Source the unchanged selector with explicit partition and udev fixtures.
Bound awk discovery to the fixture so runner filesystem size cannot delay
selection. Keep rendered-installer coverage and independent CI failures.
The image caller tests create the run directory in their fixture
image, so they did not see a bind mount to a missing mount point.
Run rh/genimage on a new and on an existing image directory. The mount
fixture fails like mount(8) when the mount point is missing, and the
package manager fixture records the mounts in place when the packages
install. Cover EL9 through the rh and rocky entry points, and EL6,
which does not mount /proc and /sys. The pull-request workflow runs
the test.
mount_chroot bind-mounts the host /run into the image root before the
run directory exists there. On the first build of a new image, mount
fails with "mount point does not exist". genimage does not check the
result, so the packages install without the host /run.
Create the directory first, as mount_chroot already does for /proc,
/sys and /dev. The mount also uses the subroutine argument instead of
the global image path. Both have the same value at every call site.
A statelite import must ask for the lock of its image directory. While
another command holds the lock, the import must fail, report the lock,
leave the tree in the bundle, and keep the image definition.
imgimport took the image lock only when the bundle had a rootimg. A
statelite bundle has a rootimgtree instead, so its import extracted the
tree while genimage, packimage, liteimg or rmimage could change the same
image. imgimport now locks the directory of the first of the rootimg,
rootimgtree, kernel and ramdisk that the bundle has.
Check that the image definition stays when another host holds the
publication lock, or when the file system refuses the publication.
Check the same when a table of the definition opens only with
autocommit. Check that the definition matches the published image when
an extra file fails to copy after the publication.
imgimport wrote the image definition before it published the root
image. When the publication failed, as when another host held the
publication lock, the import failed with the new definition and the
previous files.
imgimport now places the root image, kernel and ramdisk first, then
writes the image definition, then places the other files. A later
failure leaves the definition that matches the published boot files,
as before. Before it writes anything, imgimport also opens the tables
of the image definition without autocommit, as it writes them. A table
that it cannot open then leaves the files as they were.
The litefile lookup test called make_files, which places the litefile
of the bundle before it reads it. The test now calls place_image, the
entry point of imgimport, so it does not depend on the step that places
the litefile.
imgimport moved the boot files and kits of a bundle without a check,
and copied the extra files after it had already written the image. A
bundle without its kernel or a kit then gave a broken image without
an error. A missing extra file failed the import after it published
the root image and wrote the image definition.
imgimport now checks the boot files, kits, extra files and raw image
files that the manifest names. It does this before it writes anything,
and reports each missing file.
imgimport -n checked its nodes when it set their postscripts. By then
it had published the root image, placed the other files and written the
image definition. An undefined node then left a half-imported image.
imgimport now checks the nodes before it writes anything, and fails
with the same error.
List the native cases selected by CI, their dependencies and their isolation requirements. Distinguish rendered-artifact execution from operating-system and firmware boot qualification.
Render Subiquity configuration with real database rows and execute the emitted commands in private namespaces. Check resulting files, failure ordering and the install-monitor handshake without rewriting shell fragments.
Build a real package and repository, then execute the generated helper with representative host architectures and releases. Check its apt configuration and preserve unrelated sources.
Run each selected postinstall script in an isolated filesystem. Check SELinux settings and fstab after repeated execution instead of comparing shell source.
Render the selected profiles with an isolated database. Check the installer data and run the emitted post-install blocks against a private filesystem, including the EFI fallback and missing-loader paths.
Run the native case from the source snapshot preserved before CI package builds.
kvm_uefi_firmware.t loaded the whole kvm plugin behind stubs and set
its package variables. It now loads only xCAT::KVMUtils and keeps
each assertion: the UEFI and SeaBIOS <os> element, the disk bus and
optical drive prefix, the SMBIOS block across a reconfigure, the
refusal of an unknown firmware value, and the vm.othersettings read
that creates no vm row. It also covers the clock offset, memory and
CPU count that reconfigure_domain writes.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
kvm.pm made $parser, $node, $confdata and $updatetable package
variables so that kvm_uefi_firmware.t could set them and call
build_xmldesc and reconfigvm. That exposes plugin state to make it
testable, which xCAT::KVMUtils replaced for the other guest decisions.
xCAT::KVMUtils now holds vm_othersettings, requested_firmware,
firmware_error, guest_os, clock_offset and reconfigure_domain. Each one
takes its inputs as arguments. guest_os replaces build_oshash and the
<os> edits in build_xmldesc. reconfigure_domain replaces the XML
rewrite in reconfigvm and loads XML::LibXML when it is called. kvm.pm
calls these subs, and the four variables are file lexicals again.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
buildall now calls prepare_mock_resultdirs before the cache checks, so a
cached build also loads the mock configuration through mockbuild.config.
release_repository.t builds xCAT-release in a user namespace on a host
without mock, and the build stopped with "mock is not installed: cannot
import mockbuild.config".
The namespace fixture already stands in for /etc/mock. It now also puts a
stand-in mockbuild package on PYTHONPATH. Its load_config fails when the
chroot configuration is missing, and reports the caller as chrootuid, as
mock does when no chrootuid is set. buildrpms.pl is unchanged.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
Daniel Hilst
Vinícius Ferrão