Search results

no, the latest iteration just moved the place where to configure the amount of threads, but the underlying logic stayed the same not sure if that would help, for that we'd have to determine where the bottleneck comes from in that scenario (iow. are your disks now saturated with this?, cpu...

the logs look unsuspicious except the lpfc errors, especially these lines: lpfc 0000:03:00.0: 0:3176 Port Name 0 Optics faulted/incorrectly installed/not installed - Reseat optics, if issue not resolved, replace. ... lpfc 0000:03:00.1: 1:3176 Port Name 1 Unqualified optics - Replace with Avago...

i pinged the latest patch recently, here a link to the mali thread: https://lore.proxmox.com/pbs-devel/20250221150631.3791658-1-d.csapak@proxmox.com/

this seems to be an issue of the guest driver/graphics stack and not the virtualization anymore...

thanks for sending this. it seems there are so many errors from the lpfc driver (fibre channel) that the beginning is cut off you could give me the relevant output with journalctl -b -t kernel depending on how large it is, you might want/need to zip/tar.gz it but it seems the fc connection...

interesting.. is this reproducible on your side? (install 22h2, upgrade to 24h2) if yes, do you want to share a few more details (e.g. which windows version, specific settings, etc.) so i can look into that?

you actually have to pay NVIDIA a license to get a hold of the vGPU host and guest drivers. also note that this normally only works with some of their data center/workstation GPUs, not with consumer ones.

it seems that's not the whole output of dmesg. best to redirect into a file and either attach it here : dmesg > /tmp/dmesg.txt

yes, if you want to get support you have to have an nvidia entitlement according to their licensing terms: https://docs.nvidia.com/vgpu/18.0/grid-licensing-user-guide/index.html

can you post the output of 'dmesg' ?

the easiest way to determine demand for us is when people open a bug and comment there (as i already wrote), not all developers see all forum threads, but most are aware of high activity bugs. It seems the OP of this thread did also not leave a comment on the linked bug, so I'd encourage you to...

couldn't you connect via the pve web ui to the container and get the logs out this way? (or using 'pct enter ID' on the host cli) without more information it's not really possible to help (a shot in the dark could be: firewall settings, duplicate ip, etc..)

yes, installing the package with apt is the right way, and if you remove the dkms module it can't fail anymore (it depends on how you installed the dkms module in the first place, but if you added it straight with dkms, then removing it the same way is correct)

Hi, the easiest way to confirm is to run the pve7to8 tool again, if everything is fine, the warnings should be gone :)

Hi, the error you posted happens before any guest os is booted/involved, so i doubt that fedora is the culprit here. can you post both vm configs? (my guess is that the windows vm either does not use 'x-vga' or uses ovmf, which makes the 'x-vga' option more or less useless) in any case, try...

no bugtracker, but when we implement it, it will be rolled out in the pdm repositories like any update

just to chime in here: the reason that the 'args' part is root only is because it can contain arbitrary qemu commandline options, which would be a massive security issue if anybody could do that (e.g. one could passthrought the root disk or other sensible files to a vm) most of the time the...

yeah the implementation is far from something usable, but we'll work on it!

i moved the thread

so it works on linux? this would indicate a windows driver issue.... i meant removing that when you try to boot windows if it works then