After change ahv network port got NIC down alert

Hello, if i understood well. You just must remove the eth2 interface from the bonding. either from the prism element web interface, or from the cli on a CVM with the manage_ovs command. If you want more details/help, don't hesitate.

Regards,

You will have to remove it from PE web interface as manage_ovs commands will not work with the virtual switch created.

The alert should go away after 24 hours. NCC will alert on a NIC port going down for 24 hours, after this it will assume that its expected to be down and it should stop alerting.

https://portal.nutanix.com/kb/2566

This generally happens when you bring down an interface, then the NIC link down alert is displayed and cluster health turns red.

After the NIC is in a downstate, the NIC status will stay in the file check_cvm_health_job_state.json for 24 hours.

From NCC 4.3.0 onwards, nic_link_down_check will detect the NIC status changed to a downstate and then raise the alert. After 24 hours, if the NIC is still in the downstate and never goes in an upstate within the day, the NIC will be removed from file check_cvm_health_job_state.json and no more alert would be generated nor failure detected by nic_link_down_check.

That is correct Dylan, its just a lot of extra steps.

Manage_ovs commands to update things will not work if vswitch is created, and is a legacy way of doing things.

• You can disable vswitch
• Put the host in MM
• Update it with manage_ovs
• Take host out of MM
• Then recreate the vswitch.

That is a lot more steps than just updating that one server in the GUI and letting automated processes handle everything.

if a CLI method is preferred you can use ACLI commands such as acli net.update_virtual_switch vs# host_uplink_config=
 

Going into acli context mode and using tab will help to learn how to use the commands and options. 

Hi @DylanH & @Jon Jordan ,

Thank you both.

I had the same issue. But the alert will be gone after 24 hours. Just wait.