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
{{ message }}
Repository navigation
Unable to retrieve UUID of windows VM using wmic command #11512
Unable to get the UUID information from a Windows VM deployed on hosts with AMD processors using the: wmic path win32_computersystemproduct get uuid command, which works on hosts with Intel processors.
@DaanHoogland we have been doing some tests with different KVM host OS against 4.19 and 4.20 and found out that is not a CloudStack regression but a problem with the latest libvirt and qemu versions. Tests performed at the moment:
KVM Host Operating System / CloudStack Version
Oracle Linux 8 - libvirtd 8.0.0 0, QEMU emulator version 6.2.0 (qemu-kvm-6.2.0-53.module+el8.10.0+90638+6d8ade84.4)
We have completed the test matrix with different KVM OS against 4.19.3 and 4.20.1, it seems to be a regression from libvirt version 10 + qemu version 8 (included on Ubuntu 24) - the rest of the OSes with previous libvirt and qemu versions work fine on both CloudStack versions
While Doing that I saw CloudStack-Agent also getting removed as Qemu was being used by Agent.
Have you checked with Libvirt 10 and Qemu 7.2x or 6.2x - Does this combination works?
@rohitpals this is a failing interaction with libvirt/qemu due to changed behaviour in those programs. This is not a regression in CloudStack. We do not know the reason the integration API has changed in behaviour and where it should be fixed.
Thank you for the update.
We are aware of this workaround, and we can retrive the UUID as well.
The issue is that the 3rd Party Applications uses the "wmic" command for getting the UUID and apply licenses.
Hence we asked for the workaround.
Downgrading the OS to Ubuntu 22 is the only option?
Coz when we downgrade LIBVIRT or QEMU to older versions, the OS becomes unstable and VM provisioning job hampers.
@rohitpals , I think you can experiment with running older version of qemu on newer versions of ubuntu, but indeed the more stable option is to use ubuntu 22 as host os. Please keep us up to date on how you fare. Also keep an eye on the upstream repos for any resolutions.
@DaanHoogland Thank you for the suggestion. Will do revert.
Meanwhile, could you please let us know if in the Production infra, is it fine if we downgrade the OS to Ubuntu22 and migrate the VMs to it from other Ubuntu22 hosts? Will that cause any issues? If you are aware?
@DaanHoogland Thank you for the suggestion. Will do revert.
Meanwhile, could you please let us know if in the Production infra, is it fine if we downgrade the OS to Ubuntu22 and migrate the VMs to it from other Ubuntu22 hosts? Will that cause any issues? If you are aware?
Rohit, I think you are asking something I cannot know. There are too many other factors and I am not familiar with your environment. In general it is not safe to do something in production that you have not tested before. Please only apply suggestions you read here on test environments.
@rohitpals I have got this answer on the reported issue from libvirt project: https://gitlab.com/libvirt/libvirt/-/issues/807#note_2724045702. I have manually verified changing the machine type to: machine='pc-i440fx-8.0' and could obtain the UUID through the wmic command. Can you please try this option on a test environment?
On the KVM host, virsh edit <DOMAIN> and change the machine type
Stop and start the VM domain through libvirt, not through CloudStack as it will destroy the domain on stop and recreate it on start
Hi Nicolas,
Thank you for the update.
We already tried that but when we stop-start the VM, it resets to default.
Could you please let us know how to make it permanent when we edit the xml
file?
*Thanks & Regards,Rohit P*
@rohitpals please try virsh shutdown <DOMAIN> after editing the XML, then virsh start <DOMAIN>. This workaround is not permanent, but could help you obtaining the UUID on Ubuntu 24.
Activity
DaanHoogland commentedon Aug 26, 2025
@Pearl1594 , this is a regression on 4.20, right? or is it new/on 4.19?
nvazquez commentedon Sep 1, 2025
@DaanHoogland we have been doing some tests with different KVM host OS against 4.19 and 4.20 and found out that is not a CloudStack regression but a problem with the latest libvirt and qemu versions. Tests performed at the moment:
nvazquez commentedon Sep 1, 2025
We have completed the test matrix with different KVM OS against 4.19.3 and 4.20.1, it seems to be a regression from libvirt version 10 + qemu version 8 (included on Ubuntu 24) - the rest of the OSes with previous libvirt and qemu versions work fine on both CloudStack versions
rohitpals commentedon Sep 2, 2025
@nvazquez Could you please let us know if there is any Workaround to resolve this issue.
We can't downgrade the OS from Ubuntu 24 to ubuntu 22.
DaanHoogland commentedon Sep 2, 2025
@rohitpals , maybe you can downgrade qemu and libvirt?
rohitpals commentedon Sep 2, 2025
While Doing that I saw CloudStack-Agent also getting removed as Qemu was being used by Agent.
Have you checked with Libvirt 10 and Qemu 7.2x or 6.2x - Does this combination works?
rohitpals commentedon Sep 2, 2025
@weizhouapache I see the label as "not-a-bug"
Could you please confirm if this is not a BUG ANY MORE?
rohitpals commentedon Sep 2, 2025
@nvazquez @DaanHoogland @weizhouapache @Pearl1594
Please confirm whether this a bug or not?
DaanHoogland commentedon Sep 2, 2025
@rohitpals this is a failing interaction with libvirt/qemu due to changed behaviour in those programs. This is not a regression in CloudStack. We do not know the reason the integration API has changed in behaviour and where it should be fixed.
nvazquez commentedon Sep 2, 2025
@rohitpals we have also raised the issues on the libvirt and QEMU projects. Please find the following links: https://gitlab.com/libvirt/libvirt/-/issues/807 and https://gitlab.com/qemu-project/qemu/-/issues/3098
nvazquez commentedon Sep 2, 2025
As a workaround, you can still obtain the VM UUID through the VR user data by the URL:
http://<VR_IP>:80/latest/meta-data/instance-idrohitpals commentedon Sep 2, 2025
@DaanHoogland @nvazquez
Thank you for the update.
We are aware of this workaround, and we can retrive the UUID as well.
The issue is that the 3rd Party Applications uses the "wmic" command for getting the UUID and apply licenses.
Hence we asked for the workaround.
Downgrading the OS to Ubuntu 22 is the only option?
Coz when we downgrade LIBVIRT or QEMU to older versions, the OS becomes unstable and VM provisioning job hampers.
DaanHoogland commentedon Sep 2, 2025
@rohitpals , I think you can experiment with running older version of qemu on newer versions of ubuntu, but indeed the more stable option is to use ubuntu 22 as host os. Please keep us up to date on how you fare. Also keep an eye on the upstream repos for any resolutions.
rohitpals commentedon Sep 2, 2025
@DaanHoogland Thank you for the suggestion. Will do revert.
Meanwhile, could you please let us know if in the Production infra, is it fine if we downgrade the OS to Ubuntu22 and migrate the VMs to it from other Ubuntu22 hosts? Will that cause any issues? If you are aware?
DaanHoogland commentedon Sep 2, 2025
Rohit, I think you are asking something I cannot know. There are too many other factors and I am not familiar with your environment. In general it is not safe to do something in production that you have not tested before. Please only apply suggestions you read here on test environments.
nvazquez commentedon Sep 2, 2025
@rohitpals I have got this answer on the reported issue from libvirt project: https://gitlab.com/libvirt/libvirt/-/issues/807#note_2724045702. I have manually verified changing the machine type to:
machine='pc-i440fx-8.0'and could obtain the UUID through the wmic command. Can you please try this option on a test environment?virsh edit <DOMAIN>and change the machine typerohitpals commentedon Sep 3, 2025
nvazquez commentedon Sep 3, 2025
@rohitpals please try
virsh shutdown <DOMAIN>after editing the XML, thenvirsh start <DOMAIN>. This workaround is not permanent, but could help you obtaining the UUID on Ubuntu 24.harikrishna-patnala commentedon Sep 16, 2025
fixed by #11569