physical vmnic down reporting in vCops

[globo]

vCops reports a physical NIC linkstate down and gives for the health 0 for 1 ESXi host.

This is very strange since these are just spare NIC which are not connected to a vSwitch and are not used by the ESXi host for any operation.

We have 2 other ESXi hosts with the exact same physical config. These servers are not reporting any vmnic down in vCops.

[kitcolbert]

Hi globo,

The general idea here is to alert you to a problem with the host hardware.  You're right that we should improve the logic to take into account whether the pnic is used as an uplink or not.  But in the meantime, you can manually cancel the alerts which will reset the health of your hosts.  Select an affected host, go to the Alerts tab, find the Fault alert for this nic, and click the "Cancel fault alert" button.  Within five minutes the health of the host will be restored.

[globo]

Hi Kit,

Thanks for you anwer.

canceling the alert solved health issue. but I found it still strange that only 1 of the 3 esxi host has triggered an alert for this since they are all configured the same way. Also as you mentioned a pNIC not connected to a virtual switch should not be taken in to account.

[TomDelamalle]

Hello,

We noticed the same type of behavior.

Are there any plans to improve this in a future release or do I need to create a ticket @vmware for this?

[gradinka]

Hi TomDelamalle,

can you share scenario and which version of vc ops are you using?

[TomDelamalle]

Hello,

We are using vcops 5.6 and the scenario is that at certain occasions we see the health go down of one physical host because vcops is reporting that the onboard nics (spares which are not used) are not connected.

By canceling the alert the health is reset but now and then the alerts come back. sometimes after a reboot, sometimes without any clear reason...

[gradinka]

Most of the fault alerts are based on events which are generated by VC/ESX themselves.

so each time the alert is renewed, there is an event which your host generated saying "hey, those ncs are down"

after vcops restarts those *do* come back, yes - you have to cancel the faults again.