During NSX Host Preparation on Installation/Upgrade of NSX VIB, Legacy LACP mode portchannels over IXGBE does not rebuild leading to connectivity loss

Products

• VMware NSX

Issue/Introduction

Symptoms:

• When the ixgbe module is unloaded during NSX installation the LAG is destroyed and create.
• In the /var/log/vmkernel.log file of the ESXi host, you see entries similar to:
  
  2017-03-21T13:59:55.323Z cpu10:1205472)Mod: 5157: Unloading module <ixgbe> ...
  2017-03-21T13:59:58.749Z cpu2:1205472)Mod: 5335: Unload complete
  2017-03-21T13:59:55.907Z cpu10:1205472)Team.etherswitch: TeamESLACPLAGEventCB:6223: Received a LAG DESTROY event version :1, lagId :0, lagLinkStat
  us :NOT USED,lagName :, uplinkName :, portLinkStatus :NOT USED, portID :0x0
  2017-03-21T13:59:59.919Z cpu7:1205526)Loading module ixgbe ...
  2017-03-21T14:00:02.283Z cpu0:1205526)Team.etherswitch: TeamESLACPLAGEventCB:6223: Received a LAG CREATE event version :1, lagId :0, lagLinkStatus :NOT USED,lagName :, uplinkName :, portLinkStatus :NOT USED, portID :0x0
• If you do not see a LAG CREATE event, the port channel may be offline.
• If you force the physical links down, you may encounter the LAG CREATE at a later time frame, it should occur within seconds of the Loading module ixgbe line.
• In this case you may see only 1 of the vmnic's added to the portchannel, verify you see the following lines for all portchannel members.
  
  2017-03-21T14:00:02.283Z cpu0:1205526)lacp: LACPEnableComplete:4158: Init LAG: insert uplink vmnic3 to LAG(portset DvsPortset-0)
  2017-03-21T14:00:02.283Z cpu0:1205526)lacp: LACPEnableComplete:4158: Init LAG: insert uplink vmnic2 to LAG(portset DvsPortset-0)
  
  Note: The preceding log excerpts are only examples. Date, time, and environmental variables may vary depending on your environment.

Environment

VMware NSX for vSphere 6.4.x

Resolution

To resolve this issue, If not all uplinks are being re-added to the port channel during Init LAG, reboot the ESXi host.