installing additional packages fails if repository is enabled only in the VM

[redbeam]

Currently the method of installing a package in the VM that needs its own repository (like fedora-updates) involves downloading the package to a separate folder and installing it using the install_additional_packages function from createdisk_library.sh, which creates a local repository of packages.

If we need to remove the repository (fedora-updates), there is a problem, since it breaks rpm-ostree, which requires all installed packages to be still resolvable (#1119).

However, install_additional_packages downloads the packages on the host first, and uploads them to the VM, which fails if there are packages that are available in repositories enabled only in the VM.

[redbeam]

One possible way to avoid this bug is to enable the repositories on the host and let install_additional_packages handle the rest.

I've tested this in #1123 and for the microshift bundle, it worked:

Log

+ '[' 1 '!=' 0 ']'
+ sudo dnf -y copr enable mskvarla/9pfs
Updating Subscription Management repositories.
Enabling a Copr repository. Please note that this repository is not part
of the main distribution, and quality may vary.

The Fedora Project does not exercise any power over the contents of
this repository beyond the rules outlined in the Copr FAQ at
<https://docs.pagure.org/copr.copr/user_documentation.html#what-i-can-build-in-copr>,
and packages are not held to any quality or security level.

Please do not file bug reports about these packages in Fedora
Bugzilla. In case of problems, contact the owner of this repository.
Repository successfully enabled.
+ ADDITIONAL_PACKAGES+=' 9pfs'
+ install_additional_packages 192.168.126.11
+ local vm_ip=192.168.126.11
+ shift
+ [[ rhel = \f\e\d\o\r\a\-\c\o\r\e\o\s ]]
++ mktemp -d tmp-rpmXXX
+ local pkgDir=tmp-rpmFJa
+ mkdir -p tmp-rpmFJa/packages
+ sudo yum download --downloadonly --downloaddir tmp-rpmFJa/packages cloud-init gvisor-tap-vsock-gvforwarder hyperv-daemons 9pfs --resolve --alldeps
[...]
+ scp -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc -r tmp-rpmFJa/packages core@192.168.126.11:/home/core/
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc core@192.168.126.11 'sudo bash -x -s'
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
+ podman run --rm -v /home/core/packages:/packages:Z quay.io/centos/centos:stream9 sh -c 'dnf install -y createrepo && createrepo /packages'
[...]
Directory walk started
Directory walk done - 180 packages
Temporary output repo path: /packages/.repodata/
Preparing sqlite DBs
Pool started (with 5 workers)
Pool finished
+ podman rmi quay.io/centos/centos:stream9
Untagged: quay.io/centos/centos:stream9
Deleted: ab9123964b9384bec6e1de2dcc7511d90ca625ca5caf2be727532cf5f43612c1
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc core@192.168.126.11 'sudo bash -c '\''cat > /etc/yum.repos.d/local.repo << EOF
[local]
name=Local repo
baseurl=file:///home/core/packages/
enabled=1
gpgcheck=0
EOF'\'''
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc core@192.168.126.11 -- 'sudo rpm-ostree install cloud-init gvisor-tap-vsock-gvforwarder hyperv-daemons 9pfs'
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
Checking out tree 96fc4fa...done
Enabled rpm-md repositories: local
Importing rpm-md...done
rpm-md repo 'local' (cached); generated: 2025-07-09T13:48:47Z solvables: 180
Resolving dependencies...done
Importing packages...done
Checking out packages...done
Running pre scripts...done
Running post scripts...done
Running posttrans scripts...done
Writing rpmdb...done
Writing OSTree commit...done
Staging deployment...done
Pruned images: 1 (layers: 0)
Added:
   [...]
Changes queued for next boot. Run "systemctl reboot" to start a reboot
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc core@192.168.126.11 -- sudo rm -fr /home/core/packages
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
+ ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i id_ecdsa_crc core@192.168.126.11 -- sudo rm -fr /etc/yum.repos.d/local.repo
Warning: Permanently added '192.168.126.11' (ED25519) to the list of known hosts.
+ rm -fr tmp-rpmFJa

[praveenkumar]

One possible way to avoid this bug is to enable the repositories on the host and let install_additional_packages handle the rest.

even we enable the repo on the host then also downloading the other packages might impact and we hit 718a9c4 one.

I think what we should introduce another variable ADDITIONAL_PACKAGE_NO_DOWNLOAD and packages which needed to download from non-official repo should be part of this variable.

[redbeam]

Fixed in #1125