Previously, the script didn't distinguish between the scheduled RHEL-X
nightly pipeline and the scheduled RHEL GA pipeline when reporting
status back to GitHub. Moreover the scheduled RHEL GA pipeline on GitLab
had RHEL_MAJOR variable set to "9". As a result, the scheduled RHEL GA
pipeline status has been reported to GitHub as "Schutzbot on GitLab,
RHEL-9-nightly", which effectively overrode the RHEL-9 nightly pipeline
status for the same commit.
Fix the script reporting the status back to GitHub, to take the NIGHLY
variable value into consideration when building up the status context.
I've also modified the RHEL GA scheduled pipeline settings in GitLab to
not specify any RHEL_MAJOR, which was a relict from pre-RHEL-10 days.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Fix RHEL 10 mock confings not being created if the machine was subscribed and fix RHEL 9 being used in the RHEL 10 mock confings for a repo path.
These issues caused the rpm jobs for RHEL 10 GA to fail.
Update the terraform ref to the one that contains F42 configs. This is
needed to start testing anything on F42.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Add a workaround for the issue that podman falls back to the 'cni'
network backend when finding any container images in the local storage
when executed for the first time. Since we started embedding container
images in our CI runner images, this resulted in failures, because the
OS is missing some required CNI plugins. Until we somehow fix this in
osbuild, we explicitly set the network backend to 'netavark'.
This is relevant only for RHEL-9 / c9s, because Fedora since F40 and
el10 support only `netavark` backend.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
The original ref pointed to a commit, which used the plain Fedora 41
Cloud Base images, which don't have any test dependencies installed. Use
the latest commit from `main`.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
We are on a quest to reduce clutter on our Slack channels. Thus,
I decided to simplify the daily CI notifications:
- the link to the edge pipelines got removed, it's now in bookmarks
- several words were removed to make the message shorter
- the link to the pipeline is now a hyperlink
- the whole message should be a one liner
- less text is now bold
I've also simplified the format in which we send the message. I think
that the block format used before makes redundant line-breaks.
Unfortunately, the mentions need to be done using user IDs instead
of user names. If you ever need to find them, go to the user's profile,
click on the three dots and select "Copy member ID".
Add workarounds needed for mockbuild.sh to work on c10s / el10, due to
non-existent EPEL-10 and issues on c10s.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
The pulp client is very large and defines a lot of symbols in on
package, which causes very large memory usage on el8 aarch64 (presumably
because of 64k page sizes).
Adding a 1 GiB swapfile fixes the issue in our CI runners.
When the container with osbuild-composer gets built in our CI or by
AppSRE, we do not set the composer version to any value (as we do when
we built RPMs). As a result, the version reported by composer is always
"devel". This is not useful for debugging and determining the used
version of composer. In addition, this information now gets exposed in
Koji builds, therefore it makes sense to make it useful.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
