VM connection failing inside my UCS B environment

[Racso]

Hello,

So we are having a connectivity issue inside our virtual VMware environment

First a few things about our environment:

• Fabric Interconnects configured as EHM
• We have multiple VLANs configured on our Fabric Interconnects
• The ESXi blades have at least two vNICs for data path, and both vNICs have the same VLAN configured
• Both vNICs are configured on the same standard vSwitch and use the originating port connectivity policy
• All the vNICs have the priority Fabric Interconnect "A B"

So I did most of the classic troubleshooting, reviewing the VLANs and vNIC on the UCSM, checking the vSwitch and Port Groups on the ESXi, and we still get reports of disconnection that are solved after rebooting the VM

When checking the Fabric Interconnect MAC Address Table with "connect nxos a/b" and "show mac address table", we are not able to see the MACs of the VM in question

Is there any additional suggestion that I could check to tshoot this issue? Thanks for your help

[Kirk J]

You will need to do a sanity check on your VMnics at ESXi level and the vnics at UCSM level, and make sure the VMNIC #s are really the vnics that you think they are.  If you aren't learning macs, then there is usually a vlan misconfiguration somewhere.

Either the port group is tagging the wrong vlan,  isn't tagging at all,,, or the vnic allowed vlans aren't right (i.e. the vmnic vnic mismatch scenario)

Kirk...

[Racso]

Hi Kirk, thanks for your reply

What do you mean with sanity check?

I did check and mapped the right vNICs as the uplink of the vSwitches, and tagged the right VLANs on the vNICs

And I have no issue with the MAC of the ESXi vmNICs as they are the same of the vNICs defined on the Service Profile and are static on the Fabric, the issue we are having is with the VM MACs that are inside those ESXi

I can check the VLAN configuration once again, but I think its alright since the machines tend to connect fresh after a reboot but lose connection after a while, and its fixed with a reboot, think its more about an ARP issue

[Wes Austin]

What teaming option are you using in VMware?

https://www.cisco.com/c/en/us/support/docs/servers-unified-computing/ucs-b-series-blade-servers/200519-UCS-B-series-Teaming-Bonding-Options-wi.html 

Operating system	Supported	Not supported
VMWare ESXi	Route Based on Originating Port ID Route Based on Source MAC Hash	Route Based on IP Hash Route Based on Physical NIC Load

[Racso]

Hey Wes, we are using originating port ID,

I was thinking about deleting even one of the vNICs to discard the factor of load balancing, but it is configured properly