As builders are not re-used caching has no use and can likely only
lead to disk-size issues. This can be re-enabled when we have an
outside-of-the-builder cache.
Instead of going directly fatal inside the handlers this commit provides
a wrapper function allowing handlers to instead return an `error`; this
will allow for checking deferred errors.
The image type supports uploading to both s3 and container registries.
Right now we only care about getting it in Koji, so let's just add it to
the s3 upload section.
The container registry test also embeds containers in the image, which
the iot-bootable-container doesn't support (yet).
Don't allow unbound variables, but for the variables that are used to
determine whether or not that part of the setup should continue, default
to empty/undefined.
This is necessary to get the proper osbuild SHA to install the latest
required osbuild version into the image.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Specify the resource pool when importing the VMDK to the VCenter.
This should prevent the following error:
govc: default resource pool resolves to multiple instances, please specify
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Add any errors to job error details when an osbuild build fails.
Otherwise these won't show up in the worker log, which makes
debugging issues harder.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
If a depsolve job fails, the error details were not passed as details to
the manifest job error details. This may help with debugging failures.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
The test would loop for the job timeout limit waiting for composer
container, in case the composer API does not come up for whatever
reason. Modify the test case to wait only for 12x 10s before failing.
In case the composer API does not come up, print the logs from the
composer container and exit with non-zero code.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Add test case for verifying that distros and repo configurations work
with and without the dot-notation in the on-prem scenario. Run the test
case on the latest RHEL 8 and 9 versions in development and verify the
default distro aliases behavior on them.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Extend the `api.sh` to allow testing compose requests with distro name
which does not use dot-notation (specifically when the dot is removed
from the distro name as it used to be in the past). In addition to that,
allow also testing the distro alias using distro name without the minor
version in compose requests.
Enable these two new test variants in the CI.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Update RHEL 8.9 and 9.3 repo definitions to use the CDN repos, since
these are already GA.
Add SAP repositories to all RHEL repo configs, to be able to build the
SAP image for testing purposes.
Fix minor issues found in repos (e.g. 8.8 RT repo pointing to 8.7,
etc.).
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Using the DISTRO_CODE with dot to separate major and minor version is
breaking cloud-image-val testing. Specifically, the tool is using
regular expression to search for an uploaded image, but the expression
is not expecting any dot in the version.
This will be fixed in [1], but merging it will take some time due to CIV
CI being currently broken. Due to this, workaround the problem for now
by making sure that the TEST_ID, which is used to construct the
IMAGE_NAME, does not contain any dot in the DISTRO_CODE.
[1] https://github.com/osbuild/cloud-image-val/pull/290
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Where applicable, modify all repo config filenames to use a dot
to separate the release major and minor version. Modify test cases
to not remove dot from the distro version any more.
Existing tests will be extended (or new tests added) to explicitly test
backward compatibility and ensure that using old distro names without a
dot still works.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
While I was the one to push for TEST_MODULE_HOTFIXES to be passed as an
argument to `api.sh`, when it was implemented, this is turning our to be
impractical. I would like to extend `api.sh` to test distro names with
and without the dot to separate major and minor version and also to test
the ability to specify distro name explicitly (e.g. a distro alias).
Passing all of these new options and their combinations would make the
CLI go wild.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Any changes done to the SPEC file were not reflected in the
`rpmbuild/SPECS` directory across runs of `make rpm`. One had to delete
the SPEC file manually to be updated and used for the RPM build.
Mark the target as PHONY to ensure that the SPEC is always updated.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Weldr API used to determine the distro name from the image type, when it
was getting the repositories to use for depsolving and for the actual
depsolving (solver uses the distro name to namespace cache).
This used to be OK, but with the introduction of distro name aliases,
the distro name used to get the distro object may not be the same as the
name returned by the actual distro object. To preserve the current
behavior, the same name used to get the distro object should be used to
also get the repositories for depsolving and to namespace depsolving
cache.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
