qemu,raw: Ubuntu 20.04 EFI builds are broken

Question

qemu,raw: Ubuntu 20.04 EFI builds are broken

johananl opened this issue 2 years ago · 22 comments

What steps did you take and what happened:

PACKER_LOG=1 make build-qemu-ubuntu-2004-efi

The build fails with the following error:

2022/05/11 13:21:06 packer-builder-qemu plugin: Qemu stderr: qemu-system-x86_64: -drive file=OVMF.fd,if=pflash,format=raw,readonly=on: Could not open 'OVMF.fd': No such file or directory
==> qemu: Error launching VM: Qemu failed to start. Please run with PACKER_LOG=1 to get more info.

~~The same is theoretically true for make build-raw-ubuntu-2004-efi, however raw builds are currently broken due to #879.~~
Update: The issue reproduces for raw builds, too.

What did you expect to happen:
I expected the build to work.

Anything else you would like to add:
Looks like OVMF.fd should be a firmware file used to emulate EFI in VMs. Ideally we should create (or download) this file automatically upon build. If we can't do that, we should document that it's required and probably exempt the EFI builds from make build-{qemu|raw}-all.

Environment:

Project (Image Builder for Cluster API, kube-deploy/imagebuilder, konfigadm): image-builder

Additional info for Image Builder for Cluster API related issues:

OS (e.g. from /etc/os-release, or cmd /c ver): Ubuntu 20.04 LTS
Packer Version: v1.8.0
Packer Provider: qemu
Ansible Version: irrelevant
Cluster-api version (if using): irrelevant
Kubernetes version: (use kubectl version): irrelevant

/kind bug

cc @MaxRink

johananl commented a year ago

/assign

Answer 1 · 2022-08-09T10:55:44.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Answer 2 · 2022-08-09T13:46:05.000Z

/remove-lifecycle stale

Answer 3 · 2022-11-07T14:09:30.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Answer 4 · 2022-11-08T09:23:18.000Z

/remove-lifecycle stale

Answer 5 · 2023-02-06T09:27:24.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue or PR as fresh with /remove-lifecycle stale
Mark this issue or PR as rotten with /lifecycle rotten
Close this issue or PR with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Answer 6 · 2023-02-07T10:00:09.000Z

/remove-lifecycle stale

Answer 7 · 2023-05-08T10:55:01.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue as fresh with /remove-lifecycle stale
Close this issue with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Answer 8 · 2023-05-08T11:26:13.000Z

/remove-lifecycle stale

Answer 9 · 2023-06-19T18:44:07.000Z

From my (limited) understanding, this requires some qemu(-kvm) tooling to be installed onto the machine you're running this on.
We managed to get around this at Giant Swarm by using the docker image running the following:

apt-get update && apt-get install -y qemu qemu-kvm

Maybe it makes sense to have this built in to the standard image-builder container image to at least make sure that is useable out-of-the-box with qemu.

I'm not sure how best to handle this when running make directly other than stating it as a pre-requisite. Do you have any ideas?

Answer 10 · 2023-06-19T18:45:18.000Z

Actually, looks like this is now added to the Docker image: 6b2a0ec

Answer 11 · 2023-08-10T20:57:46.000Z

@johananl is this still an issue now that #879 is closed? (See also #1122.)

Answer 12 · 2023-08-14T11:28:16.000Z

The OVMF.fd error still reproduces for me @mboersma. IIUC #879 didn't affect this issue but rather the ability to try to reproduce this issue for raw builds (because you need to be able to run a raw build in order to determine if the EFI raw build is broken).

Thanks for the info @AverageMarcus. AFAICT you're talking about the QEMU/KVM dependency which is off course necessary, however I'm not sure how satisfying this dependency fixes the problem with the missing OVMF.fd file.

Answer 13 · 2023-08-14T11:30:13.000Z

This seems very relevant: https://github.com/tianocore/tianocore.github.io/wiki/How-to-run-OVMF

Answer 14 · 2023-08-14T12:03:37.000Z

I've reproduced the problem using the Docker-based build, too:

docker run -it --rm --net=host -e PACKER_LOG=1 registry.k8s.io/scl-image-builder/cluster-node-image-builder-amd64:v0.1.17 build-qemu-ubuntu-2004-efi

Answer 15 · 2023-08-14T12:16:31.000Z

I'll try to build and use OVMF using the instructions above and report here.

Answer 16 · 2024-01-26T08:30:58.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue as fresh with /remove-lifecycle stale
Close this issue with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

Answer 17 · 2024-01-29T16:15:06.000Z

/remove-lifecycle stale

Answer 18 · 2024-04-28T16:55:48.000Z

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

Mark this issue as fresh with /remove-lifecycle stale
Close this issue with /close
Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale