You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With upgrade of my system to Fedora 39 and qemu 8.1, provisioning of new vagrant w2k22 boxes using the windows-vagrant scripts stopped working. I've traced it to this qemu issue - https://gitlab.com/qemu-project/qemu/-/issues/2008
The powershell command "Get-CimInstance -Class Win32_ComputerSystemProduct" returns an empty string because of this issue, and this causes scripts to fail.
As a workaround, I change machine type on the windows-2022.pkr.hcl from "q35" to "q35,smbios-entry-point-type=32". I don't know how fast this will be fixed in qemu and backported to Linux distros. If it is not fixed, consider making this a permanent change.
The text was updated successfully, but these errors were encountered:
Good to known there is a workaround for the recent Fedora.
The only QEMU 8.1 that I've tested with was the QEMU flavor that ships with Proxmox 8.1.3 (the pve-qemu-kvm 8.1.2-6 package), and it works fine. I can install the VM using packer, launch a VM (using https://github.com/rgl/terraform-proxmox-windows-example), and inside the launched VM:
PS C:\Users\vagrant> Get-CimInstance -Class Win32_ComputerSystemProduct
IdentifyingNumber :
Name : Standard PC (Q35 + ICH9, 2009)
Vendor : QEMU
Version : pc-q35-8.1
Caption : Computer System Product
It's conceivable that Proxmox have patched their version of qemu to fix this, given that they reported the qemu issue. I've opened a fedora bug as well - https://bugzilla.redhat.com/show_bug.cgi?id=2255791
With upgrade of my system to Fedora 39 and qemu 8.1, provisioning of new vagrant w2k22 boxes using the windows-vagrant scripts stopped working. I've traced it to this qemu issue - https://gitlab.com/qemu-project/qemu/-/issues/2008
The powershell command "Get-CimInstance -Class Win32_ComputerSystemProduct" returns an empty string because of this issue, and this causes scripts to fail.
As a workaround, I change machine type on the windows-2022.pkr.hcl from "q35" to "q35,smbios-entry-point-type=32". I don't know how fast this will be fixed in qemu and backported to Linux distros. If it is not fixed, consider making this a permanent change.
The text was updated successfully, but these errors were encountered: