High CPU load on windows machines on XCP-ng

Hi, I have this setup:

I have the windows update button enabled in xen-orcestra and all updates are installed but it still says (HVM).

I have tried to install the drives manually from this page: This page and This page.
My windows 2016 server takes up to 20 minutes to boot and the CPU load is at 100%. Even with all 8 cores assigned to it, it still takes about the same time to boot.

When the cpu load finally slows down it still jumps up quite a bit only by opening a folder.

From device manager:

I have no problem with my linux machines.

Could this only be related to the fact my server is using old HW? (Dell ML350 G5 - 2 x Xeon E5410)

was this a clean load? I have seen systems that were migrated from another hypervisor or from a physical machine have issue due to them still having the old hardware drivers installed but not showing in the device manager.

I did a clean Install from a windows iso.
Also tried a clean install of windows 10 with the same results. When just browsing directories it uses about 45% CPU with 4 cores.
Is there a fix, maby remove all drivers and install them again?

That is not a very fast processor and if you don’t have very fast hard drives for storage, that may be what is wrong.

Probably. :stuck_out_tongue:
They are stock 6 times 146gb 10k SAS hdd. 4 are 15k. In raid 5. And 2 in raid 1.
Then maby i just need to stick with linux on that server :slight_smile:

Those CPUs that you are running in that server were released in Quarter 4, 2007. They aren’t going to be able handle Windows 10 very well, even if it was a dedicated install. I’ve been shopping around for a server to run XCP-ng on and I came across an HP ProLiant DL360e on eBay for $300 with 32GB of RAM, 2x500GB SATA drives and dual Xeon E5-2420 6 core/12 thread CPUs, granted they are only 1.9Ghz with a boost clock of 2.4Ghz, but they were released Q2 of 2012.