Visit Open-E website
Results 1 to 8 of 8

Thread: Bad Performance with 3ware 9550SXU-4LP

  1. #1

    Default Bad Performance with 3ware 9550SXU-4LP

    Hi,

    after running the open-e DSS for over two years now, sind 2-3 weeks I got a performance issue. The problem is, I cannot describe since when this issue really exists.

    I made all the updates for DSS on a regularly basis. Actually, all my six clients freeze or dump during the access to the NAS server. Copying of 20GB takes longer than 3 hours. There are a lot of "breaks" during the copy process, anyway it continues copying afterwards. Again, I had no problems before.

    The server uses a Hardware RAID 5, with 3ware 9550SXU-4LP. The Write Cache is activeted. Rebuild/Migrate Rate and Verify Rate are set to "Faster I/O". StorSave is set to Perfofmance.

    Firmware of the 3ware card is Linux 2.6.25.13-oe32.
    Firmware of open-e DSS is 5.0.DB49000000.3278

    Does anyone have any idea what could have happen to my network/dss/3ware card?

    Any help appreciated.

    Thanks

  2. #2

    Default

    Dear zernik,

    In many occasions the NIC was the reason bhind a poor performance. You can check if there are any dropped packets from the test.log file that can be found in the GUI (Status->Hardware->Logs)

    Regards,
    SJ

  3. #3
    Join Date
    Jan 2008
    Posts
    86

    Default

    G'day Zernik,
    I agree with SJ, the NIC/Switch is a good place to look.
    Try starting a continuous ping to the box and/or another device on the same switch to see if you get dropped or high latency packets at the same time as the breaks in the 20GB copy.

    How are you accessing the data? NFS/iSCSI/CIFS?

    I don't use 3Ware, but I assume they have a logging facility? Does it show anything odd?

    Rgds Ben

  4. #4

    Default

    Hi Sh-J,

    indeed, in the tests.log there is an entry for that:

    Code:
    eth1    Link encap:Ethernet  HWaddr 00:30:48:8A:59:E1  
              inet addr:192.168.0.1  Bcast:192.168.0.255  Mask:255.255.255.0
              UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
              RX packets:116038240 errors:0 dropped:95 overruns:0 frame:0
              TX packets:160030293 errors:0 dropped:0 overruns:0 carrier:0
              collisions:0 txqueuelen:1000 
              RX bytes:3660420155 (3.4 GiB)  TX bytes:985991496 (940.3 MiB)
    95 droped packages. Is this high? What is the issue? How can I fix it?

    Thanks

  5. #5

    Default

    Here is the post again, bold was displayed in white color:

    eth1 Link encap:Ethernet HWaddr 00:30:48:8A:59:E1
    inet addr:192.168.0.1 Bcast:192.168.0.255 Mask:255.255.255.0
    UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
    RX packets:116038240 errors:0 dropped:95 overruns:0 frame:0
    TX packets:160030293 errors:0 dropped:0 overruns:0 carrier:0
    collisions:0 txqueuelen:1000
    RX bytes:3660420155 (3.4 GiB) TX bytes:985991496 (940.3 MiB)

  6. #6

    Default

    Hi Ben,

    appreciate your help. I access the NAS via SMB. NFS is off.

    All pings were successfull. Some of them are longer than 1ms:

    Code:
    Antwort von 192.168.0.1: Bytes=32 Zeit<1ms TTL=64
    Antwort von 192.168.0.1: Bytes=32 Zeit=3ms TTL=64
    Antwort von 192.168.0.1: Bytes=32 Zeit<1ms TTL=64
    What else could I do?

    Thanks
    Paul

  7. #7
    Join Date
    Oct 2007
    Location
    Toronto, Canada
    Posts
    108

    Default

    Quote Originally Posted by zernik
    What else could I do?
    Find the source of the packet slowdown.

    I would suggest the following order of attack (working out from the DSS Server):

    - Connect a client directly to the DSS server (you may need a cross-over network cable) using the port that the DSS is currently connected to the network, try all the ports that are available on the DSS

    - Change the switch ports that the DSS server is connected to

    - Review each of the devices between the clients and the DSS server, change the switch ports used for LAN connectivity.

    - Change the switch ports that the test client is connected to

    In between each step you want to test connectivity.

    Sean

    P.S. Change only one thing/connection at a time otherwise you will not be able pinpoint the source of the problem.

  8. #8

    Default

    Yepper - Just like old school "Change only one thing/connection at a time, otherwise you will not be able pinpoint the source of the problem." Thanks Sean!!

    You had it in Bold [B] - i took care of it.
    All the best,

    Todd Maxwell


    Follow the red "E"
    Facebook | Twitter | YouTube

Posting Permissions

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