I ran into this issue as well after upgrading nova containers built off the latest stable/wallaby
After deployment of nova-libvirt, new uefi machines no longer boot and the following is logged
2022-03-22 17:02:32.102 7 ERROR nova.compute.manager [req-9b5bf117-f72b-4ea6-b059-e937e2db6212 89fee540585c9929befdb0c40c845dc47552126acf7b8b18fafce5393cdd818c b8d86bc97cf54605ae41d37ac70178b4 - d3acf2f34f8f47ca987b500937c7be58 d3acf2f34f8f47ca987b500937c7be58] [instance: 6aaa5ee5-aa10-4be5-b943-d8dea6a5080a] Failed to build and run instance: nova.exception.UEFINotSupported: UEFI is not supported
Existing uefi machines could still reboot, but would no longer boot if shut down at the hypervisor.
I had to re-deploy the older nova-libvirt container as a workaround for now.
# The old centos8s container, based off master from around 5 months ago.
()[nova@0cc7e1c86c17 /]$ ls /nova-base-source/
nova-23.0.3.dev23
# The new centos8s container, based off stable/wallaby from 2-3 days ago.
()[root@67c6853d02ba /]# ls /nova-base-source/
nova-23.2.1.dev1
I've tried building containers off several other different branches and versions of nova (including master), but all result in the same error. Only deploying that old container again as a bandaid solution, for now, allows uefi instances to boot again.
I ran into this issue as well after upgrading nova containers built off the latest stable/wallaby
After deployment of nova-libvirt, new uefi machines no longer boot and the following is logged
2022-03-22 17:02:32.102 7 ERROR nova.compute. manager [req-9b5bf117- f72b-4ea6- b059-e937e2db62 12 89fee540585c992 9befdb0c40c845d c47552126acf7b8 b18fafce5393cdd 818c b8d86bc97cf5460 5ae41d37ac70178 b4 - d3acf2f34f8f47c a987b500937c7be 58 d3acf2f34f8f47c a987b500937c7be 58] [instance: 6aaa5ee5- aa10-4be5- b943-d8dea6a508 0a] Failed to build and run instance: nova.exception. UEFINotSupporte d: UEFI is not supported
Existing uefi machines could still reboot, but would no longer boot if shut down at the hypervisor.
I had to re-deploy the older nova-libvirt container as a workaround for now.
# The old centos8s container, based off master from around 5 months ago. 0cc7e1c86c17 /]$ ls /nova-base-source/
()[nova@
nova-23.0.3.dev23
# The new centos8s container, based off stable/wallaby from 2-3 days ago. 67c6853d02ba /]# ls /nova-base-source/
()[root@
nova-23.2.1.dev1
I've tried building containers off several other different branches and versions of nova (including master), but all result in the same error. Only deploying that old container again as a bandaid solution, for now, allows uefi instances to boot again.