11 Replies Latest reply on Feb 2, 2019 5:50 AM by TheBobkin

    Very High pNIC inbound packet loss rate counters... host issue or switch issue?

    DRAGONKZ Novice



      I've just rebuilt my all flash vSAN cluster on version 6.7 and swapped to 10Gb NICs and switches and I am seeing very high pNIC inbound packet loss rate counters on all my hosts (under host, monitor, performance, vsan - physical adapters).


      Packet loss counters range from 40% to 600%


      I've struggled to find anything during my searches to help point me in the direction of where to look.


      This is a 4 node home lab so unfortunately I don't have the luxury of logging a support ticket.


      Can anyone help point me in the direction of where to troubleshoot further? Hosts? Switches?...etc.





      In case the specs of my environment help I've got them below... all items are listed on the HCL.


      2 x Dell 8024F switches stacked with 4 ports

      2 x 10Gb DAC cables from NICs to switches (1 cable in each NIC goes to each switch)


      4 x HP DL20 Gen9, each of which has:

      HPE 546SFP+ Dual Port 10Gb NIC

      HPE H240 HBA

      HPE 400GB SAS 12G SSD cache disk (HPE branded HGST HUSMMxxxx)

      HPE 600GB SATA SSD capacity disk (HPE branded Intel S3500)



      vDS configuration/VMKernal Configuration wise for vSAN:


      Dedicated port group, with own VLAN

      MTU 9000 (Tried 1500 to start with... my environment supports both)

      Route based on physical NIC load

      Both NIC 1 and 2 are active uplinks