Description of problem: Anaconda's gating tests are failing during the package installation with the following error: "An error occurred during the transaction: Error in PREIN scriptlet in rpm package unbound-anchor." (see [0]) The gating tests are failing since October 11. The last successful run was on September 19. Version-Release number of selected component (if applicable): Probably unbound-anchor-1.17.0-1.fc38.x86_64.rpm How reproducible: Always Steps to Reproduce: # tmt reproducer git clone https://src.fedoraproject.org/forks/packit/rpms/anaconda testcode git -C testcode checkout -b testbranch 5a7836f84c54d146627a664474cde9bc58a1259a cd testcode tmt --context arch=x86_64 --context distro=fedora-38 --context trigger=commit run --all --verbose provision --how virtual --image Fedora-Rawhide prepare --how shell --script ' ' plan --name ^/plans/integration$ Actual results: The PREIN scriptlet exists with a non-zero exit code. Expected results: All scriptlets exit with the zero exit status (see [1]). Additional info: [0] https://artifacts.dev.testing-farm.io/fccee813-5773-4b40-9507-ef1969206499/ [1] https://docs.fedoraproject.org/en-US/packaging-guidelines/Scriptlets/
Hmm, strange. Preinstall just ensures the group and user for unbound exists. There should be no reason for its failure on normal system. preinstall scriptlet (using /bin/sh): getent group unbound >/dev/null || groupadd -r unbound getent passwd unbound >/dev/null || \ useradd -r -g unbound -d /etc/unbound -s /sbin/nologin \ -c "Unbound DNS resolver" unbound
I have issues starting reproduction machine. tmt fails for some reason to spawn rawhide system on my f37 system. Not sure how can I debug what is the actual problem. $ LC_ALL=C.UTF-8 sudo tmt --context arch=x86_64 --context distro=fedora-38 --context trigger=commit run --all --verbose provision --how virtual --image Fedora-Rawhide prepare -vvv --how shell --script ' ' plan --name ^/plans/integration$ /var/tmp/tmt/run-009 Found 1 plan. /plans/integration summary: Integration tests for anaconda discover how: fmf directory: /home/pemensik/rhel/tests/testcode hash: 5a7836f filter: tag: integration summary: 1 test selected /tests/dirinstall/dirinstall/text provision how: virtual user: root image: Fedora-Rawhide memory: 2048 MB disk: 10 GB connection: session arch: x86_64 qcow: Fedora-Cloud-Base-Rawhide-20221129.n.0.x86_64.qcow2 name: tmt-009-SAEhtZih key: /var/tmp/tmt/run-009/plans/integration/provision/default-0/id_ecdsa progress: booting... libvirt: QEMU Driver error : vnitřní chyba: qemu neočekávaně ukončilo monitor: 2022-12-01T15:41:44.935951Z qemu-system-x86_64: -device {"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x2","drive":"libvirt-2-format","id":"virtio-disk0","bootindex":1,"write-cache":"on"}: PCI: slot 2 function 0 not available for virtio-blk-pci, in use by virtio-net-pci,id=(null) Instance startup failed, retrying in 5 seconds... libvirt: QEMU Driver error : vnitřní chyba: qemu neočekávaně ukončilo monitor: 2022-12-01T15:41:50.425676Z qemu-system-x86_64: -device {"driver":"virtio-blk-pci","bus":"pci.0","addr":"0x2","drive":"libvirt-2-format","id":"virtio-disk0","bootindex":1,"write-cache":"on"}: PCI: slot 2 function 0 not available for virtio-blk-pci, in use by virtio-net-pci,id=(null) finish how: shell overview: 1 script found script: command -v journalctl && journalctl -a || true The guest is not available.
Found alternative reproduction, just use Rawhide machine to install anaconda, prepare the same ks.cfg and use: cat > ks.cfg << EOF url --metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-rawhide&arch=x86_64 repo --name=fedora-modular --metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-modular-rawhide&arch=x86_64 # The repository configuration (url, repo) needs to be added here. # It varies by the product and version we are running on / testing lang en_US.UTF-8 keyboard --vckeymap=us --xlayouts='us' rootpw --plaintext redhat timezone --utc Europe/Prague shutdown %packages %end EOF anaconda --dirinstall /var/tmp/dirinstall.UrZKlu/install_dir --kickstart $PWD/ks.cfg --text --noninteractive It reproduces described issue. Problem is I would need more details from dnf transaction. I think any failures should be printed to stderr. Is there perhaps way to request anaconda to add --verbose flag to dnf during install? Is it possible to use kickstart file for that? Is there log with more details somewhere in install_dir?
All the require information should be found in logs in /tmp/. Probably want the dnf.log and dnf.librepo.log. Also a note it started with Fedora 38, on Fedora 37 this worked just fine.
Sorry I meant you want `packaging.log`. Attaching it from the installation.
Created attachment 1929324 [details] packaging.log
Created attachment 1929325 [details] dnf.librepo.log
I see there is an issue with ``` dnf.rpm: groupadd connet open /etc/group ``` Could be an issue with adding groups on Rawhide?
(In reply to Jiri Konecny from comment #6) > Created attachment 1929324 [details] > packaging.log Thanks! found inside: 13:36:15,525 INF dnf.rpm: groupadd: cannot open /etc/group useradd: group 'unbound' does not exist error: %prein(unbound-anchor-1.17.0-1.fc38.x86_64) scriptlet failed, exit status 6 13:36:15,526 ERR dnf.rpm: Error in PREIN scriptlet in rpm package unbound-anchor 13:36:15,573 INF dnf.rpm: error: unbound-anchor-1.17.0-1.fc38.x86_64: install failed warning: user unbound does not exist - using root warning: group unbound does not exist - using root warning: user unbound does not exist - using root warning: group unbound does not exist - using root 13:36:16,419 INF dnf.rpm: Created symlink /etc/systemd/system/graphical.target.wants/udisks2.service → /usr/lib/systemd/system/udisks2.service. Running in chroot, ignoring request. Running in chroot, ignoring request. Could you please verify group file is present soon enough for packages to evaluate? It is part of setup package. Is setup package order enforced to be before any service adding a new user? I am not sure even using systemd-sysusers would allow creating users at this stage. It seems to me much more packages might have issues with this error. I doubt I should have Requires(pre): setup in the unbound-libs package.
Hi, I honestly not sure how this is handled but I don't think Anaconda is in charge or configuring this somehow special. Maybe DNF team could help us here. Jardo, are you able to help us here?
Adding RPM team members to CC to check if this scriptlet failure should be fatal for the transaction or not. Thanks jmracek for the suggestion! :)
(In reply to Martin Kolman from comment #11) > Adding RPM team members to CC to check if this scriptlet failure should be > fatal for the transaction or not. Thanks jmracek for the suggestion! :) I think more important would be adding more details into useradd/groupadd when the adding failed. If it is permission problem, it should clearly state it. Message: groupadd: cannot open /etc/group is not much helpful. Errno of the failure would help a lot, because it is unclear to me what the problem is and how to debug it during the install process.
Maybe I have found a fix. Not sure whether it is correction of ownership of unbound-libs with user created there. Or usage of systemd-sysusers helped instead. Not sure which one. With following kickstart with my copr unbound, installation passes successfully. cat >ks-copr.cfg <<EOF url --metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-rawhide&arch=x86_64 repo --name=fedora-modular --metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-modular-rawhide&arch=x86_64 repo --name=copr-unbound --baseurl=https://download.copr.fedorainfracloud.org/results/pemensik/unbound/fedora-rawhide-x86_64/ # The repository configuration (url, repo) needs to be added here. # It varies by the product and version we are running on / testing lang en_US.UTF-8 keyboard --vckeymap=us --xlayouts='us' rootpw --plaintext redhat timezone --utc Europe/Prague shutdown %packages %end EOF
FEDORA-2023-f1a7ba0307 has been submitted as an update to Fedora 37. https://bodhi.fedoraproject.org/updates/FEDORA-2023-f1a7ba0307
FEDORA-2023-f1a7ba0307 has been pushed to the Fedora 37 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2023-f1a7ba0307` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2023-f1a7ba0307 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
FEDORA-2023-43a9d5fbe6 has been submitted as an update to Fedora 36. https://bodhi.fedoraproject.org/updates/FEDORA-2023-43a9d5fbe6
FEDORA-2023-43a9d5fbe6 has been pushed to the Fedora 36 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2023-43a9d5fbe6` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2023-43a9d5fbe6 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
FEDORA-2023-f1a7ba0307 has been pushed to the Fedora 37 stable repository. If problem still persists, please make note of it in this bug report.
FEDORA-2023-43a9d5fbe6 has been pushed to the Fedora 36 stable repository. If problem still persists, please make note of it in this bug report.