Fedora's timeout for ssh-keyscan is short by default, but RHEL's is much
longer. Set the timeout to 5 seconds for consistency.
Signed-off-by: Major Hayden <major@redhat.com>
The subpackages are not really meant to be used with a different version of
the osbuild-composer package. This commit enforces.the usage of a matching
version in the spec file.
When osbuild crashes (e.g. when cp fails because of the machine running
out of disk space), it doesn't produce a machine-readable result. Due to
our suboptimal handling of the result struct (this is my fault), this can
lead to result == nil. However, composer expects that result != nil in all
cases because it uses the Success flag to assess the compose state. If
result == nil, it just crashes terribly.
Exit the whole worker process when a job was canceled, because osbuild
does not clean up all child processes when receiving SIGKILL.
Change the service to restart osbuild-worker also on success, and
decrease the restart timeout.
On some environments (like RHEL gating) there's no virtualization available.
This commit adds -disable-local-boot argument to osbuild-image-tests. When
this argument is present, the local booting is skipped. This doesn't affect
the cloud booting, the test binary still tries to do that. If no credentials
are available, the fall back to local booting will be skipped if
-disable-local-boot is given.
When osbuild-composer crashed, it left temporary directories in
`/var/cache`. Use `/var/tmp` for these output directories, because
systemd will clean these up (we set PrivateTmp=true).
Also, put the store into `/var/cache/osbuild-store`. The worker does not
checkpoint anything. The store is only used as a cache for rpms. That
can be shared between multiple workers and successive runs of a single
worker.
Speed up the deployment during CI runs and avoid the need for lots of
extra packages to run Ansible. This gets us closer to running bare
images for each distro that better simulates a customer environment.
Signed-off-by: Major Hayden <major@redhat.com>
On RHEL, `VERSION_ID` shows up as `8.2` or `8.3`, but we need it to be
just `8`.
Remove the period and everything after it in `$VERSION_ID`.
Signed-off-by: Major Hayden <major@redhat.com>
This patch introduces support for s390x architecture. It has some
specific requirements like using kernel-cmdline stage before the RPM
stage because RPM scriptlets require these values to be present in the
system already. It also uses zipl stage and bootloader as opposed to all
the other architectures.
Also allow a list of them to be passed. Generating 'all' image types
makes less sense now that different distros/arches support different
ones.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Initial definition for RHEL for edge package set on x86.
This is based marginally on the upstream Fedora IoT package set
but slimmed down and adjusted for RHEL package naming.
Based on Tom's patch.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Gather logs directly for AWS + qemu tests using `composer-cli compose
log` and also gather the metadata. This will hopefully help diagnost the
problems we are seeing in #754.
Signed-off-by: Major Hayden <major@redhat.com>
* Ensure we get the output of `composer-cli compose start` for
diagnostics
* Hide the annoying `genisoimage` output
* Simplify the `IMAGE_EXTENSION` check
* Make it more clear what is happening while we wait for ssh
Signed-off-by: Major Hayden <major@redhat.com>
This updates the test-cases to reflect the changes in the previous few
commits.
The osbuild rpm stage options format was changed, and the build root on
RHEL8 now includes SELinux policy, and is labelled correctly.
None of these changes affect our output images, so the image info is
unchanged.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Allow individual test-cases or sub-sets of test-cases to be generated
more easily.
We allow explicit skipping of image-info generation (and hence the
osbuild run), and also individual image types to be specified.
Also drop distros and image types that are no longer supported.
Signed-off-by: Tom Gundersen <teg@jklm.no>
We require the right SELinux labels to be set on some of the tools we
use, make sure the policy is installed and the labels applied in the
buildroot, in the same way as in the target image.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Rather than passing an array of checksums, pass an array of objects
with package properties. One is the `checksum`, which is required, and
in addition to that we now pass the boolean check_gpg, which indicates
that that RPM should have its GPG signature verified.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Currently each repo is annotated with this information, use the mapping
we have from packages to their originating repos to annotate each
PackageSpec with the same information.
This information is not serialized, and not yet used. It may be used to
expose this information to osbuild's org.osbuild.rpm stage.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Use the new dnfPackageSpec only for communicating with dnf-json. This
allows us to drop the RepoID from the public type, as that is only meant
to be used internally in `Depsolve()`.
Signed-off-by: Tom Gundersen <teg@jklm.no>
The sources weldr API already supports this, so hook it up to be
represented on disk and in our internal state tracking too.
This does not yet hook this up to be respected by osbuild, which
currently takes this to be unconditionally set to true.
Signed-off-by: Tom Gundersen <teg@jklm.no>
When gsl with version * was specified in the blueprint,
composer depsolved both x86_64 and i686 version of gsl.
This test case should prevent this from happening.
gsl is used because it has x86_64 and i686 versions on both RHEL and Fedora.
Also, gsl-devel package exists, which is not dependant on gsl and shouldn't
be depsolved.
blueprint.GetPackages() method was used to depsolve blueprints prior the
dnf-to-rpm switch. However, it got dropped during the switch. This commit
makes weldr use it again.
The nice effect of this change is that we can drop getPkgNameGlob function and
have only one function for getting package name-versions from a blueprint.
Also, blueprint.GetPackages() works better with * version. Previously, we had
issues with composer depsolving bash of version * to both x86_64 and i686
versions of bash package. GetPackages() converts the package to name-version
of just bash, which dnf-json correctly depsolves to just one architecture. On
the contrary, the previous method converted bash to name-version bash-*.*.*,
which confused dnf-json.
Note that conversion to bash-* is also wrong because this will cause dnf-json
to install all packages with prefix "bash-*".
This makes it easier to use the test binary with the `-run` argument.
Instead of the full path:
-test.run TestImages//usr/share/tests/osbuild-composer/cases/rhel_8.2-x86_64-openstack-boot.json
this only requires the actual name:
-test.run TestImages/rhel_8.2-x86_64-openstack-boot.json
When edd7b37ea added `--output-directory` to the invocation of osbuild,
it also removed `--store`.
This was a mistake: osbuild's default store is `.osbuild`, which is not
what we want. Restore the old behavior of passing a temporary directory,
but use the same for each test run.
Treating stdout and stderr separately makes it hard to match what
happened when. It's also easy to miss when `-v` is passed to the test
binary.
Print the output to stdout when osbuild fails, because the test
framework we're using does not print errors if they're too large.
Also, don't special-case exec.ExitError. Output might be useful in any
case.
