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
I'm trying to experiment with creating a bootc and bootupd compatible container image standalone with DNF (no layering from bootc base images nor bootstrapping from rpm-ostree), however when I try to run bootc backend generate-update-metadata so bootc can install bootupd properly, it fails with:
The logger seems to not log enough data to point exact file paths (preferrably at TRACE level) to say which files are being accessed, so this might be more of an issue in the long run.
Looking at the code, the top-level call to generate_update_metadata() is wrapped with .context("generating metadata failed"), but deeper filesystem operations inside that call (e.g. file reads/writes or directory accesses) aren't consistentlywrapped with .context() to include path details. As a result, when something like an ENOENT occurs, the error bubble-up loses valuable context, especially the actual file path being accessed.
Thanks for the report!
I am not quite sure to catch your meaning, IMU, you can not install kerenl on normal fedora container, but agree that we should do improvement on bootupd to get clear output for the error.
This is not to fix the issue, add context to print more info for
coreos#915
For example:
`error: generating metadata failed: Opening /usr/lib/bootupd/
updates/EFI: No such file or directory (os error 2)`
This is not to fix the issue, add context to print more info for
coreos#915
For example:
`error: generating metadata failed: Opening /usr/lib/bootupd/
updates/EFI: No such file or directory (os error 2)`
I'm trying to experiment with creating a bootc and bootupd compatible container image standalone with DNF (no layering from bootc base images nor bootstrapping from rpm-ostree), however when I try to run
bootc backend generate-update-metadata
so bootc can install bootupd properly, it fails with:The logger seems to not log enough data to point exact file paths (preferrably at TRACE level) to say which files are being accessed, so this might be more of an issue in the long run.
Looking at the code, the top-level call to
generate_update_metadata()
is wrapped with.context("generating metadata failed")
, but deeper filesystem operations inside that call (e.g. file reads/writes or directory accesses) aren't consistently wrapped with.context()
to include path details. As a result, when something like an ENOENT occurs, the error bubble-up loses valuable context, especially the actual file path being accessed.Also, here's my Containerfile:
The text was updated successfully, but these errors were encountered: