Ontap select : Half duplex and NFS issues

[martin_thibert]

We are trying to run a VM on an NFS datastore presented by Ontap Select 2 nodes cluster.  Everything is working except that the VM has access to the disks every minutes for 30 seconds.  The ethernet ports on the ontap select appliance are shown has half duplex even if it's configured for full duplex.  ESX host VMnic are Fullduplex. 

Anyone has idea on how to fix this issues?

Thk's!

Ontap select 9.5RC1

VMware ESXi, 6.7.0, 10302608

standard vswitch 

  Node: ONTAP-02
                                        Port: e0b
                                        Link: up
                                         MTU: 9000
             Auto-Negotiation Administrative: false
                Auto-Negotiation Operational: false
                  Duplex Mode Administrative: full
                     Duplex Mode Operational: half
                        Speed Administrative: 10000
                           Speed Operational: -
                 Flow Control Administrative: none
                    Flow Control Operational: none
                                 MAC Address: 00:a0:b8:f6:e4:fc
                                   Port Type: physical
                 Interface Group Parent Node: -
                 Interface Group Parent Port: -
                       Distribution Function: -
                               Create Policy: -
                            Parent VLAN Node: -
                            Parent VLAN Port: -
                                    VLAN Tag: -
                            Remote Device ID: -
                                IPspace Name: Default
                            Broadcast Domain: Default
                          MTU Administrative: 9000
                          Port Health Status: healthy
                   Ignore Port Health Status: false
                Port Health Degraded Reasons: -

[GregoryH]

Sounds like this may be cosmetic and there are BUGs relating to this, but they are supposed to have been fixed. (See below) . Have you tested the throughput to see if it is actually running at half? Which type of VMnics are you using? E1000 or VMXNET3. 

https://mysupport.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=1110020

[martin_thibert]

The half duplex is effectively cosmetic, but we still have issue with the NFS.  We are using vmxnet3.  Storage access to the NFS datastore is on for 25-30sec. and off for 25-30sec. 

I have the issue while doing storage vmotion on and off the NFS datastore and also when generating IO from the guest OS VM.

I don't have the issue when generation IO from the guest Os on the local datastore.