Symptoms:
- The Uplink Port on an ESXi 5.5 host flaps when connected to a vSphere Distributed Switch (VDS).
- This issue does not occur when the Uplink Port is connected to a Standard Virtual Switch (VSS).
- This issue can occur when a NIC using a Broadcom bnx2 driver is attached to a NIOC-enabled vSphere Distributed Switch.
- The /var/log/vmkernel.log file contains entries similar to:
<6>bnx2 0000:03:00.1: vmnic3: CPU fw versions: TXP: 'txp6.2.1c' TPAT: 'tpa6.2.1c' RXP: 'rxp6.2.1c' COM: 'com6.2.1c' CP: 'cp6.2.1c'
YYYY-MM-DDT20:38:23.882Z cpu20:32852)<3>bnx2 0000:03:00.1: vmnic3: NIC SerDes Link is Down
YYYY-MM-DDT20:38:23.882Z cpu20:32852)IRQ: 540: 0x51 <cnic> exclusive, flags 0x10
YYYY-MM-DDT20:38:23.882Z cpu20:32852)VMK_VECTOR: 218: Registered handler for interrupt 0xff51, flags 0x10
or
YYYY-MM-DDT10:30:48.959Z cpu1:32852)<3>[cnic_stop_bnx2_ooo_hw:7307(vmnic3)]hw rx_cons=0 != sw rx_cons=0 rx_prod=511
<6>bnx2 0000:04:00.1: vmnic3: CPU fw versions: TXP: 'txp6.2.1b' TPAT: 'tpa6.2.1b' RXP: 'rxp6.2.1b' COM: 'com6.2.1b' CP: 'cp6.2.1b'
YYYY-MM-DDT10:30:49.348Z cpu1:32852)<3>bnx2 0000:04:00.1: vmnic3: NIC Remote Copper Link is Down
YYYY-MM-DDT10:30:49.349Z cpu1:32852)IRQ: 540: 0x4b <cnic> exclusive, flags 0x10
YYYY-MM-DDT10:30:49.349Z cpu1:32852)VMK_VECTOR: 218: Registered handler for interrupt 0xff4b, flags 0x10
YYYY-MM-DDT10:30:49.349Z cpu10:33412)netsched: NetSchedMClkNotify:2329: vmnic3: link down notification
Note: The preceding log excerpts are only examples. Date, time, and environmental variables may vary depending on your environment.