Announcement

Collapse
No announcement yet.

Network Timeout Errors

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Network Timeout Errors

    We are testing a new OEM platform and seeing up to 190 Network timeout errors on the Network using the default test period at 100% duty cycle. When running the test at 50% duty cycle we are seeing less, but not zero errors.
    We are using a Nortel switch to windows 2003 R2 with a 3Ghz 8Cpu Xenon server as a target.
    We feel this is excessive. Does this sound like a driver issue? Or is this test excessive for this configuration?

    2 Software Versions:
    Burnin Test ver. 5.3 Pro (1012) Win 32
    Burnin Test ver. 6.0 Pro (1019) Win 32
    Operating System:WindowsEmbedded POS ready 2009, Ver2.0, Service pack3
    Processor: Intel Atom CPU N270 at 1.6Ghz
    Memory: 1.60Ghz, 1GbRam
    Chipset: Intel 945GSE
    Network: Realtek RTL 8168D/8111D Family PCi-E GBE NIC
    2 Network Drivers:
    1
    - Realtek Ver. 5.720.327.2009
    2- Realtek Ver. 5.754.310.2010

    Please Help
     

  • #2
    Network timeouts are often caused by overloading.

    So turn off all tests except the network test in BurnInTest.
    Try using a cross over cable instead of a switch if you suspect the switch.

    There is also the advanced network test option. See the User's Guide for details.

    Comment


    • #3
      Time out errors

      Network timeouts are often caused by overloading.

      So turn off all tests except the network test in BurnInTest.
      Yes we did, but we still receive excessive errors.

      Try using a cross over cable instead of a switch if you suspect the switch.
      We used a crossover cable and we also ran a loopback test with no errors.

      We noticed the errors mostly occur when running the Burnin test to a DHCP IP server.

      Any Ideas? We also tried windows XP pro on the same machine, same driver but we still receive excessive errors.

      Thanks, Mike

      Comment


      • #4
        So you are testing between two machines with a switch in the middle, and get errors.

        Then you remove the switch and replace the connection between the two machines with a direct (cross over) cable, and the errors disappear.

        Thus it would seem sensible to blame the swtich or initial cabling.

        and we also ran a loopback test with no errors


        If you are refering to the internal loop back test with the 127.0.0.1 IP address, then this isn't a good test of the hardware. It really just tests if the device driver is installed.

        Comment

        Working...
        X