nagilum99
commented
3 years ago I've seen this several times at various systems. But the problem seems to be inside the guest tools and as you post Citrix - this is not the right place, as this is XCP-ng and not Citrix. You either need to switch to XCP-ng gues tools or make a bug report at Citrix.
Also: You should upgrade to XPC-ng 8.2 and check if there are newer tools (which I don't have in mind). IIRC I updated the tools and the problem didn't reappear on a customer machine.
kuepe-sl
I have a machine with xcp-ng as host, running a few VMs. One of them is a Windows 10 VM, which gives me trouble.
When this VM is running for about 2 weeks, xenguestagent.exe starts eating up all RAM until the VM becomes completely unresponsive and needs a forced reboot. It is always happening after approximately two weeks. (between 13 and 17 days of VM uptime)
I found out that at the point where it starts eating up all RAM, there are hundreds of warnings appearing in the Windows Event Log: Microsoft → Windows → WMI-Activity. I attached two event logs here: WMI-Activity-high-RAM-load_2021-02-25.zip When I took log "a", there was still about ~5 GB RAM left. When I took log "b" all RAM was already filled, so there may be different errors there.
The host machine is an AMD Ryzen Threadripper 3960X (24 cores) with xcp-ng 8.1.0-2 installed. The guest machine is running a Windows 10 Education N 1909 x64, with 8 cores and 20 GB of RAM. I have Citrix tools installed:
xenguestagent.exe reports v9.0.0.14
Do you have any advise on how I can fix the issue? Not having to reinstall Windows would be nice.