@olivierlambert The box was fully up to date. The problem fixed by itself, but it was noticed during 2 weeks on fully up to date box.

@olivierlambert Yes, all updates were installed. But. After about 1 hour of my post, situation with monitoring behaviour fixed by itself without any interventions.

I`ve looked this bug during 2 weeks or so.

I think, that my post was a reason of fix

Hi, people. I`m using xcpng and xo for more then 10 years. And after one last installation of 8.3 on new server found a strange behaviour of CPU monitoring. Monitoring shows a falling down of cpu usage down to zero for short time period every 4 minutes. But VM monitoring does not show this falling down. And i could not find any other indications of this problem.
CPU = Intel(R) Xeon(R) Gold 6338.

XCP-NG

One of VMs in this box.

@olivierlambert Now i can confirm absolutely exactly, that XCPNG upgrade or changing of mgmt interface is not a reason of this situation. I checked right now some hosts with bond interface and all these hosts lost possibility to add new networks(vlans) by XO

@olivierlambert

Pifs looks good

@olivierlambert

Hi. Founded a problem after last update of xoa (may be it is coincidence).

1. Change mgmt interface from bond to vlan (other network block)
2. Upgrade XCPNG (host rebooted). This step is not needed. Without upgrade i have the same situation.
3. Lost possibility to add any new vlans in xoa interface: list of pifs is empty, no errors
4. it is possible to add new vlan manually or by windows center
5. Before this procedure i could add vlans. Only hosts with bond are affected.

Is it a bug of xoa or of my brains?