VMs innaccessible, event log shows "Lost uplink redundancy on virtual switch"

[agreenbhm]

We are running ESXi 4. A few times we've had various VM on this host innaccessible. The event log shows the following message:

Lost uplink redundancy on virtual switch "vSwitch0". Physical NIC

vmnic1 is down. Affected portgroups:"VM Network", "Management

Network", "VM Network", "VM Network", "VM Network", "VM

Network", "VM Network", "VM Network", "VM Network", "VM

Network", "VM Network", "VM Network", "VM Network".

warning

10/2/2009 8:15:38 AM

This entry occurs every few minutes until the problem "fixes" itself. We have two NICs built into the motherboard of the server for all of the host's networking. The NICs are teamed using default settings. They are both set as active adapters. Why is it that these NICs are not working with the failover function? My understanding was that without doing much configuration at all (besides adding both NICs to be active in the virtual switch) that we would have load balancing and failover capabilities. It seems that when this NIC is down, the VMs are innaccessible. However, we are still able to get into the management network (which is on vmnic0 only). Since vmnic0 is part of the same virtual switch, and that switch is accessible through vmnic0 (using vSphere Client), shouldn't the VMs be accessible, also? I think the answer is yes, but I'd like some input as to why this isn't working that way.

Thanks,

Drew Green

[zemotard]

Hi, I exactly have same trouble since V4 migration.

Do you have solved this problem ?

Regards

[agreenbhm]

Unfortunately I haven't found a solution. Fortunately it hasn't happened again since I last posted (knocks on wood).

[sqian]

As per your description, failover should work and network traffic should be able to failover to vmnic0.

Since the problem "fixes" itself, did you check if the lost uplink's back or failover worked?

When this problem happens again, VMs get inaccessible, you may login the host console, and type 'esxtop'

1. esxtop

then press 'n' to monitor which vmnic is using by the VMs actually. It will tell us if failover function takes effect at that point.