Generated with:
go run ./cmd/gen-manifests -output ./test/data/manifests -cache /var/tmp/manifest-cache/
Includes some changes to ec2-sap manifests due to using the same repos.
It needs to be listed in format-request-map.json, and the repo snapshots
required to depsolve the packages need to be added to repos.json
NOTE: The snapshots are from https://github.com/osbuild/rpmrepo/
This can be shared between cloud providers so move it out of the EC2 SAP
config into its own file and drop the X86_64 from the name (there is
nothing arch specific in it, even if it is only ever used on X86).
Podman doesn't work when installed from 9.1 repositories on 9.0 image.
I found that upgrading the whole system helps. Sadly, that requires removing
python-unversioned-command for some reason.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
Exclude unwanted packages from the EC2-SAP image. These packages have
been pulled into RHEL-9 image due to the fact that we moved away
from using `@core` package group by default and as a result we dropped
explicit package excludes. However the SAP image includes the
`@Server` package group, which pulls in these unwanted packages, thus
we need to explicitly exclude them in the SAP package set.
Related to COMPOSER-1829
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
`amdgpu` module is causing error to be printed in the system log on AWS
instances. After investigation, it turns out that it is not needed.
Disable it by default on all AWS images.
Related to COMPOSER-1807
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
`amdgpu` module is causing error to be printed in the system log on AWS
instances. After investigation, it turns out that it is not needed.
Disable it by default on all AWS images.
Related to COMPOSER-1807
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
F37 no longer ships sil-scheherazade-fonts, but
sil-scheherazade-new-fonts instead. Let's change this. The repos for
test manifests must have been updated in order to get the new package.
Co-authored-by: Ondřej Budai <ondrej@budai.cz>
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
Previously, this just happened silently and let to extremely odd errors. Let's
just print the error to simplify debugging the next time.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
8fdd158799 modified the Cloud API to resolve
ostree commits using a separate job. This change caused the API handler
to call PackageSets without any ostree options (because they are not resolved
yet).
Unfortunately, the new implementation of PackageSets initializes the manifest.
The initialization checks the options and if the type is iot-installer and
it doesn't have the fetch checksum for IoT, it just returns an error.
To work around this (we need an initialized manifest to create the chains),
this commit just gives the initialization method a dummy checksum. The ostree
options currently don't have any effect on the package sets, so this should
be fine.
In order to make this workaround at least slightly sane, a warning is printed,
there's a new test just for this behaviour and a long comment to remember to
delete these lines.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
The new terraform runners support an extended lists of tags to get a
greater precision on the stats we can extract from AWS cost center.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
Don't list specific RHEL repos which should be copied over to the repo
overrides directory, but instead copy all of them, similar to how it's
done for Fedora and CentOS Stream. This is less error prone when support
for a new RHEL release is added to composer.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
This can be shared between cloud providers so move it out of the EC2 SAP
config into its own file and drop the X86_64 from the name (there is
nothing arch specific in it, even if it is only ever used on X86).
This fixes a race condition where the context might have been canceled
or timed out in between the preliminary check and trying to dequeue, and
consequently returning the wrong error.
Instead of doing the preliminary check, just check for the context
errors when trying to dequeue.
This fixes a bug where the listen function would keep trying to use a
closed, unrecoverable connection to listen for a notification. This
continued failing, which essentially made the queue instance useless.
The `LVMVolumeGroup.Clone()` method could end up dereferencing a `nil`
pointer in the `lv` variable, if there would be a `nil` logical volume
in the LVM volume group. Such situation would be an error of its own.
There is no point in checking if the cloned logical volume is not `nil`
and casting it to another variable. The logic should check if the cloned
logical volume is `nil` and panic in such situation. The following code
can then cast the clone to a different variable without issues and there
is no risk of dereferencing a `nil` pointer.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
The `PartitionTable.Clone()` method could end up dereferencing a `nil`
pointer in the `part` variable, if there would be a `nil` partition in
the partition table. Such situation would be an error of its own.
There is no point in checking if the cloned partition is not `nil` and
casting it to another variable. The logic should check if the cloned
partition is `nil` and panic in such situation. The following code can
then cast the clone to a different variable without issues and there is
no risk of dereferencing a `nil` pointer.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
Fedora-37 was using "ci-ssd" instance in RHOS-01 which does not support
nested KVM. This SHA switches them to use "ci" instances which do
support nested kvm.
When the store is written to disk it simplifies the ImageBuild details
into a simple image type string. This works fine for composes that match
the host's distro but isn't enough detail to load composes made for
other distros, especially if the image type name isn't supported on the
host. This results in cross distro compose results being lost after a
reboot.
This fix uses the distro information from the compose's blueprint to
determine which distro the image type should be loaded from. It assumes
that the architecture matches the hosts' arch -- this is currently
always true but in the future if cross-arch builds are added it will
need to be addressed in a different way.
newComposeFromV0, newComposesFromV0, and newStoreFromV0 now take a
pointer to the full distro registry instead of an Arch, this allows them
to access the correct image types for the distro selected by the
blueprint. When loading the composes from disk the blueprint distro is
loaded from the registry before checking the image type string.
This means that we do not have to change the store version or on disk
format, the only thing changing is how it decides to populate the
ImageBuild when reloading the store.
A number of tests use a fake test distro using fake architecture names.
These tests have been adjusted to use a fake distro registry with
overridden host architecture that matches the fake one.
The store needs to know what default distribution and arch to use for
images built with a blueprint that has no distro set. Store the host's
architecture name along side the host's distribution details.
SetHostArchName() allows this host arch to be overridden for testing,
which uses arch names that do not match any host.
If a job is unresponsive the worker has most likely crashed or been shut
down and the in-progress job been lost.
Instead of failing these jobs, requeue them up to two times. Once a job is lost
a third time it fails. This avoids infinite loops.
This is implemented by extending FinishJob to RequeuOrFinish job. It takes a
max number of requeues as an argument, and if that is 0, it has the same
behavior as FinishJob used to have.
If the maximum number of requeues has not yet been reached, then the running
job is returned to pending state to be picked up again.
The consumer key/cert is used to uniquely identify a system against a
candlepin instances. They're useful for any Red Hat (ostree) content
which requires (cert) authentication.
Let's default to not providing the Resource Group Location when testing
Cloud API with Azure and leave it up to the implementation to determine
the correct location to use.
Signed-off-by: Tomáš Hozza <thozza@redhat.com>
