@dinhngtu Thanks for your help. I solved to install XS Tools uninstalling old XS version 9.0, running XenClean and installing version 9.4. Also Disk stats now are correct. Now I get these good performances: [image: 1747727185969-a6beabc3-793e-41d7-9f03-4689db372275-immagine.png] [image: 1747727202526-20d58e2a-e9f9-4d7b-ba44-844dca2008fa-immagine.png]

@Forza said in Epyc VM to VM networking slow: olivierlambert said in Epyc VM to VM networking slow: If we become partners officially, we'll be able to have more advanced accesses with their teams. I still have hope, it's just that the pace isn't on me. Hi, is there anything new to report on this? We have very powerful machines, but unfortunately limited by this stubborn issue. Can you test https://xcp-ng.org/forum/topic/10862/early-testable-pvh-support ? We observe very significant improvements on AMD EPYC with PVH. We're still pin-pointing the issue with HVM, the current hypothesis is a issue regarding memory typing (grant-table accessed as uncacheable(UC) which is very slow) related to grant-table positionning in HVM.

Can I get a link to the download?

@AshleyDe In the last few days, I've had these same issues with a few Server2025 virtual machines (I assume are close ish to Win11 24h2).

Great news Closing this as fixed. Thanks everyone for your feedback!

@plaidypus You can configure the disk within the guest OS to match whatever formatting you want. I would suspect the performance gains would be included as well, but you'd have to test and verify.

@sborrill And for completeness, try pvshim (a.k.a. pv-in-pvh) using xl: Download https://ftp.netbsd.org/pub/NetBSD/NetBSD-10.1/amd64/binary/kernel/netbsd-INSTALL_XEN3_DOMU.gz Create /root/nbpv containing: name = "nbpv" type="pvh" pvshim=1 pvshim_cmdline="pv-shim console=xen pv-linear-pt=true" kernel = "/root/netbsd-INSTALL_XEN3_DOMU.gz" memory = 1024 Boot VM successfully with xl create -c nbpv So pv-in-pvh does work (does require pv-linear-pt=true on shim command line though).

This only affects UEFI VMs running Ubuntu. A default Ubuntu installation includes the linux-modules-extra-<kernel-version>-generic package, which provides the bochs driver—required for proper VGA display in the Xen Orchestra console. If this package is removed, the console display becomes scrambled. Run the following command to install the missing package (replace <kernel-version> with the version in use by the OS): sudo apt install -y linux-modules-extra-<kernel-version>-generic source: https://docs.xcp-ng.org/vms/#distorted-display-console-on-ubuntu-uefi-vms

@tjkreidl Interesting that ASRock is also using the NUC name. ASUS also has the new NUC15. This new ASRock box uses the Ultra 7 255H which has 6P/8E/2LPE cores (no HT).

I'm having similar issue with A400 on xcp-ng8.3 Proprietary driver fails with following message when running nvidia-smi : NVRM: GPU 0000:00:05.0: RmInitAdapter failed! (0x24:0x72:1568) [ 44.619030] NVRM: GPU 0000:00:05.0: rm_init_adapter failed, device minor number 0 [ 45.095040] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint. [ 45.144703] nvidia-uvm: Loaded the UVM driver, major device number 241. system is actually loading the driver : [ 6.026970] xen: --> pirq=88 -> irq=36 (gsi=36) [ 6.027485] nvidia 0000:00:05.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem [ 6.029010] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 550.144.03 Mon Dec 30 17:44:08 UTC 2024 [ 6.063945] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 550.144.03 Mon Dec 30 17:10:10 UTC 2024 [ 6.118261] [drm] [nvidia-drm] [GPU ID 0x00000005] Loading driver [ 6.118265] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:00:05.0 on minor 1 xl pci-assignable-list gives : 0000:43:00.0 0000:43:00.1 and gpu is assigned as passthrough,, but when listing test VM i have empty list of devices.. [23:06 epycrep ~]# xl pci-list Avideo-nvidia [23:35 epycrep ~]# Not sure if i want to try more before switching gpu to something else. Any hint where to look for ? Server is gigabyte g292-z20 wih epyc 7402p and single gpu for testing. IOMMU enabled.

@msupport Right! I'm sure it's a great product, the goal for this is to have no licenses, unlimited users, ported for several hypervisors. It's just a pet project I've wanted to do since I started in IT, and if other people besides me can use it, great! I know one of my employers may work on making it a polished project if the Omnissa renewal is too expensive. And if I can build it, why pay for it? I might not be a programmer by trade, but there's no reason I can't learn and ask around for help!

@ATSandrearicci Check the compatibility [image: 1744296853670-0877c1cb-2ac5-4afb-a993-a96fd66c7474-image.png] You must configure the license for the client on the license server: [image: 1744296959393-01b3b16d-8f06-4fe9-94ce-87b5c1cc1fcf-image.png] On the windows 11 client in the Nvidia control panel you have to specify the license server: [image: 1744297049623-8d883042-fc27-4678-8a05-e3c96b7a38a4-image.png] Check the authorization on the file: [image: 1744298130266-796ee95b-956e-4e9e-b178-cf60e92a0c97-image.png]

@CodeMercenary I had a simillar issue to you, check the time and the date on the host. If this might be any help to you: https://xcp-ng.org/forum/topic/10687/after-an-update-the-nic-s-has-disappeared-but-still-works-somehow/10?_=1743764792775

Ok, found a solution. Seems there is nothing I could do via cli or xo-ce to see the virtual GPUs, had to install xcp-ng center, specifically from: https://github.com/xcp-ng/xenadmin/files/13800584/XCP-NG-Center-2023-Release.zip Using the center, I was able to assign the GPU to the VM and everything worked well from then on. Now to figure out licensing....

@gduperrey I thought I had. I was replying to that thread and the forum software prompted me to click to reload. It continued to show the contents of the thread above as I clicked to post my reply, so if it opened a new thread, I count that as a bug in the forum, not user error

@Denson After enabling HA, the host has to be manually rebooted, which I think you're already aware of. OK, well, that leaves pretty much a network issue. Hmm. The statefile created OK on the shared SR within the pool? The "not found" error message doesn't give you much to go on, unfortunately.

@Andrew, @bvitnik and @DustinB In my tests, I did the following. I did this process twice and it worked. To simulate a hardware failure on the master node, I simply turned it off. If the master pool is down or unresponsive due to a hardware failure, follow these steps to restore operations: Use an SSH client to log in to a slave host in the pool. Run the following command on the slave host to promote it as the new pool master: xe pool-emergency-transition-to-master Confirm the change of the pool master and verify the hosts present in it: xe pool-list xe host-list Even if it is down, the old master node will appear in the listing. Remap the pool in XCP-ng or XO using the IP of the new master node. After resolving the hardware issues on the old master node, start it up. When it finishes booting, it will be recognized as a slave node. In testing, I did not need to run any other commands. However, if the node is not recognized, try typing on it after accessing it via SSH: xe pool-recover-slaves I didn't understand why it worked. It seemed like "magic"!

@bvitnik Thank you -- very useful information!