test: Add more distros for composefs test #1810
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request refactors the OS image mapping to support different image types (base and buildroot-base) and updates the build scripts to use this new structure for composefs tests. The changes are logical and enable more flexible testing across different distributions. However, I've noted that RHEL entries were removed from the image map, which could be an unintended breaking change. Please see my specific comment on that.
cgwalters
left a comment
cgwalters
left a comment
There was a problem hiding this comment.
Looks sane to me, though I see C9S failing with what looks like filesystem corruption errors though. Let's just merge as is and debug that as a followup.
BTW though, one thing we should do soon is split up a "build rpm" phase for the GHA because now we're building from source fully twice per OS.
(In general of course lots of overlap between TF and GHA as discussed)
Ah well no can't do that because that's a required job. |
|
Okey, AI tells me the issue on C9S is older kernel and older virtiofs/virtiofsd userspace. The composefs install does many small-file metadata ops (walk /etc/pki/..., copy many small cert files, systemd unit symlinks, etc.). That pattern reliably triggers older virtiofs/virtiofsd bugs that return EUCLEAN (“Structure needs cleaning”) during stat/cp operations. |
|
Maybe. In this case virtiofsd is on the host, but yeah it could be the guest virtiofs support. Anyways...well, let's skip it for now with a TODO? Or we could do this https://docs.github.com/en/actions/reference/workflows-and-actions/workflow-syntax?versionId=free-pro-team%40latest&productId=actions&search-overlay-input=continue-on-error+docs&search-overlay-ask-ai=true#example-preventing-a-specific-failing-matrix-job-from-failing-a-workflow-run Or if we want to narrow in on if it's actually a virtiofs problem, we could run the test in TF (though we'd need to control the filesystem, which really leads to us supporting a kickstart or Ignition flow) |
Do you mean in |
|
Yeah, I file an issue for tracking and use |
henrywang
force-pushed
the
distros
branch
3 times, most recently
from
66637cd to
175b43a
Compare
And ignore test-integration-cfs failure on centos-9 Signed-off-by: Xiaofeng Wang <[email protected]>
The build-sealed script introduced in PR bootc-dev#1810 referenced BOOTC_buildroot_base which is only defined in GitHub Actions CI, causing failures when running 'just build-sealed' manually. Changes: - Add buildroot_base parameter to build-sealed script with default value of quay.io/centos/centos:stream10 - Define buildroot_base in Justfile from BOOTC_buildroot_base env var - Pass buildroot_base as 4th argument to all build-sealed invocations - Reorder arguments to put buildroot_base before optional secureboot This allows manual execution while maintaining CI compatibility. Assisted-by: Claude Code (Sonnet 4.5) Co-Authored-By: Claude <[email protected]> Signed-off-by: Wei Shi <[email protected]>
The build-sealed script introduced in PR bootc-dev#1810 referenced BOOTC_buildroot_base which is only defined in GitHub Actions CI, causing failures when running 'just build-sealed' manually. This allows manual execution while maintaining CI compatibility. Signed-off-by: Wei Shi <[email protected]>
The build-sealed script introduced in PR #1810 referenced BOOTC_buildroot_base which is only defined in GitHub Actions CI, causing failures when running 'just build-sealed' manually. This allows manual execution while maintaining CI compatibility. Signed-off-by: Wei Shi <[email protected]>
2 participants
This PR adds CS9, Fedora 42/43/44 distros for composefs test.