Page 1 of 2 12 LastLast
Results 1 to 25 of 30
  1. #1

    Hardware RAID test (3ware 8006-2LP)

    Hello,

    After seeing a topic a week or go discussing RAID cards I decided to give a hardware raid card a go to see if the performance will increase in one of our boxes.

    Just for the simplicity of the test, I have put them into a RAID0 formation for purely performance tests and no redundancy. I choose a 3ware RAID card and went for the 2 port 8006-2LP option rather than the 9600 (as they had the 8006-2lp and risers in stock and what I've always been told is that SATA1 and SATA2 is really a selling point rather than any performance increase but we will leave that argument there). Because we run mainly Windows systems, I have put on Windows Server 2003 x64 R2. What I am finding after installing it all is it seems pretty "slow".

    The rest of the hardware is a Dual, Quad Xeon (E5410x2), 8GB ram on a Tyan motherboard. Hard drives are 160GB Western Digital 7200 RPM so I can't see quite why it feels like its not running at a peak level.

    Does anyone have any applications or software to give this RAID array a test as I really don't want to order any more or roll them out on to the network to find that software raid would be a better improvement. I did try a burn in app which tests everything but it according to the 20 seconds I ran it, in average it only transferred at 2mbs.. That cant be right..

    I think one possibility is the RAID drivers arn't installed correctly as its still coming "Unknown Devices" in Device Manager and it seems It wont let me manually install the drivers for the 3ware device as it doesn't like the OS even though I have the correct ones and it installed Windows with it fine (a bit longer than normal anyway)

    Oh well thoughts please

    -Chris

  2. #2
    Join Date
    Mar 2006
    Location
    Reston, VA
    Posts
    3,131
    so heres a question is your raid done initializing? that can put a HUGE performance hit on write IO. Also make sure drives have their jumpers set to sata2?
    Yellow Fiber Networks
    http://www.yellowfiber.net : Managed Solutions - Colocation - Network Services IPv4/IPv6
    Ashburn/Denver/NYC/Dallas/Chicago Markets Served zak@yellowfiber.net

  3. #3
    Join Date
    Jun 2001
    Location
    Denver, CO
    Posts
    3,302
    3Ware 8006-2LP only supports SATA1 speeds. In any event, my feeling is that 9650-2LP is a far better card in terms of performance. As for testing your drives, I do have something in the office for that, but I can't remember the name of it off the top of my head and I'm going to be out this week. I will see if I can remember ...
    Jay Sudowski // Handy Networks LLC // Co-Founder & CTO
    AS30475 - Level(3), HE, Telia, XO and Cogent. Noction optimized network.
    Offering Dedicated Server and Colocation Hosting from our SSAE 16 SOC 2, Type 2 Certified Data Center.
    Current specials here. Check them out.

  4. #4
    If you can base a good argument for the 9650 then I would be happy to listen but as I am aware the difference between SATA1 and SATA2 is the drives have the potential of going to 3Gbs a second when the drive cant even max out a SATA1 port when you look at standard 7200 RPM SATA drives.

    -Chris

  5. #5
    Join Date
    Aug 2006
    Location
    Ashburn VA, San Diego CA
    Posts
    4,615
    Sounds like a driver issue. You definately shouldn't see any Question marks.

    I see a mininum of 50MB/sec reads/writes on Linux systems with the 8006-2lp. I've never tested throughput on a Windows system, but I don't notice any slowness at all.
    Fast Serv Networks, LLC | AS29889 | DDOS Protected | Managed Cloud, Streaming, Dedicated Servers, Colo by-the-U
    Since 2003 - Ashburn VA + San Diego CA Datacenters

  6. #6
    Join Date
    Jun 2002
    Location
    Waco, TX
    Posts
    5,623
    9650-2LP has 128MB RAM, speeds are good, I have an older multi port 8000 series and on RAID5 it does good to get 15MB/s but the 9500/9600's get far faster on RAID5 and RAID10.

  7. #7
    Join Date
    Aug 2000
    Location
    Sheffield, South Yorks
    Posts
    3,627
    8006-2LP is out of steam at about 48-54MB/s, even with WD Raptors on it. The 9650 PCI-e card, 2 port, will get nearer to 80MB/s with the same drives. The 8006-2LP is getting v. long in the tooth these days.
    Karl Austin :: KDAWS.com
    The Agency Hosting Specialist :: 0800 5429 764
    Partner with us and free-up more time for income generating tasks

  8. #8
    Dare I say it, would it be better to throw it over to software RAID for RAID0?

    -Chris

  9. #9
    Join Date
    Jan 2004
    Location
    Pennsylvania
    Posts
    942
    Quote Originally Posted by TheGreat1 View Post
    If you can base a good argument for the 9650 then I would be happy to listen but as I am aware the difference between SATA1 and SATA2 is the drives have the potential of going to 3Gbs a second when the drive cant even max out a SATA1 port when you look at standard 7200 RPM SATA drives.

    -Chris
    HUGE difference between SATA1 and SATA2 - NCQ.
    Matt Ayres - togglebox.com
    Linux and Windows Cloud Virtual Datacenters powered by Onapp / Xen
    Instant Setup, Instant Scalability, Full Lifecycle Hosting Solutions

    www.togglebox.com

  10. #10
    Have you got any proof/reviews to back that up, just because the port changed to 300mbs doesn't mean the drive can push even anywhere close to that..

    Last time I checked most drives only push 120mbs if that.. counting SATA1 not being near the limit. But anyway that's a different argument for a different topic.

    -Chris

  11. #11
    Join Date
    Jun 2001
    Location
    Denver, CO
    Posts
    3,302
    The NCQ makes a big difference. Additionally, rebuilds on 9650SE-2LP are significantly faster and far less service impacting.
    Jay Sudowski // Handy Networks LLC // Co-Founder & CTO
    AS30475 - Level(3), HE, Telia, XO and Cogent. Noction optimized network.
    Offering Dedicated Server and Colocation Hosting from our SSAE 16 SOC 2, Type 2 Certified Data Center.
    Current specials here. Check them out.

  12. #12
    Join Date
    Jan 2004
    Location
    Pennsylvania
    Posts
    942
    Quote Originally Posted by TheGreat1 View Post
    Have you got any proof/reviews to back that up, just because the port changed to 300mbs doesn't mean the drive can push even anywhere close to that..
    I suppose I wasn't clear enough. What I meant was SATA2 supports NCQ which gives a huge performance improvement, especially in servers where there are a lot of random reads/write.

    The 150/300 speed change itself is not that big of a deal.
    Matt Ayres - togglebox.com
    Linux and Windows Cloud Virtual Datacenters powered by Onapp / Xen
    Instant Setup, Instant Scalability, Full Lifecycle Hosting Solutions

    www.togglebox.com

  13. #13
    Join Date
    Jan 2003
    Location
    Chicago, IL
    Posts
    6,957
    Also, do you have write cache turned on? The 8006 cards still have that, don't they?
    Karl Zimmerman - Founder & CEO of Steadfast
    VMware Virtual Data Center Platform

    karl @ steadfast.net - Sales/Support: 312-602-2689
    Cloud Hosting, Managed Dedicated Servers, Chicago Colocation, and New Jersey Colocation

  14. #14
    Join Date
    Jul 2006
    Location
    Montreal, Canada
    Posts
    497
    I'm not sure what the OP needs however here are benchmarks for a RAID-10 setup if you'd like to compare. The controller is a 9550SX-4LP and the setup is 4x 500GB Seagate Barracuda HDDs.

    # hdparm -tT /dev/sda

    /dev/sda:
    Timing cached reads: 13472 MB in 2.00 seconds = 6744.51 MB/sec
    Timing buffered disk reads: 304 MB in 3.01 seconds = 100.97 MB/sec
    VEXXHOST, Inc.
    Secure & Reliable Hosting Solutions for Developers & Enterprises
    High Performance OpenStack® Public Cloud | Cloud Servers | Web Hosting

  15. #15
    Quote Originally Posted by TheWiseOne View Post
    I suppose I wasn't clear enough. What I meant was SATA2 supports NCQ which gives a huge performance improvement, especially in servers where there are a lot of random reads/write.

    The 150/300 speed change itself is not that big of a deal.
    Sorry, I didn't realise I sounded so defensive there

    To just to clarify, looking for a command to test the performance of the raid array on Windows to see how effective it is. Just doesn't seem quite fast enough for the spec it is. Windows install took a while which generally says to me something isn't right with the RAID array.

    I am looking at a driver issue TBH so might either downgrade to 32bit Windows Enterprise and see if that improves at all.

    Any thoughts and comments are greatly appreciated.

    Is there a huge difference with the 9650SE model over the 8006 model in a RAID0 environment? Or would software Raid (on the specification of the server) just be good enough to cope?

    -Chris

  16. #16
    Join Date
    Oct 2001
    Posts
    1,319
    Sounds to me like you might be using old drives - you're using a Dual Quad Xeon, 8G of ram but 7200 RPM 160GB drives?

    Perhaps you should do a speed test on the harddrives themselves without RAID and then compare it to the harddrives with raid.

    We use 8006-2LPs - no major complaints. Never tried 9650SE-2LP. 3ware just works.
    Avi B

  17. #17
    Hi Avi,

    Where using small drives (they are brand new though) as the 300GB it creates with the RAID is all that is required. We think from looking around on the 3ware site that there are no working drivers for Windows Server 64bit Enterprise R2 as its not supported on their list for the 8006-2LP.

    Ill try the 32bit version and see if there is any improvement. Do you know what software I can use to do a speed test tho?

    -Chris

  18. #18
    Join Date
    Oct 2001
    Posts
    1,319
    Hm... we use hdparm but that is for Linux. If you want the hardware performance you may be able to boot a Knoppix cd and use hdparm. If you want to get the performance from within Windows, I don't know.
    Avi B

  19. #19
    Join Date
    Aug 2005
    Location
    Maidenhead, UK
    Posts
    330
    Would highly recommend the 9650-2LP cards, definitely worth every penny.

  20. #20
    FWIW, the 8xxx series cards have had data corruption issues when a drive fails - haven't experienced this myself but have heard it from quite a few sources, YMMV.

  21. #21
    Join Date
    Jun 2002
    Location
    Waco, TX
    Posts
    5,623
    Quote Originally Posted by cacheflymatt View Post
    FWIW, the 8xxx series cards have had data corruption issues when a drive fails - haven't experienced this myself but have heard it from quite a few sources, YMMV.
    Raid0 would fail on any controller if a drive fails.

  22. #22
    Join Date
    Apr 2003
    Location
    San Jose, CA.
    Posts
    1,624
    This is what I usually use for judging drive/array performance in Windoze http://www.simplisoftware.com/Public...request=HdTach
    Daved @ Lightwave Networking, LLC.
    AS1426 https:/www.lightwave.net
    Primary Bandwidth: EGIHosting (NLayer, NTT, HE, Cogent)
    Xen PV VPS Hosting

  23. #23
    Exactly what I was looking for.. Thanks

    -Chris

  24. #24
    Join Date
    Mar 2008
    Location
    USA - Little Rock, AR
    Posts
    45
    Quote Originally Posted by ukgaz View Post
    Would highly recommend the 9650-2LP cards, definitely worth every penny.
    I would second this.. Definitely a great card!

    -Robert
    BitPath Networks, LLC - Professional VPS Hosting Services
    cPanel, DirectAdmin, and Plesk VPSs on the Virtuozzo Platform

  25. #25
    Hi there, coming late in the discussion, trying to choose between the 9650SE-2LP and the 8006-2LP for simple RAID 1.

    Also, what good current huge drives should I hook on them for reliability and reasonable performance?

    Thanks!

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •