maintained speed throughout copy
WHS to Kelsey (Vista 64) eSATA 30 MB/S - initiated from Kelsey - started >60 MB/s, finished ~30 MB/sec

1 GB folder of 351 files


Kelsey (Vista 64) eSATA to WHS 28 MB/S - steady speed throughout copy
WHS to Kelsey (Vista 64) eSATA 14 MB/S - steady speed throughout copy



Hmmm... Was hoping to see a bit higher than that. Any chance you could move that eSATA drive over to the Eric system and run the same test? Better yet any chance you could move the drive from being in the eSATA enclosure to an internal drive? Technically eSATA should be just as fast but I am not quiet sure if that is ACTUALLY true when dealing with moving data across the network. I have never owned a eSATA enclosure so I am not sure if it makes a difference. Just trying to eliminate any possible bottlenecks.

You know since seeing how low your performance is I am considering testing WHS with Vista SP2 on my server and main PC. I have extra IDE drives I could use for the server to test with. Also I have been meaning to reinstall Vista. It has been so long since I set it up I don't recall what problems I had with it right out of the box. Haven't tested SP2 yet either.

00Roush

[00Roush]

This is a a response from another forum, with a parallel scenario. May be worth a try.
I've had similar problems with my WHS / W7 RTM setup and these seem to have helped, though I'm still early in testing.

http://support.microsoft.com/kb/948066

Try disabling this feature by changing the following registry setting to 0xffffffff (DWORD):
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\Network Throttling Index

You can also change
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\SystemR esponsiven ess
to 10 from 20.

You will need to reboot after making this change.

Other things to try (though it looks like you've adjusted adapter settings):
Network loss due to the PC network card:
- Right click on My Network Places and choose properties.
- Double click on the Local Area Connection icon to bring up the status page.
- Check that the Speed is reported as 100Mbps or 1Gbps.
- Click Properties. Check the brand/model of the network interface:
-- nForce based interface: Click Configure and switch to the Advanced tab. Disable checksum offloading features.
-- Intel based interface: Click Configure and switch to the Advanced tab. Test with alternative Interrupt Mitigation settings. onfigure are switch to the Advanced tab. Increase the Receive Buffer size (if this option is present). Test with the Interrupt Mitigation/Moderation setting both On and Off (if this option is present).

Ben

Just wanted to bring up a couple of points. First would be that normally these changes would only affect network performance while any type of media is being played. There was a thread (http://forums.smallnetbuilder.com/showthread.php?t=899) a while back about this. Next would be that HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\SystemR esponsiveness registry key sets the amount of CPU resources set aside for non media tasks. Default is 20%. This means that when media is being played all other tasks combined are limited to 20% cpu usage. So I believe this should be made bigger not smaller to allow other non media tasks more cpu usage. I currently have mine set at 50.

00Roush

[mtber]

00Roush

yep, I should be able to move the eSata to Eric and run the tests, will take me a through the end of the weekend to get a chance top do it.

I'm wishing that I had started looking at all of this stuff and established benchmarks before putting that older 160 GB in Kelsey. (I do have Passmark benchmarks on all systems before, haven't run them since putting the 160 in Kelsey. Are they of any value? - I'm assuming the IOZone and manual file copy data is "better".

Boy, I didn't realize that my HD's were so below par. Seems that just a couple of years ago, if you had a 7200 rpm drive I was happy. - although I guess the same would be said of the other major components.

I have ordered a Seagate Barracuda 7200.12 ST3500418AS 500GB to put in Dorene for my planned Win 7 upgrade. I don't need 500 GB in that machine now that I have the WHS, but it seems that it's not really significantly cheaper to go with less capacity from a price/performance perspective.

wathman,

I'll check the RAM timings and experiment with the virtual memory when I get a chance.


Ben,

The network speed is reported as 1 GB.

The network interface driver fields and settings on Kelsey are below. There are several CheckSum entries, would I disable them all? What about the Large Send Offload? (I see that I haven't disabled the IPv6 on this system either.)

Marvell PHY
Driver Date - 5/3/2007
Driver version - 65.7.4.0 - Nvidia Driver
Vlan ID - 1
Interupt moderation - Enabled
IP CheckSum Offload - Rx&Tx enabled
TCP Checksum offload (IPv4) - Rx&Tx enabled
TCP Checksum offload (IPv6) - Rx&Tx enabled
UDP Checksum offload (IPv4) - Rx&Tx enabled
UDP Checksum offload (IPv6) - Rx&Tx enabled
Flow Control - Rx&Tx enabled
Jumbo Frame - 1514 bytes (no options to disable, just chnage size)
Large Send Offload (IPv4) - Enabled
Large Send Offload v2 (IPv4) - Enabled
Large Send Offload v2 (IPv6) - Enabled
Priority & Vlan - Priority Enabled
Receive side scaling - Enabled
Transmit buffers - Enabled
Low Power State Link Speed - Enabled
Network Address - not present
Speed & Duplex - auto negotiation
Wake on LAn Capabilities - Enabled



Thanks,

John

[Jay_S]

I'm wishing that I had started looking at all of this stuff and established benchmarks before putting that older 160 GB in Kelsey. (I do have Passmark benchmarks on all systems before, haven't run them since putting the 160 in Kelsey. Are they of any value? - I'm assuming the IOZone and manual file copy data is "better".
Not better really, but probably not objectively comparable. In my experience, IOZone is ruthless in ways the other HDD benchmarks aren't.
Boy, I didn't realize that my HD's were so below par. Seems that just a couple of years ago, if you had a 7200 rpm drive I was happy. - although I guess the same would be said of the other major components.
I hear ya - this was a hard lesson for me as well. A 7200rpm drive is going to give you latency performance you can't get with slower rpm drives. But throughput depends on rpm + platter density (GB / platter). My 7200.9 was a spare I had from an old build. It replaced a failing 7200.11 in my htpc. The upside is that it's still under warranty! Since I rip directly to (and play from) my server, the slow HDD isn't a big deal.
I have ordered a Seagate Barracuda 7200.12 ST3500418AS 500GB to put in Dorene for my planned Win 7 upgrade. I don't need 500 GB in that machine now that I have the WHS, but it seems that it's not really significantly cheaper to go with less capacity from a price/performance perspective.
The 500GB 7200.12 is the one you want :) - all 500GB is on a single platter. High platter density is good for throughput, and with only 1 platter power consumption is low.

Having praised the 7200.12, magnetic storage is a dead end. Especially in your/my environment (desktop clients with mass available networked storage). SSDs are the future. They're expensive now (and NAND prices are crazy volatile), but no storage tech has had such a disruptive effect in a long time. Slow magnetic storage has been "good enough" for ... well, as long as magnetic storage has been around. Today you can get better performance than 2-drive RAID0 with a single SSD, with 10x the random I/O, 1/10th the power consumption, 1/100 the seek time, zero noise - in a smaller form factor to boot! I can't wait for SSD's to drop in price. Fingers are crossed for some Intel or Indilinx love around black friday.

[mtber]

The SSDs do look fantastic for this type of application - when the prices become reasonable. Thinking about that, I don't feel too bad about "wasting” the unneeded capacity of the 7200.12 500 GB drive. Once the SSDs come down to reasonable prices, I can put one in Dorene and move the fast 500 GB to expand the WHS storage – win/win.

The 7200.12 500 GB seems to be a sweat spot right now and hard to pass up at $55 with free shipping. A couple of months ago a comparable drive would have been $100.

Here’s the Passmark Performance test comparison for Kelsey (kind of hard to read the jpeg). The 28-Aug data was with a Seagate Barracuda 7200.10 ST3500630AS 500GB, Disk Mark 329 (slower than the 384 reported in Passmark's tables). The “This Computer” is the Seagate Barracuda 7200.9 ST3160812AS7200160GB, Disk Mark 409. Looks like she is actually marginally better off than when I had the 500’s in there. But, slow is slow.

[mtber]

duh, no eSata port on Eric, so won't be able to easily make that comparison.

[00Roush]

duh, no eSata port on Eric, so won't be able to easily make that comparison.

LOL! That's okay.

I have actually been doing some testing with WHS and Vista SP2. So far it has been interesting. I also now have a eSATA enclosure and a couple more 320 GB WD Carviar Blue drives to play with. Read speeds were seemed to be good but write speeds had some problems. I think I found out what my problem is so I am going to reload WHS and Vista to see where I am at. On a side note... From what I can tell the eSATA drive works just as fast as a internal drive. I will post up some of my results after some more testing.

During my testing I have come across quite a bit of information talking about problems with Windows Server 2003 SP2 network offloading features. WHS is based on this verison of Windows so the same problems could be in WHS. Not sure if you want to try it but these changes might help.

To manually disable RSS and TCP Offload yourself, follow these steps:

1. Click Start, click Run, type regedit, and then click OK.
2. Locate the following registry subkey:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\Tcpip\Parameters
3. Right-click EnableTCPChimney, and then click Modify.
4. In the Value data box, type 0, and then click OK.
5. Right-click EnableRSS, and then click Modify.
6. In the Value data box, type 0, and then click OK.
7. Right-click EnableTCPA, and then click Modify.
8. In the Value data box, type 0, and then click OK.
9. Exit Registry Editor, and then restart the computer.

This is straight out of this (http://support.microsoft.com/kb/948496) Microsoft KB.

You know I just realized a couple of things...

First was that I have not seen any file transfer tests between Eric and Kelsey's eSATA drive using a large single file.

Next would be that it might be helpful to see Iperf results between some of your other machines. Maybe do a Eric\Kelsey or Eric\Dorene test.

Last was when you ran the Iozone tests on your WHS drive you were using the C drive I assume. This means you were only testing the Seagate 7200.12 drive. My understanding is that WHS writes to the other disks before it writes to the drive that has the OS on it. This means that most likely the other two drives are actually the ones that need to be tested. The problem is that with WHS the storage setup is well... goofy. By default the data drives are not assigned a drive letter so testing them becomes a little more difficult. I think you can assign a drive letter with the computer management control panel but I am not sure if that works out. I will test this out and see if it works.

I hope we are not driving you nuts with all of this. By the way your presentation of your results has been excellent. Everything is well organized and easy to read/understand. Especially considering how much data has been collected in this thread.

00Roush

[00Roush]

Okay I have done a bit more testing. First up I was able to run Iozone tests on my data drive in WHS. It turns out that WHS does not assign drive letters to data drives but instead maps them to folders. For example my data drive was mapped to c:\fs\4. I found this by going to Control Panel\Administrative Tools\Computer Management then double clicking on Disk Management near the bottom.

On my system I had the following listed:
( E: ) (This was a USB flash drive)
DATA (This was my data drive)
DATA ( D: ) (This was the 2nd partition on the system disk that was created by WHS which is also used for data)
SYS ( C: ) (This was the 20 GB system partition created by WHS)

Once here I right clicked on the drive marked "DATA" that did not have a drive letter attached and selected "Change Drive Letter and Paths...". A box popped up that showed where the drive was currently mounted. Then I just used that path with Iozone.

iozone -Rab c:\iozone\results.xls -i 0 -i 1 -+u -f c:\fs\4\test.txt -y 64k -q 64k -n 256m -g 8G -z

Both read and write speeds were between 104000 and 95000 from 2 GB to 8 GB file sizes with this drive.

File copy speeds were between 90 and 100 MB/sec for a few test files ranging in size from 1.2 GB to 20.1 GB in size. I have the exact copy times and file sizes if you want them. The problem I mentioned with write speed was caused by a storage controller driver. I found the standard windows version works best for my setup.

Here is my setup:
Client:
Athlon 64 5400+ (2.8 Ghz) CPU
2 x 2GB DDR2 800 RAM
Asus M3A78-T Motherboard
Marvell Yukon 88E8056 PCI-E Gigabit Ethernet Onboard Motherboard
320 GB WD Caviar SE16 Hard Drive (OS drive)
320 GB WD Caviar Blue Hard Drive in a eSATA enclosure (Data drive)
Vista SP2 x64 with all the updates

Server
Opteron 165 (1.8 Ghz) CPU
2 x 1GB DDR 400 RAM
Asus A8R32-MVP Deluxe Motherboard
Marvell Yukon 88E8053 PCI-E Gigabit Ethernet Onboard Motherboard
320 GB WD Caviar Blue Hard Drive (20 GB partition for OS)
320 GB WD Caviar SE16 Hard Drive (Data drive)
WHS SP2 trial with all the updates
Server

My setup would most likely be considered a best case scenario. Everything was freshly installed and no other programs were installed.

00Roush

[mtber]

00Roush,

Here’s some IOZone testing results between Kelsey (Vista 64) eSata, Eric (Vista 32), and Dorene which has been newly configured with a 500 GB Seagate 7200.12 and is now running 64 bit Win 7 (quite an upgrade compared to the XP/5400 rpm IDE).

Wait, that can't be right - I couldn't have run "from" Kelsey eSata, I was just running from regular Kelsey, duh.

I need to re-run the Eric to Kelsey eSata as the 8 GB point looks spurious.

In re-reading your post I see that you asked for IPerf between machines, do these IOZone results achieve the same end?

The bold lines are “local” IOZone runs.

I gathered data on manual file transfers between Eric and Kelsey eSATA, but need to repeat them as there are some inconsistencies.

I’ve also attached a summary of my manual file copy tests to date.

Nope, not driving me nuts (the computers may be driving me nuts) I haven’t had time to keep up with all of the suggestions, but do appreciate the help.

All IOZone and manual copy tests to WHS were done by copies to the Public folder, in the case of IOZone by mapping that share to a drive letter. Maybe it would be worthwhile for me to map that public folder through remote desktop on the WHS and re-run the IOZone local test to that mapped drive.


John

[mtber]

I checked, and the settings below were already set on my WHS.

To manually disable RSS and TCP Offload yourself, follow these steps:

1. Click Start, click Run, type regedit, and then click OK.
2. Locate the following registry subkey:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\Tcpip\Parameters
3. Right-click EnableTCPChimney, and then click Modify.
4. In the Value data box, type 0, and then click OK.
5. Right-click EnableRSS, and then click Modify.
6. In the Value data box, type 0, and then click OK.
7. Right-click EnableTCPA, and then click Modify.
8. In the Value data box, type 0, and then click OK.
9. Exit Registry Editor, and then restart the computer.

[mtber]

00Roush,

I did some testing to follow-up on your WHS testing in regard to which disks are actually being written to.

Attached is a plot of local tests results run through Remote Desktop. I ran the same IOZone test on the C: and D: drives and then mapped the Public folder to a drive letter and ran the same test.

The Read data seems pretty clear, both the Public and D: data are really slow compared to the C: data. The Write data are less clear with D: being slow and Public being better, but not at straight C: performance.

I’m not sure what to make of the data other than to confirm your point that the WHS isn’t using my fast Seagate 7200.12 for client I/O operations. That is a real BUMMER!

This is very discouraging and suggests that it’s crippling to move “older” slower drives to the WHS storage pool to allow folder duplication. That’s part of the whole point of setting up the WHS, who wants to have to buy all new drives to expand storage when you have multiple otherwise decent 500 GB SATA drives laying around.

This data also would seem to change the equation for all of the other IOZone data that I have been generating for client to WHS. I’ll need to take some time to sift through it and see if it is consistent and makes sense.

A question though, since I’m typically working with much smaller files, < 10 MB, does any of this 2- 8 GB IOZone data really matter except when I just occasionally do big file copies? I understand that it’s needed diagnostically to determine if I have any configuration issues that are affecting overall performance. I’m just wondering if I’m chasing my tail at this point.

Thanks,

John

[00Roush]

Well I am not sure if I was clear in the previous post about the drives to test. With WHS testing D: or your public folder does not actually mean you would be testing your data drives specifically. D: is actually the second partition on your Seagate 7200.12 drive but it could also be a combination of all three drives. Same goes for your Public folder... it is a single folder that is linked to all three drives. This could also mean that if you have file duplication on it could have affected your Iozone results. I am not sure exactly though as the way WHS deals with the storage pool is unusual. This is why I recommended using the Disk Management part of the Computer Management control panel to find out where (what folder) WHS mapped your data drives. That way you can test them directly without interference from the drive extender software. But file duplication still might be in effect. With that said your results still might be correct. I mean testing the Public folder should give you actual read\write speeds to\from that folder. If that is true then you have definitely found a bottleneck.

Now you are hitting on some points that really irk me with WHS. Like the fact that you have no control over where your data gets written. Also the fact that by default WHS installs itself on the largest drive (fastest in your case) in the system. Then this drive is set to be used the least amount. This means if you want to use a old small drive as the OS drive you have to only have that drive hooked up when installing WHS. Also using this drive could bring down your overall performance as it might be considerably slower than some of the other drives in the storage pool. Basically what it comes down to is if you want consistent high performance out of WHS all of your drives should have similar read\write speeds. At least from what I can tell.

You do make a good point about the different file sizes. Testing the larger file sizes allows us to see performance without interference from files being cached. In reality this is what would happen if you tried to open a 10 MB file that had not recently been opened. It would have to be read directly from the disk and not memory so we want to see the actual speed data can be read from the disk. A single large file also helps with consistent resuslts. As you mentioned it is used as a diagnostic tool to determine if you have any configuration issues that are affecting overall performance.

Here is my thought process... I test for the best case scenario getting rid of all variables that might interfere with performance. This is helpful because I know from here performance goes down. So using Iperf I test the best case for network throughput. Then using Iozone I can test the best case local disk throughput. These results can give me a good idea of what I can expect for best case performance for network file transfers. The higher these results the better the chance that every other scenario will also have higher performance. For example... my best case scenario for large file transfers is 50 MB/sec. (Iperf of > 900 Mbps and Iozone of ~55 MB/sec read write) Now for a scenario of transfering 100 files at 10 MB a piece. Actual tranfser speed comes in at about 40 MB/sec or about 80% of the maximum. Now lets say the my best case scenario for a large file is 100 MB/sec. There is a chance that I still might see around 80% of maximum transfer speed so about 80 MB/sec. Realistically it would probably be more around 60-70 MB/sec though.

In your case it really comes down to what works for you. If photos seem to open quickly and you are able to move files around without much wait I'd say you doing good.

00Roush

[00Roush]

Something else I came across is if you don't like exactly how WHS works out of the box you can change it. I mean I accidentally installed WHS as a basic Windows 2003 install without all of the WHS extras. Which basically means it would work just like Windows Server 2003 Small Business Server.

Almost forgot in your current setup you could do a reinstall of WHS with just one of the 500 GB drives attached so that it would be your OS drive. That way you could have your fast Seagate drive as part of the storage pool that gets used. Just a thought... you know if you get bored. :D

00Roush

[benogil]

Along the lines of what OOrush is saying ( I think ), on the AVS forum under htpc, is a 149 page media server thread. It's just a thought, but folks in the last few pages have been playing with disabling the backup services for WHS, and just using it as a host OS for Flexraid and using the OS for streaming app's ( squeezecenter, ps3mediaserver, etc. )

Ben

[mtber]

no, not getting bored - yet


The client back-up is a key fetaure for me and one of the reasons I decided to go with the WHS, so in my case turning it off is not an attractive solution.

I ran the Iperf tests on the various clients (have dropped John from the mix for now). I've compiled the results from the IPerf runs (red bars), IOZone runs on the local machine (blue solid and strip) and IOZone machine to machine data (burgundy solid & strip) in the attached graph. The IOZone data is the average for 4G & 8GB transfers (where the results have reached “steady state”).

There are some IOZone machine to machine results that I need to generate to fill in the gaps.

A few things that I have noted in examining the data:


Huge difference in WHS local performance between going to the system disk (C)and the Public share or “D:” drive (as noted in previous posts) – reads are particularly poor


Eric (Vista 32) to Kelsey (Vista 64) – results are about as good as they will get given Kelsey (Vista 64) local results, performance limited by Kelsey (Vista 64) hard disk (Kelsey to Eric similar, but not quite as good)


Best results are between two of the faster disks Dorene (Win 7), Seagate Barracuda 7200.12 500GB, to Kelsey eSata, WDC WD5001AALS-00L3B (500GB).


The Kelsey (Vista 64) and Dorene (Win 7) to WHS IOZone results are similar somewhat correlate to the poor WHS D results (although the drive map was to Public).


Even though the WHS has a very fast system disk (Seagate Barracuda 7200.12 ST31000528AS 1TB), the two extended storage disks (Seagate Barracuda 7200.10 ST3500630AS 500GB) are dragging the whole system down.
Still some tweaks to try based on suggestions from this thread, but it seems that my WHS is hopeless, at least in regard to large file transfer performance. (because I’m not going out and buying a bunch of fast extra disks!)

John