NASLite Network Attached Storage

I am posting my experience with NASlite+ for others to use as a reference and to hopefully help the developers with some hard throughput comparisions. Maybe others can run tests with the same parameters and post what kind of results they get and we can work to figure out what the best NICs and IDE interfaces are to acheive high-throughput numbers.

I will be installing OpenFilter 1.1p7 on the same server I installed Naslite+ on and see what kind of results I get from it.

-felipe


Using diskwriggler as a benchmarking tool from within Windows XP with the following parameters.

diskwriggler -NTSC -C -t -n 2000 -o Z:\
Where Z: is mapped to the share. Final filesize is about 1.3GB
See http://freshmeat.net/projects/diskwriggler/ for more info.

---------------------
Win XP Pro SP2 - Dell Poweredge SC1425 Dual Xeon 2.8Ghz, 1GB RAM with Intel Pro/1000 MT adapter, Adaptec SATA Mirrored 160GB 8MB Cache
to
NASlite+ 1.1 - Visionman (Gigabyte) P4 2.8GHz, 512MB RAM with Linksys EG1032 PCI Gigabit adapter (Marvell Chipset), 2 - Hitachi 80GB ATA133 2MB Cache HD

3.30 MB/s write (high fluctuation 2.1 to 4.9 MB/s)
20.40 MB/s read (started slow 7-15 in the first 15 seconds then when up to 22-24 MB/s)

Notes: Onboard Realtek disabled in BIOS. Write data rate fluctated wildly relative to its poor write speed. HD light on write would turn on in about 6 second intervals for less than a second indicating that it was flushing less data from the buffers to disk due to the low throughput on the network. The Marvell-based interface was properly detected at boot and was shown to be in 1000mbps mode by the following:

Mar 13 16:17:31 nas user.warn klogd: eth0: network connection up using port A
Mar 13 16:17:31 nas user.warn klogd: speed: 1000
Mar 13 16:17:31 nas user.warn klogd: autonegotiation: yes
Mar 13 16:17:31 nas user.warn klogd: duplex mode: full
Mar 13 16:17:31 nas user.warn klogd: flowctrl: symmetric
Mar 13 16:17:31 nas user.warn klogd: role: master
Mar 13 16:17:31 nas user.warn klogd: irq moderation: disabled
Mar 13 16:17:31 nas user.warn klogd: scatter-gather: enabled

The adapter was detected with the following entries in the log:
Mar 13 16:17:31 nas user.warn klogd: sk98lin: Network Device Driver v6.22
Mar 13 16:17:31 nas user.warn klogd: (C)Copyright 1999-2004 Marvell(R).
Mar 13 16:17:31 nas user.info klogd: PCI: Found IRQ 11 for device 01:00.0
Mar 13 16:17:31 nas user.warn klogd: eth0: SK-9521 10/100/1000Base-T Adapter
Mar 13 16:17:31 nas user.warn klogd: PrefPort:A RlmtMode:Check Link State
Mar 13 16:17:31 nas user.info klogd: forcedeth.c: Reverse Engineered nForce ethernet driver. Version 0.25.

---------------------
Same Config but with Realtek Chipset 100mb integrated in place of Linksys/Marvell GbE.

9.76 MB/s write (very little fluctuation)
4.80 MB/s read (somewhat choppier - between 3 and 6 MB/s)

Notes: Unlike others reporting fluctuating network throughput the Realtek received data from the XP system with the Intel Pro/1000 at a near flatline rate for the whole test. The drive activity light would light up for about 2 seconds every 5 seconds or so indicating that Naslite or the drive was buffering data to write. The read test from the Naslite back to XP was somewhat less smooth and much slower than the other tests. The drive light flickered with varying intesity the whole time.


---------------------
Win XP Pro SP2 - Dell Poweredge SC1425 Dual Xeon 2.8Ghz, 1GB RAM with Intel Pro/1000 MT adapter, Adaptec SATA Mirrored 160GB 8MB Cache
to
Win XP Pro SP2 - Dell Poweredge SC1425 Dual Xeon 2.8Ghz, 1GB RAM with Intel Pro/1000 MT adapter, Adaptec SATA Mirrored 160GB 8MB Cache

22.46 MB/s write (very little fluctuation)
23.03 MB/s read (little fluctuation 22.5 - 24.75 MB/s)

Notes: HD Light during write would flicker and go solid in about 3 second intervals representative of the faster flushing of the buffer and transfer rate. The HD light stayed on just about solid during the read phase.

----------
Further commentary:
I realize that the Hitachi drives only had 2MB Cache and its a slower computer but for the practical view of this test the numbers should not be that different. I would expect at most maybe a 10-20% loss for the drive's maybe not keeping up the sustained write speed.

I also found the 4GB limit as a filesize of 4,194,256 KB when writing using 7000 frames in diskwriggler. It just quits with an error when the limit is hit. This limit does not exist writing to the other WinXP system.

One thing to consider is that immediately after boot, NASLite+ initiates an extended SMART self-test for each storage drive. During that time, the HD’s performance is greatly affected and the performance results will be inaccurate. In order to determine the network throughput accurately, the drives must have completed the SMART self-test before proceeding.

Thanks for the hint. It would have been useful for me to read through previous posts regarding the SMART test. I will try again after letting the SMART test finish and only put one drive online so it completes faster.

The SMART issue does not explain the lopsided performance of the two different NICs though. What do you think?

Why the 4G limit via SMB/CIFS? I thought the CIFS VFS (not the older SMBfs) in Samba had worked past this limitation.

-felipe

----------------
Test series #2
----------------
This time I let the SMART test finish which took about 40 mins. on two 80GB EIDE HDs.

Please note that transfer rate snapshots are in 5 second increments.

Win XP Pro SP2 - Dell Poweredge SC1425 Dual Xeon 2.8Ghz, 1GB RAM with Intel Pro/1000 MT adapter, Adaptec SATA Mirrored 160GB 8MB Cache
to
NASlite+ 1.1 - Visionman (Gigabyte) P4 2.8GHz, 512MB RAM with Linksys EG1032 PCI Gigabit adapter (Marvell Chipset), 2 - Hitachi 80GB ATA133 2MB Cache HD

3.63 MB/s write (high fluctuation 1.8 to 5.2 MB/s)
16.95 MB/s read (15.5 to 18 MB/s)

----------------
Win XP Pro SP2 - Dell Poweredge SC1425 Dual Xeon 2.8Ghz, 1GB RAM with Intel Pro/1000 MT adapter, Adaptec SATA Mirrored 160GB 8MB Cache
to
NASlite+ 1.1 - Visionman (Gigabyte) P4 2.8GHz, 512MB RAM with Broadcom BCM5705KFB based Gigabit Ethernet , 2 - Hitachi 80GB ATA133 2MB Cache HD

31.90 MB/s Write (34-43 MB/s except one point where it dipped to 11 MB/s)
15.95 MB/s Read (14-18.5 MB/s except one point at 6 MB/s)

Second run:
30.27 MB/s Write (28-42 MB/s except one point where it dipped to 9.28 MB/s)
17.32 MB/s Read (Very tight except for one 11 MB/s point)


This is what NASlite+ reports for the Broadcom card:
Mar 16 16:35:35 nas user.info klogd: eth0: Tigon3 [partno(BCM95705A50-D) rev 3001 PHY(5705)]

(PCI:33MHz:32-bit) 10/100/1000BaseT Ethernet 00:.....
Mar 16 16:35:35 nas user.info klogd: eth0: HostTXDS[1] RXcsums[1] LinkChgREG[1] MIirq[1] ASF[0] Split[0] WireSpeed[1] TSOcap[0]


My conclusion: Marvell-based GigE adapters are very poor performers with NASlite+.
The Realtek-based 100mbps adapter performs at the limit of 100baseT for write (9.77 MB/s WOW!) and for read (4.8 MB/s)
The Broadcom 5705 works quite well for write performance (31 MB/s) and mediocre at read (17 MB/s)

I hope this information proves useful to others planning a NASlite+ system. I wish I had more time to test other adapters but will not have the opportunity.

I must say that I think NASlite+ is quite elegant and functional. I would like to see the following features in a near future version of NASlite+.
1. Ability to disable the extended SMART tests especially for doing benchmarking.
2. Built-in software RAID-5/Mirror/Raid-0 would be nice though likely to create a performance hit when used.
3. Support for more than one interface on different subnets/networks
4. A console-based and web-based I/O perfomance meter for the disks and network interfaces. The data collection must be switchable on/off.
5. Overcome the 4GB file size limit under SMB/CIFS.

Cheers,
-felipe
http://www.felipevidal.com

Thanks Felipe, that's very interesting...

From: ASUS P4P800-E DELUXE P 4/Northwood 3.2 Ghz ,Memory 2GB , Maxtor 6Y160P0 160GB Ultra ATA133 7200RPM 8MB Hard Drive
Onboard LAN Chipset Marvell 88E8001 LAN Speed 10/100/1000Mbps

To: Naslite File Server IBM 300PL 6565 PC
Operating System Linux 2.4.27.NASLite #6 Wed Oct 12 12:16:31 UTC 2005 i686
Processor (CPU) vendor_id : GenuineIntel model name : Pentium III (Coppermine)cpu MHz: 596.009 cache size: 256 KB
Memory 320MB RAM

Disk is a 300 GB Seagate Ultra ATA/100 7200 RPM 8MB Cache

Network Interface Card is a D-Link DGE-530T 10/100/1000Mbps Gigabit Adapter

The adapter was detected with the following entries in the log:

# Nov 2 16:35:53 max user.info kernel: Intel(R) PRO/1000 Network Driver - version 5.2.52-k3
# Nov 2 16:35:53 max user.info kernel: Copyright (c) 1999-2004 Intel Corporation.
# Nov 2 16:35:53 max user.info kernel: dgrs: SW=$Id: dgrs.c,v 1.13 2000/06/06 04:07:00 rick Exp $ FW=Build 550 11/16/96 03:45:15
# Nov 2 16:35:53 max user.warn kernel: FW Version=$Version$
# Nov 2 16:35:53 max user.info kernel: pcnet32.c:v1.30c 05.25.2004 tsbogend@alpha.franken.de
# Nov 2 16:35:53 max user.info kernel: ThunderLAN driver v1.15
# Nov 2 16:35:53 max user.info kernel: TLAN: 0 devices installed, PCI: 0 EISA: 0
# Nov 2 16:35:53 max user.info kernel: dmfe: Davicom DM9xxx net driver, version 1.36.4 (2002-01-17)
# Nov 2 16:35:53 max user.info kernel: ns83820.c: National Semiconductor DP83820 10/100/1000 driver.
# Nov 2 16:35:53 max user.warn kernel: sk98lin: Network Device Driver v6.22
# Nov 2 16:35:53 max user.warn kernel: (C)Copyright 1999-2004 Marvell(R).
# Nov 2 16:35:53 max user.info kernel: PCI: Found IRQ 9 for device 00:0e.0
# Nov 2 16:35:53 max user.info kernel: PCI: Sharing IRQ 9 with 00:12.0
# Nov 2 16:35:53 max user.warn kernel: eth0:
# Nov 2 16:35:53 max user.warn kernel: PrefPort:A RlmtMode:Check Link State
# Nov 2 16:35:53 max user.info kernel: forcedeth.c: Reverse Engineered nForce ethernet driver. Version 0.25.
# Nov 2 16:35:53 max user.info kernel: Uniform Multi-Platform E-IDE driver Revision: 7.00beta4-2.4
# Nov 2 16:35:53 max user.info kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
# Nov 2 16:35:53 max user.info kernel: VP_IDE: IDE controller at PCI slot 00:02.1
# Nov 2 16:35:53 max user.info kernel: VP_IDE: chipset revision 6
# Nov 2 16:35:53 max user.info kernel: VP_IDE: not 100% native mode: will probe irqs later
# Nov 2 16:35:53 max user.info kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
# Nov 2 16:35:53 max user.info kernel: VP_IDE: VIA vt82c596b (rev 12) IDE UDMA66 controller on pci00:02.1
# Nov 2 16:35:53 max user.info kernel: ide0: BM-DMA at 0xfff0-0xfff7, BIOS settings: hda:DMA, hdb:pio
# Nov 2 16:35:53 max user.info kernel: ide1: BM-DMA at 0xfff8-0xffff, BIOS settings: hdc:DMA, hdd:pio
# Nov 2 16:35:53 max user.warn kernel: hda: ST3300831A, ATA DISK drive
# Nov 2 16:35:53 max user.warn kernel: blk: queue c0349280, I/O limit 4095Mb (mask 0xffffffff)
# Nov 2 16:35:53 max user.warn kernel: hdc: CRD-8480C, ATAPI CD/DVD-ROM drive
# Nov 2 16:35:53 max user.warn kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
# Nov 2 16:35:53 max user.warn kernel: ide1 at 0x170-0x177,0x376 on irq 15
# Nov 2 16:35:53 max user.warn kernel: hda: attached ide-disk driver.
# Nov 2 16:35:53 max user.warn kernel: hda: host protected area => 1
# Nov 2 16:35:53 max user.info kernel: hda: 586072368 sectors (300069 MB) w/8192KiB Cache, CHS=36481/255/63, UDMA(66)

Start of TEST:

Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:\Documents and Settings\Carlos>diskwriggler -NTSC -C -t -n 2000 -o S:\tmp
Frame Resolution : 720 x 486
Bytes / Pixel : 2.00 (packed) (8 bit, 4:2:2 sampling)
Frame Size : 699,840 bytes
File Type : All frames in one container file (1,399,680,000 bytes).

Writing 2000 x 699840 byte frames into file S:\tmp/framestream.dat


Write Summary : 2000.00 frames in 126.44 secs
: 14.36(min) 15.82(avg) 16.15(max) FPS
: 9.58(min) 10.56(avg) 10.78(max) MB/s

Reading 2000 x 699840 byte frames from file S:\tmp/framestream.dat

Read Summary : 2000.00 frames in 174.77 secs
: 6.01(min) 11.44(avg) 12.41(max) FPS
: 4.01(min) 7.64(avg) 8.28(max) MB/s


BTW My hub is a LINKSYS 8-Port 10/100/1000 Gigabit Switch Model SD2008

Should I have expected better performance!?



These are from THE SERVER SYSTEM LOG:
# Nov 2 16:35:53 max user.info kernel: eth0: -- ERROR --
# Nov 2 16:35:53 max user.warn kernel: Class: internal Software error
# Nov 2 16:35:53 max user.warn kernel: Nr: 0x19e
# Nov 2 16:35:53 max user.warn kernel: Msg: Vpd: Cannot read VPD keys
# Nov 2 16:35:53 max user.warn kernel: eth0: network connection up using port A
# Nov 2 16:35:53 max user.warn kernel: speed: 1000
# Nov 2 16:35:53 max user.warn kernel: autonegotiation: yes
# Nov 2 16:35:53 max user.warn kernel: duplex mode: full
# Nov 2 16:35:53 max user.warn kernel: flowctrl: symmetric
# Nov 2 16:35:53 max user.warn kernel: role: slave
# Nov 2 16:35:53 max user.warn kernel: irq moderation: disabled
# Nov 2 16:35:53 max user.warn kernel: scatter-gather: enabled
# Nov 2 16:35:53 max user.info kernel: eth0: -- ERROR --
# Nov 2 16:35:53 max user.warn kernel: Class: internal Software error
# Nov 2 16:35:53 max user.warn kernel: Nr: 0x19e
# Nov 2 16:35:53 max user.warn kernel: Msg: Vpd: Cannot read VPD keys
# Nov 2 16:35:53 max user.info kernel: eth0: -- ERROR --
# Nov 2 16:35:53 max user.warn kernel: Class: internal Software error
# Nov 2 16:35:53 max user.warn kernel: Nr: 0x19e
# Nov 2 16:35:53 max user.warn kernel: Msg: Vpd: Cannot read VPD keys
# Nov 2 16:40:01 max user.info kernel: eth0: -- ERROR --

These messages continue to the end of the file!!!

Is my adapter not being recognised???

Greetings all,

Very knew to it all and my testing is somewhat more basic at this point. But I have the following system which I bought on ebay yesterday for $29.00:

Dell Optiplex P3 - 500mhz
Onboard LAN (dont know at what speed it runs yet)
128 meg ram

I installed three old hardrives I had laying around from windows machines (which incidental I had taken from those windows machines cause they wouldnt boot anymore...)

1 x 20 gb
2 x 40 gb

all hdd are of different manufacture.

I had the system up and running in about 30min after some initial attempts that didnt recognise any drives installed. ( I added one drive at a time)

No to my rudimentary testing...

From my windows based PC I downloaded 3 x dvd ISO files that were approx 4.5 gb each to each harddrive at the same time. So allowing a few seconds between each drag and drop they all started at pretty much the same time.

so three downloads at once. One file to each hardrive.

I used this to monitor the speed:
http://www.analogx.com/contents/downloa ... rk/nsl.htm
has anyone had any experience with this. Im unsure how to interperet the results.


in real terms it took 35 minutes to download all three files to the individual hdd.

So thats around 13.5 gb in 35 minutes.

Is that OK?

I guess if I did the same thing but downloaded all three files to the one drive at the same time the time would blow out?

Any way thats what Ive done so far for what its worth but so far Im loving it. Cant wait to get some decent sized drives and move all my mp3 and video files over to them.

Hi felipevidal, do you know what 1GB network cards use this Broadcom 5705 chip ? Thanks,Wolfgang

Hi All,

Ok so I finally worked out how to use wriggler and here are my results. My onboard lan runs at 100. I would be interested to see other results for comparison. My PC details are posted somehwere above in a previous post. So what speeds are people getting?


C:\Downloads\diskwriggler>diskWriggler -NTSC -C -t -n 2000 -o Z:\
Frame Resolution : 720 x 486
Bytes / Pixel : 2.00 (packed) (8 bit, 4:2:2 sampling)
Frame Size : 699,840 bytes
File Type : All frames in one container file (1,399,680,000 bytes).

Writing 2000 x 699840 byte frames into file Z:\/framestream.dat
69 frames in 5.06 secs ==> 13.63 FPS (9.10 MB/s)
74 frames in 5.00 secs ==> 14.80 FPS (9.88 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
74 frames in 5.02 secs ==> 14.75 FPS (9.85 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
70 frames in 5.00 secs ==> 14.00 FPS (9.34 MB/s)
71 frames in 5.02 secs ==> 14.16 FPS (9.45 MB/s)
68 frames in 5.02 secs ==> 13.56 FPS (9.05 MB/s)
74 frames in 5.00 secs ==> 14.80 FPS (9.88 MB/s)
71 frames in 5.05 secs ==> 14.07 FPS (9.39 MB/s)
70 frames in 5.03 secs ==> 13.91 FPS (9.29 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
75 frames in 5.05 secs ==> 14.86 FPS (9.92 MB/s)
75 frames in 5.06 secs ==> 14.81 FPS (9.89 MB/s)
72 frames in 5.03 secs ==> 14.31 FPS (9.55 MB/s)
67 frames in 5.00 secs ==> 13.40 FPS (8.94 MB/s)
74 frames in 5.00 secs ==> 14.80 FPS (9.88 MB/s)
71 frames in 5.00 secs ==> 14.20 FPS (9.48 MB/s)
71 frames in 5.05 secs ==> 14.07 FPS (9.39 MB/s)
75 frames in 5.03 secs ==> 14.91 FPS (9.95 MB/s)
74 frames in 5.00 secs ==> 14.80 FPS (9.88 MB/s)
75 frames in 5.03 secs ==> 14.91 FPS (9.95 MB/s)
74 frames in 5.05 secs ==> 14.66 FPS (9.79 MB/s)
63 frames in 5.02 secs ==> 12.56 FPS (8.38 MB/s)
43 frames in 3.22 secs ==> 13.36 FPS (8.92 MB/s)

Write Summary : 2000.00 frames in 139.00 secs
: 12.56(min) 14.39(avg) 14.91(max) FPS
: 8.38(min) 9.60(avg) 9.95(max) MB/s

Reading 2000 x 699840 byte frames from file Z:\/framestream.dat
47 frames in 5.08 secs ==> 9.26 FPS (6.18 MB/s)
52 frames in 5.16 secs ==> 10.08 FPS (6.73 MB/s)
51 frames in 5.11 secs ==> 9.98 FPS (6.66 MB/s)
50 frames in 5.06 secs ==> 9.88 FPS (6.59 MB/s)
38 frames in 5.13 secs ==> 7.41 FPS (4.95 MB/s)
51 frames in 5.05 secs ==> 10.11 FPS (6.74 MB/s)
51 frames in 5.09 secs ==> 10.01 FPS (6.68 MB/s)
47 frames in 5.05 secs ==> 9.31 FPS (6.22 MB/s)
51 frames in 5.05 secs ==> 10.11 FPS (6.74 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
50 frames in 5.00 secs ==> 10.00 FPS (6.67 MB/s)
50 frames in 5.05 secs ==> 9.91 FPS (6.61 MB/s)
52 frames in 5.03 secs ==> 10.34 FPS (6.90 MB/s)
50 frames in 5.00 secs ==> 10.00 FPS (6.67 MB/s)
51 frames in 5.00 secs ==> 10.20 FPS (6.81 MB/s)
51 frames in 5.00 secs ==> 10.20 FPS (6.81 MB/s)
50 frames in 5.08 secs ==> 9.85 FPS (6.57 MB/s)
51 frames in 5.05 secs ==> 10.11 FPS (6.74 MB/s)
51 frames in 5.02 secs ==> 10.17 FPS (6.79 MB/s)
51 frames in 5.03 secs ==> 10.14 FPS (6.77 MB/s)
51 frames in 5.03 secs ==> 10.14 FPS (6.77 MB/s)
50 frames in 5.00 secs ==> 10.00 FPS (6.67 MB/s)
50 frames in 5.05 secs ==> 9.91 FPS (6.61 MB/s)
51 frames in 5.05 secs ==> 10.11 FPS (6.74 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
51 frames in 5.06 secs ==> 10.07 FPS (6.72 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
50 frames in 5.05 secs ==> 9.91 FPS (6.61 MB/s)
50 frames in 5.02 secs ==> 9.97 FPS (6.65 MB/s)
51 frames in 5.00 secs ==> 10.20 FPS (6.81 MB/s)
52 frames in 5.00 secs ==> 10.40 FPS (6.94 MB/s)
52 frames in 5.09 secs ==> 10.21 FPS (6.81 MB/s)
52 frames in 5.02 secs ==> 10.37 FPS (6.92 MB/s)
50 frames in 5.03 secs ==> 9.94 FPS (6.63 MB/s)
48 frames in 5.02 secs ==> 9.57 FPS (6.39 MB/s)
51 frames in 5.00 secs ==> 10.20 FPS (6.81 MB/s)
52 frames in 5.05 secs ==> 10.30 FPS (6.88 MB/s)
44 frames in 4.28 secs ==> 10.28 FPS (6.86 MB/s)

Read Summary : 2000.00 frames in 200.83 secs
: 7.41(min) 9.96(avg) 10.40(max) FPS
: 4.95(min) 6.65(avg) 6.94(max) MB/s

My NASlite fileserver:

Operating System

Linux 2.4.27.NASLite #6 Wed Oct 12 12:16:31 UTC 2005 i686

Uptime

21:35:02 up 21:30, load average: 0.00, 0.00, 0.00

Processor (CPU)

vendor_id : GenuineIntel
model name : Pentium III (Coppermine)
cpu MHz : 596.015
cache size : 256 KB
bogomips : 1189.47

Memory
total used free shared buffers
Mem: 321744 317872 3872 0 1908

Disks

Filesystem Size Used Available Use% Device
System 7.7M 3.7M 4.1M 47% RAM Disk
Storage Area 275.1G 207.3G 67.8G 75% NAS Disk-1

Network Interface

ns83820.c: 0x22c: 49001186, subsystem: 1186:4900
ns83820 v0.20: DP83820 v1.3: 00:50:ba:38:ed:1e io=0xfeaff000 irq=9 f=sg
link now 1000 mbps, full duplex and up.
Link encap:Ethernet HWaddr 00:50:BA:38:ED:1E
inet addr:192.168.0.108 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1474402 errors:0 dropped:0 overruns:0 frame:0
TX packets:1372528 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1500751459 (1.3 GiB) TX bytes:1482396952 (1.3 GiB)

The server recognised the Gigabit Adapter!!!!!!


After the SMART test stopped I ran DISKWRIGGLER again. Here are the results:

Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:\Documents and Settings\Carlos>diskwriggler -NTSC -C -t -n 2000 -o s:/tmp
Frame Resolution : 720 x 486
Bytes / Pixel : 2.00 (packed) (8 bit, 4:2:2 sampling)
Frame Size : 699,840 bytes
File Type : All frames in one container file (1,399,680,000 bytes).

Writing 2000 x 699840 byte frames into file s:/tmp/framestream.dat


Write Summary : 2000.00 frames in 122.42 secs
: 15.70(min) 16.34(avg) 16.55(max) FPS
: 10.48(min) 10.90(avg) 11.04(max) MB/s

Reading 2000 x 699840 byte frames from file s:/tmp/framestream.dat


Read Summary : 2000.00 frames in 315.19 secs
: 2.70(min) 6.35(avg) 14.77(max) FPS
: 1.80(min) 4.24(avg) 9.86(max) MB/s


What disappointment

Perhaps the speed problem is with the limitations of the MOBO?!

But none of the speed tests done here so far have been blisteringly fast so maybe its just the way it is.

I think for brevity here apart from listing what equipment is used it might be better to just post your read and write final results...what say you all??

I've been doing some large tranfers to my naslite+ server lately, and I am noticing that SMB tranfers to Naslite are slower than SMB transfers to a windows box. FTP transfers are much faster. On my 100Mb lan, typical FTP performance is 11MB/s, while typical SMB performance is about 8MB/s. SMB transfers to a windows host are right about 9-10MB/s.

Is samba generally slower than the native windows implementation? Or is the performance gap NASLite specific?

I upgraded to V1.5 today (not user if that is the issue), and decided to run diskwriggler with the following - diskwriggler -NTSC -C -t -n 2000 -o w:\ from windows XP. My write summary is

Write Summary : 2000.00 frames in 132.63 secs
: 8.75(min) 15.08(avg) 16.64(max) FPS
: 5.84(min) 10.06(avg) 11.11(max) MB/s

which is fine. My read has been running forever with the following

8 frames in 5.34 secs ==> 1.50 FPS (1.00 MB/s)
7 frames in 5.77 secs ==> 1.21 FPS (0.81 MB/s)
8 frames in 6.53 secs ==> 1.22 FPS (0.82 MB/s)
6 frames in 5.22 secs ==> 1.15 FPS (0.77 MB/s)
8 frames in 5.63 secs ==> 1.42 FPS (0.95 MB/s)
10 frames in 5.56 secs ==> 1.80 FPS (1.20 MB/s).

Any ideas? Even music is stuttering at times across the network. I'm running on an 866MHz Dell Dimensions 4100

I decided to post my results to show what a low power box is capable of. I use my NASLite+ server for serving up movies and general backup on my home LAN.

Software: NASLite+ v1.5

Hardware:

CPU = PentiumII 350 Deschutes @ 66-MHz bus speed, so the processor speed is in fact 233-MHz (3.5 x 66).

Motherboard = Intel AL440LX PentiumII motherboard

RAM = 1 Crucial 256MB SDRAM DIMM

NIC = 3COM 3C905 TX 10/100

Hard Drives = 1 Western Digital 2500JB; 250GB, 7200RPM, 8MB cache

Foppy = 1.44MB generic floppy drive

CD-ROM = 24x Mitsumi IDE

Video Card = S3 Virge PCI card, usefull when i connect a monitor to check BIOS settings, etc.

This setup was put together from spare parts that I had laying around on hand. I have 3 IDE ports open waiting for more hard drives. That is the reason for booting the CD from an add-on IDE controller.

I boot from CD-ROM connected to a Promise Ultra100 TX2 IDE controller. I had to use a custom CD to boot from this controller. I used "the ultimate boot cd" and added the boot.img to automatically boot from the NASLite+ ISO image.

Test
---------

command = diskwriggler -NTSC -C -t -n 2000 -o z:/tmp

results =

Frame Resolution : 720 x 486
Bytes / Pixel : 2.00 (packed) (8 bit, 4:2:2 sampling)
Frame Size : 699,840 bytes
File Type : All frames in one container file (1,399,680,000 bytes).

Writing 2000 x 699840 byte frames into file z:/tmp/framestream.dat
80 frames in 5.07 secs ==> 15.79 FPS (10.54 MB/s)
81 frames in 5.04 secs ==> 16.08 FPS (10.73 MB/s)
81 frames in 5.05 secs ==> 16.05 FPS (10.71 MB/s)
81 frames in 5.07 secs ==> 15.98 FPS (10.67 MB/s)
80 frames in 5.04 secs ==> 15.88 FPS (10.60 MB/s)
80 frames in 5.03 secs ==> 15.91 FPS (10.62 MB/s)
80 frames in 5.04 secs ==> 15.88 FPS (10.60 MB/s)
80 frames in 5.02 secs ==> 15.95 FPS (10.64 MB/s)
77 frames in 5.01 secs ==> 15.38 FPS (10.26 MB/s)
80 frames in 5.03 secs ==> 15.91 FPS (10.62 MB/s)
80 frames in 5.02 secs ==> 15.95 FPS (10.64 MB/s)
81 frames in 5.05 secs ==> 16.05 FPS (10.71 MB/s)
81 frames in 5.06 secs ==> 16.02 FPS (10.69 MB/s)
80 frames in 5.01 secs ==> 15.98 FPS (10.66 MB/s)
81 frames in 5.06 secs ==> 16.02 FPS (10.69 MB/s)
80 frames in 5.01 secs ==> 15.98 FPS (10.66 MB/s)
81 frames in 5.06 secs ==> 16.02 FPS (10.69 MB/s)
81 frames in 5.07 secs ==> 15.98 FPS (10.67 MB/s)
81 frames in 5.06 secs ==> 16.02 FPS (10.69 MB/s)
81 frames in 5.06 secs ==> 16.02 FPS (10.69 MB/s)
81 frames in 5.03 secs ==> 16.11 FPS (10.75 MB/s)
75 frames in 5.02 secs ==> 14.95 FPS (9.98 MB/s)
76 frames in 5.05 secs ==> 15.06 FPS (10.05 MB/s)
81 frames in 5.04 secs ==> 16.08 FPS (10.73 MB/s)
80 frames in 4.97 secs ==> 16.11 FPS (10.75 MB/s)

Write Summary : 2000.00 frames in 125.90 secs
: 14.95(min) 15.89(avg) 16.11(max) FPS
: 9.98(min) 10.60(avg) 10.75(max) MB/s

Reading 2000 x 699840 byte frames from file z:/tmp/framestream.dat
53 frames in 5.03 secs ==> 10.54 FPS (7.04 MB/s)
55 frames in 5.03 secs ==> 10.94 FPS (7.30 MB/s)
57 frames in 5.14 secs ==> 11.10 FPS (7.41 MB/s)
59 frames in 5.04 secs ==> 11.71 FPS (7.82 MB/s)
57 frames in 5.19 secs ==> 10.99 FPS (7.33 MB/s)
36 frames in 6.07 secs ==> 5.93 FPS (3.96 MB/s)
56 frames in 5.16 secs ==> 10.86 FPS (7.25 MB/s)
57 frames in 5.07 secs ==> 11.25 FPS (7.51 MB/s)
57 frames in 5.01 secs ==> 11.38 FPS (7.60 MB/s)
57 frames in 5.02 secs ==> 11.36 FPS (7.58 MB/s)
57 frames in 5.06 secs ==> 11.27 FPS (7.52 MB/s)
54 frames in 5.02 secs ==> 10.76 FPS (7.18 MB/s)
57 frames in 5.01 secs ==> 11.38 FPS (7.60 MB/s)
55 frames in 5.16 secs ==> 10.66 FPS (7.12 MB/s)
58 frames in 5.19 secs ==> 11.18 FPS (7.46 MB/s)
58 frames in 5.01 secs ==> 11.58 FPS (7.73 MB/s)
56 frames in 5.02 secs ==> 11.16 FPS (7.45 MB/s)
57 frames in 5.08 secs ==> 11.23 FPS (7.49 MB/s)
56 frames in 5.07 secs ==> 11.05 FPS (7.38 MB/s)
58 frames in 5.21 secs ==> 11.14 FPS (7.43 MB/s)
58 frames in 5.03 secs ==> 11.54 FPS (7.70 MB/s)
57 frames in 5.07 secs ==> 11.25 FPS (7.51 MB/s)
56 frames in 5.07 secs ==> 11.05 FPS (7.38 MB/s)
58 frames in 5.13 secs ==> 11.31 FPS (7.55 MB/s)
58 frames in 5.05 secs ==> 11.49 FPS (7.67 MB/s)
56 frames in 5.02 secs ==> 11.16 FPS (7.45 MB/s)
56 frames in 5.09 secs ==> 11.01 FPS (7.35 MB/s)
58 frames in 5.01 secs ==> 11.58 FPS (7.73 MB/s)
57 frames in 5.05 secs ==> 11.29 FPS (7.54 MB/s)
57 frames in 5.12 secs ==> 11.14 FPS (7.43 MB/s)
54 frames in 5.18 secs ==> 10.43 FPS (6.96 MB/s)
52 frames in 5.01 secs ==> 10.39 FPS (6.93 MB/s)
58 frames in 5.08 secs ==> 11.42 FPS (7.62 MB/s)
57 frames in 5.18 secs ==> 11.01 FPS (7.35 MB/s)
59 frames in 5.09 secs ==> 11.60 FPS (7.74 MB/s)
39 frames in 3.47 secs ==> 11.22 FPS (7.49 MB/s)

Read Summary : 2000.00 frames in 182.15 secs
: 5.93(min) 10.98(avg) 11.71(max) FPS
: 3.96(min) 7.33(avg) 7.82(max) MB/s



Hope this helps for users of older systems.

Here are results from an even lower powered machine

Micronics motherboard with 200MHz processor
32MB ram
GA311 Gigabit ethernet card
Hitachi 250GB 7200rpm 8MB buffer on secondary IDE as slave.

100Mbit network

Write Summary : 2000.0 frames in 193.34 sec
10.34 FPS avg
6.90 MB/s avg

Read Summary:
8.15 FPS avg 7.46 FPS min
5.44 MB/s avg 4.98 MB/s min

It looks like the lower power processor is having trouble keeping up or the secondary IDE is slowing things down.