andrewwong
Contributor
Contributor

NFS behaviour with vSwitch vs. DirectPath NIC

I have had this NFS issue on ESX 5-6.5, and I have never been successful in working out which part of ESX is causing the issue.

I think it's something to do with the vSwitch uplink.

My home has a ESX 6.5 server with Debian/Nas4Free VMs, and a QNAP NAS.

To test the issue I have 2 separate networks between Debian and QNAP box - one using vSwitch and one via Directpath NIC.

1. vSwitch with 1 uplink e1000 interface , 2 VMs connected

2. 1 Intel dual port NIC direct path to Debian VM on 192.168.9.1 directly connected to QNAP NAS (Debian VM also has a VMXNET NIC to 192.168.8.6)

3. 1 Nas4Free VM using the vSwitch with 192.168.8.251

4. 1 QNAP NAS at 192.168.8.250 and 192.168.9.2

From my Debian VM, if I try to mount NFS volume from Nas4Free, it works fine. If I try to mount NFS volume from QNAP it times out (also rpcinfo -p timesout, showmount -e takes forever)

From my Debian VM, if I try to mount the QNAP NFS volume using the 192.168.9.2 address ,it works perfectly.

So what can people suggest about the vSwitch / NIC Uplink that causes NFS to not work ?

Appreciate some ideas.

Thanks,


Andrew

0 Kudos
0 Replies