@acebmxer said: connectNbdClientIfPossible yes connectNbdClientIfPossible is the nearest error message from the real cause Now, why does it fails ? nbd should be enabled on the network used for backups , do you have a default backup network defined ? xo must have at least one VIF on this network

@MajorP93 said: @florent Thanks! I did not have time (yet) to look into the heap export due to weekend being in between. I will update to current master and provide you with the heap export in the following days. Best regards no problem , the sooner we'll have these exports, the sooner we will be able to indentify what is leaking in your usage, since we already plugged the leaks we reproduced in our labs

@yannsionneau said: Can you retry with an up-to-date xcp-ng 8.3 please? FYI on recent XCP-ng 8.3 versions the pci-passthrough will enable the ROM expansion bar. The guest VM will have access to it, so no need to pass it via qemu anymore. See my comment on GitHub: https://github.com/xcp-ng/xcp/issues/786#issuecomment-4281846490 Regards, Yann I tried, yes. I confirmed as well on github that the patch does what it suppose to do - it exposes ROM BAR, but that alone is not sufficient to get our cards working. More in my other topic.. I believe it will still fix a lot of issues, and it could potentially fix amd gpu passthrough for some card, but not Phoenix, Raphael and this generation of Ryzen iGPUs (not sure specifically about Barcelo).. But it is a good progress anyway as I wouldn't be able to fix this ever on my own, so I am really glad it is done. Now I will probably get back to the topic and try to patch whatever else is needit and give it to people in a form of rpm package for the time being.. we will see, it is all about time

Yeah, as far as I understand, this looks like something RPU wasn't really designed to handle, and I'm not deep enough in the internals to say for sure how tricky it would be to add. My instinct is this sits more in feature request territory than a bug, so I'd suggest filing it on https://feedback.vates.tech with something like "RPU: graceful shutdown/restart for PCI-pinned VMs" and your actual topology in the description. The product team will understand the constraint much better with a concrete Ceph setup to look at. Someone here probably knows if there's a cleaner workaround than manually draining and shutting down the pinned VMs before each host update, but that's what I'd fall back on in the meantime.

@gleveque Hello Can you retry with an up-to-date XCP-ng 8.3 host? PCI expansion ROM are now enabled when doing pass-through. The guest VM can now access the ROM bar. Regards, Yann