composer doesn't support f34 or f35, but it should be possible to build
these even with the f33 distro definition. Introduce f34 and f35 repos
and aliases for f33.
Composer does not have 1:1 mapping of what can be the Host Distro name
and the names of supported distributions held in the Distroregistry.
The fact that the host distro `Name()` method as passed to the Weldr API
does not return the same name as what is used as distro name for
repository definitions. This makes it hard to use `distro.Distro` and
`distro.Arch` directly and rely on the values returned by them as their
name.
Add `New*HostDistro()` to all distro definitions, accepting the name
that should be returned by the distro's `Name()` method. This is useful
mainly if the host distro is Beta or Stream variant of the distro.
Change the distroregistry.Registry to contain host distro as a separate
value set when creating it using `New()` function. This value is
returned by `Registry.FromHost()` method. Determining the host distro is
handled by the `NewDefault()` function. Move the distro name mangling to
distroregistry package. Add relevant unit tests.
Signed-off-by: Tomas Hozza <thozza@redhat.com>
The `distribution` struct defined in multiple distributions contained
unused `imageTypes` field. Remove it to simplify code.
Signed-off-by: Tomas Hozza <thozza@redhat.com>
imageTypeS2 implements the distro.ImageType interface buts generates a
Manifest matching the new osbuild v2 schema.
Two new image types are added to the rhel84 distro (x84_64 and aarch64)
for generating OCI containers contain an Edge (ostree) commit and, when
run, start a web serer to serve the commit.
The image type uses the new PackageSets map to define packages (and
excludes) for the image. The old methods (Packages() and
BuildPackages()) are implemented for compatibility with the old
workflow.
The image also defines an extra package set for the container that will
serve the package: "httpd" (and its dependencies).
The distro.ImageType interface has a new method: Exports()
It should return a list of names or IDs of artefacts that should be
exported from osbuild when the job is complete.
For the old image types, this is simply set to "assembler".
The same test is run in distro/distro_test.go. The redundancy was probably
caused by a bitrot in several commits.
I decided to remove the test from distro implementations to reduce the amount
of duplicated code.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
This replaces Packages() and BuildPackages() by returning a map of
package sets, the semantics of which is up to the distro to define.
They are meant to be depsolved and the result returned back as a
map to Manifest(), with the same keys.
No functional change.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Explicitly set the kernel to boot into.
Also change the blueprint/kernenl handling:
Rather than only falling back to the default kernel name for
getting the package list, let GetKernel() always return the
correct result so we can rely on this being consistent.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Rather than setting this automagically, expose it to the caller. For
now the only caller we have simply passes it back in, so this is a
noop.
In follow-up commits this will be used to resolve the parent commit.
This is tested by verifying that the generated manifests do not
change.
Signed-off-by: Tom Gundersen <teg@jklm.no>
The kernel now comes from the blueprint packages even when it's not
specified. Removing from the base packages of the image types avoids
duplication and allows for alternative kernels to be specified without
also including the default.
The latter is necessary for RHEL for Edge and Fedora IoT images (ostree
commits) that fail to build when multiple kernels are installed.
ImageType tests modified to fix expected package order.
Test for each distro that runs through all architecture - image type
combinations and calls the Manifest() method with a kernel boot option
customization and checks if the ostree image types produce the expected
error.
Kernel boot parameters have no effect on ostree type images (Fedora IoT
and RHEL for Edge). Catch this and fail early in the pipeline creation
and communicate the issue to the user.
Imagine this situation: You have a RHEL system booted from an image produced
by osbuild-composer. On this system, you want to use osbuild-composer to
create another image of RHEL.
However, there's currently something funny with partitions:
All RHEL images built by osbuild-composer contain a root xfs partition. The
interesting bit is that they all share the same xfs partition UUID. This might
sound like a good thing for reproducibility but it has a quirk.
The issue appears when osbuild runs the qemu assembler: it needs to mount all
partitions of the future image to copy the OS tree into it.
Imagine that osbuild-composer is running on a system booted from an imaged
produced by osbuild-composer. This means that its root xfs partition has this
uuid:
efe8afea-c0a8-45dc-8e6e-499279f6fa5d
When osbuild-composer builds an image on this system, it runs osbuild that
runs the qemu assembler at some point. As I said previously, it will mount
all partitions of the future image. That means that it will also try to
mount the root xfs partition with this uuid:
efe8afea-c0a8-45dc-8e6e-499279f6fa5d
Do you remember this one? Yeah, it's the same one as before. However, the xfs
kernel driver doesn't like that. It contains a global table[1] of all xfs
partitions that forbids to mount 2 xfs partitions with the same uuid.
I mean... uuids are meant to be unique, right?
This commit changes the way we build RHEL 8.4 images: Each one now has a
unique uuid. It's now literally a unique universally unique identifier. haha
[1]: a349e4c659/fs/xfs/xfs_mount.c (L51)
Attempt to clarify the structure of our tests. Each test case is now
encapsulated in a script in `test/cases`. Each of these scripts should
be runnable on a pristine machine and be independent of each other. It
is up to the test-orchestractor to decide if they should be run
consequtively instance, or in parallel on separate instances. Each
script can execute several tests and call whatever helper binaries
is desired. However, each case should be assumed to always run as one.
No tests should be run directly from git, but should rather be installed
onto the test system using rpm and run from there. This moves towards
unifying our two types of test cases.
The new structure of is now:
`test/cmd`: the executors, one for each test-case. This is installed
into `/usr/libexec/test/osbuild-composer`.
`test/data`: data and config used by the tests. This is installed into
`/usr/share/tests/osbuild-composer`.
`schutzbot`: configuration of the actual test run. In particular, this
is where the distros and repositories to test against are
configured.
This is very much still work-in-progress, and is only the first step
towards simplifying schutzbot. Apart from moving files around, this
should be a noop.
Signed-off-by: Tom Gundersen <teg@jklm.no>
By labeling `cp` with `system_u:object_r:install_exec_t:s0` we allow it
to copy labels unknown to the host.
See also corresponding commit in osbuild:
e80130a830
Previously, cloud-init was installed on qcow and openstack images but was not
enabled, preventing a user from logging in to these images. Cloud-init, cloud-final,
cloud-init-local and cloud-config services are enabled to make cloud-init work
and to maintain consistency with official fedora 31 and fedora 32 images.
The image tests have been updated to reflect these changes.
Fixes#698
Although the vhdx format is smaller than raw, AWS often has some issues
booting instances from an imported vhdx snapshot. The raw format seems
to boot consistently at AWS.
Fixes#754.
Signed-off-by: Major Hayden <major@redhat.com>
Rather than getting a set of base packages from the ImageType, and then
appending the requested packages from the blueprint, pass the blueprint
into the new Packages() function, and return the full set of packages to
be depsolved.
This allows us to also append packages based on other customizations
too, and use that to append chrony when the timezone is set. This
matches the behavior anaconda had, and there was a TODO item to do this,
which had been overlooked.
Fixes#787.
Signed-off-by: Tom Gundersen <teg@jklm.no>
How exactly the final file system layout looks like is determined
by the installer (anaconda) and thus can not be known at commit
creation time. Thus creating an /etc/fstab file is unnecessary and
the information in it probably wrong. The file wont be used though
because it will be overwritten during the installation process.
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>
The input options of qemu assembler do not require filesystem for every
partition. This patch makes it optional and omits the object in case it
is empty.
osbuild schema definition:
162797f206/assemblers/org.osbuild.qemu (L63)
Partition type, as defined in osbuild qemu assembler, is of type string
because it is either UUID or identifier. See here for details:
162797f206/assemblers/org.osbuild.qemu (L88)
Change type of the "Type" struct member to match the definition in
osbuild. This patch is needed for ppc64le support.
Rather than Manifest() returning an osbuild.Manifest object, introduce a
new distro.Manifest object which represents it as an opaque, JSON
serializable object. This new type has the following properties:
1) its serialization is compatible with the input to osbuild,
2) any valid osbuild input can be deserialized into it, and
3) marshalling and unmarshaling to and from JSON is lossless.
This means that even as we change the subset of valid osbulid manifests
that we support, we can still load any previous state from disk, and it
will continue to work just as before, even though we can no longer
deserialize it into our internal notion of osbuild.Manifest.
This fixes the underlying problem of which #685 was a symptom.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Rename the `Fedora32` type to simply `distribution`, to avoid the
stutter. Move `New()` to the bottom of the file, which is the only
non-generic part. Also make the linter happy.
No functional change.
Signed-off-by: Tom Gundersen <teg@jklm.no>
We do not properly test, and do not have properly defined use-cases for
the ext4-filesystem, partitioned-disk, nor tar image types. Drop them to
focus on delivering the things we car properly test.
Signed-off-by: Tom Gundersen <teg@jklm.no>
When generating an osbuild manifest for an image type, we take a
customizations struct, which specifies the image-type-independent
customizations to apply. We also take the size argument, which is
specific to the image build and not part of the blueprint.
Introduce a new argument ImageOptions, which for now just wraps the size
argument. These options are specific to the image build/type, and
therefore does not belong with the other customizations.
For now this is a non-functional change, but follow-up commits will
introduce more types of image options.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Prior this commit the ami image type produced raw.xz images. This was bad for
two reasons:
- The upload was broken because AWS doesn't support tar.xz format
- XZ compression is terribly slow
This commit changes the format to vhdx, which is supported by AWS and also
quite quick. See https://github.com/osbuild/osbuild-composer/issues/257
why vhdx was chosen.
Fixes#257
We need to make partition uuids stable to be able to run image-info
tests on images using gpt. For example all aarch64 images use gpt.
Also change the type of the (part) UUID to string because that's what
the other UUIDs use and it is easier to work with.
There is no point in having the grub2 stage in pipelines for image types
that are not bootable. The current version is probably a result of
previous refactoring where the member variable was named `IncludeFSTab`.
Moving the grub2 stage into the conditional branch should also fix test
generation on aarch64.
Finally it is necessary to regenerate test cases for non-bootable image
types.
Dracut is unfortunately very host-dependant by default. The package
dracut-config-generic forces it use a generic configuration instead of a
configuration generated from the host environment.
This change should make the image generation more reproducible. For example
it was not possible to boot ami images built on Travis on AWS prior this
commit.
Also, the tests were re-generated in this commit.
In PR#395 we discussed the spelling of archs vs. arches and we agreed to
use arches. This patch only renames the public method `ListArchs`in the
`Distro` interface.
The main reason here is not to make sure we have the right set of build
packages, but rather to make sure the pointer hierarchy (distro->arch->
image type) works as expected.
This patch uses the same approach as fedora31 few commits ago. It moves
the arch pointer to the imageType struct and sets it while creating the
distro->arch->image type hierarchy in the setImageType function. The
rest of the patch is just renaming.
It uses the same approach as fedora31 few commits ago. The pointer to
distro is moved to "arch" and "arch" now contains a map of allowed
image types for each arch.
Also include setArches and setImageTypes helper functions to ease the
creation of pointers from the structures to a parent structure while
creating the distro->arches->image types hierarchy.
This information is now provided only when an architecture is specified,
so it is necessary to first obtain object implementing the Arch interface
then object implementing the ImageType interface and then you can get
the filename and mime type.
Tests are changed accordingly to the new API.
