NERSC logo National Energy Research Scientific Computing Center
  A DOE Office of Science User Facility
  at Lawrence Berkeley National Laboratory
 

Dell 5224 Switch


    Network issues

      This is a summary of the issues with the Dell 5224 switches.

      It appears that we were chasing two problems. The first is that there appears to be some kind of forwarding corruption, or a problem with learning. In some cases a host can only talk after being pinged from the switch itself (by logging into the switch and issuing a ping command from the switch cli). The other problem is that ports 13 through 24 were unable to send traffic that would exit the switch through a vlan tagged port that was a member of an aggregated link. We called dell support, and they agreed to send us beta code that fixed these problems.

      Let me know if you need further info.

      --eli

    Test Configuration

      SWITCH/PORTS SETUP:

        Dell 5224 (gusgs06, Ports 1-4) <=(4 ports trunked)=> Extreme 7i
        VLAN 310: Ports 5-12,21-24
        VLAN 311: Ports 13-16
        VLAN 312: Ports 17-20

        guscn25 (eth2) <--> Dell (Port 5)
        guscn26 (eth2) <--> Dell (Port 6)

      Software Version:

        #show version
        Unit1 Serial number :
        Service tag :CMS0221
        Hardware version :A02
        Number of ports :24
        Main power status :up
        Redundant power status :not present
        Agent(master)
        Unit id :1
        Loader version :1.0.0.0
        Boot rom version :1.0.0.0
        Operation code version :3.1.0.10

      TEST NODES: guscn25, guscn26

        Both nodes are configured with:
        • Supermicro P4DP6 motherboards with six PCI-X slots, two of which are 133 MHz capable
        • dual 2.2 GHz Pentium IV Prestonia Xeon CPUs
        • 2 GB of DDR PC2100 ECC memory
        • dual on-board Intel PRO/100 Ethernet interfaces
        • one Intel PRO/1000 XT 133 MHz PCI-X Gigabit Ethernet NIC

      TEST SOFTWARE:

        iperf version 1.7.0 (13 Mar 2003) pthreads (default settings)

    Test Results

      Same subnet (routed within a Dell switch):
        guscn25-ge0 -> guscn26-ge0T1: 633 Mb/s, 632 Mb/s
        T2: 989 Mb/s, 989 Mb/s
        T4: 990 Mb/s, 990 Mb/s
        guscn26-ge0 -> guscn25-ge0T1: 630 Mb/s, 630 Mb/s
        T2: 989 Mb/s, 989 Mb/s
        T4: 990 Mb/s, 990 Mb/s
      Between two subnets (Dell <-> E7i <-> Dell):
        guscn25-ge0 -> guscn26-ge0T1: 360 Mb/s, 362 Mb/s
        T2: 671 Mb/s, 672 Mb/s
        T4: 989 Mb/s, 989 Mb/s
        guscn26-ge0 -> guscn25-ge0T1: 361 Mb/s, 360 Mb/s
        T2: 689 Mb/s, 689 Mb/s
        T4: 989 Mb/s, 989 Mb/s
      Summary: iperf was able to achieve ~990 Mb/s with GigE but it required more than one iperf process/thread. With only one iperf process (T1), the performance was lower (633 Mb/s if the network traffic was within the same Dell switch and 360 Mb/s if the traffic needed to hop through Extreme 7i switch). With enough number of iperf processes, it will saturate the GigE pipe.