Introduce a DistroRegister object. For now this does not introduce
any functional changes, as the object is always instantited to be
the same. However, in follow-up patches it will get options.
Signed-off-by: Tom Gundersen <teg@jklm.no>
In RHEL golang dependencies must be vendored, whereas on Fedora they
must be packaged separately. Add conditionals accordingly.
Some macros are not yet available in RHEL, so fall back to older
versions. We may want to just use macros avilable everywhere
unconditionally, but I left that for a follow-up.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Commit b1c5ef2a introduced support for retrieving logs from osbuild.
This commit finishes the second part - actually returning the logs
from /compose/logs route.
dnf-json relies on dnf's ability to cache repository metadata. This is
important, because the API calls it quite often to serve requests for
package lists and depsolves.
However, osbuild's dnf stage always fetches new metadata, because it
doesn't have access to the host's cache. Since metadata is valid for
some time, even after a repository changed, the checksum we put in
the pipeline might be old.
Force a new metadata download when producing the pipeline. This is still
not perfect, but greatly reduces the probability of putting stale
metadata into the pipeline.
When restarting composer, we were not handling the compose states
correctly.
This resolves that as follows:
* any running composes are marked as failed,
* any waiting composes are put back in the pending jobs queue
As a consequence of needing the ability to reinitialize the job
queue, we must include the depsolved pipeline in the compose object.
This is the correct thing to do, as the semantics we currently
adhere to is that pipelines are depsolved when the compose is
started (and restarting composer should not affect this by for
instance re-depsolve the pipeline).
Resolves rhbz#1784062.
Signed-off-by: Tom Gundersen <teg@jklm.no>
When group names are passed on to dnf, they must be prefixed with an
ampersand, or they are treated as a regular package, potentially
causing the build to fail.
Add a testcase to verify this behavior.
This resolves rhbz#1784035.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Right now, there is no way to get at the name from a Distro instance.
We will need this to include the distro's name in the job we pass to the
worker, for instance.
On architectures that require EFI, we must create the ESP partition
and use a GPT partition table. We must also install either the UEFI
or the legacy version of GRUB2 in the image.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Move to the new options format, allowing more flexible partition
tables. The pipeline changes, but the result should be the same.
This requires a yet-to-be-released version of osbuild.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Allow bootloader specific packages to be defined per architecture,
and allow repositories to depend on the architecture.
This does not altert he pipelines we produce, part from the ami
image now contains the grub2-pc package, rather than the grub2
package. This should make no difference.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Make sure we catch expected errors and fail gracefully. Also, make
sure the output id is printed on successufl osbuild run so it can
be introspected from outside the test suite.
Handle the case where osbuild suceeds but does not return an
output_id.
Signed-off-by: Tom Gundersen <teg@jklm.no>
These are used to verify that our pipeline generation is stable, and
that the piplines can generate images that boot.
Signed-off-by: Tom Gundersen <teg@jklm.no>
We must make sure not to depend on the modules loaded in the host
kernel. This makes the initrd the same between different hosts.
Signed-off-by: Tom Gundersen <teg@jklm.no>
When we get an ssh connection, before the image is fully booted,
systemctl is-system-started returnse "starting", treat this as a
failed connection, and keep retrying.
Some distros may not support the wait switch correctly.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Image generation is very much a batch job, latency is not important
and because it is so CPU and IO intensive, it really affects the
overall responsiveness of the system if left on the default settings.
Set scheduling to idle/batched for the worker. The composer service
is still left in the normal scheduling classes as its latency is
important for the UI.
Signed-off-by: Tom Gundersen <teg@jklm.no>
This contains new API for getting multi-arch to work. We expect more additions
before the next release, but this way we can get things working in parallel.
Signed-off-by: Tom Gundersen <teg@jklm.no>
