Comments
Oh, hah, I know what it is. It's a bug that I fixed already, but the fix is in Cockpit 356, and that never made it into F42. It's stuck at 355. The 356 update didn't get stable before it was obsoleted by the 357 update which is still in testing. Meanwhile, PackageKit 1.3.4 landed in F42, so now F42 is affected by the bug.
Basically, we need to get the Cockpit 357 update for F42 stable ASAP. I'll karma it and try and get other folks to do so.
This breaks Cockpit. It'll be broken until 357 or later goes stable.
Ah, this seems to be a Cockpit issue, nothing to do with this update. Not sure why it didn't show up before.
The failure here is kinda odd: it seems to cause cockpit to error out when/after installing the update. It's consistent and reproducible, at least inside openQA. I'll try to reproduce manually so I can use the browser console to figure out what's going on.
If the changes are 'correct' and FreeIPA needs to adjust to them, then this needs to be a multi-package update which includes builds of freeipa and any other affected package, that adapt to the changes.
The openQA failures are nothing to do with anything commented above, but because the tomcat package is missing /usr/bin/tomcat:
[adamw@toolbx fedora-toolbox-43 tmp]$ rpmdiff tomcat-10.1.46-3.fc44.noarch.rpm tomcat-10.1.52-2.fc45.noarch.rpm | grep "^removed */usr/bin"
removed /usr/bin/tomcat
removed /usr/bin/tomcat-digest
removed /usr/bin/tomcat-tool-wrapper
It actually removed a ton more files than just that:
[adamw@toolbx fedora-toolbox-43 tmp]$ rpmdiff tomcat-10.1.46-3.fc44.noarch.rpm tomcat-10.1.52-2.fc45.noarch.rpm | grep "^removed */"
removed /etc/logrotate.d/tomcat.disabled
removed /etc/sysconfig/tomcat
removed /etc/tomcat
removed /etc/tomcat/conf.d/module-start-up-parameters.conf
removed /etc/tomcat/jaspic-providers.xsd
removed /etc/tomcat/tomcat-users.xsd
removed /usr/bin/tomcat
removed /usr/bin/tomcat-digest
removed /usr/bin/tomcat-tool-wrapper
removed /usr/lib/systemd/system/tomcat@.service
removed /usr/libexec/tomcat
removed /usr/libexec/tomcat/functions
removed /usr/libexec/tomcat/preamble
removed /usr/libexec/tomcat/server
removed /usr/share/doc/tomcat
removed /usr/share/doc/tomcat/LICENSE
removed /usr/share/doc/tomcat/NOTICE
removed /usr/share/doc/tomcat/RELEASE-NOTES
removed /usr/share/tomcat
removed /usr/share/tomcat/bin/bootstrap.jar
removed /usr/share/tomcat/bin/catalina-tasks.xml
removed /usr/share/tomcat/conf
removed /usr/share/tomcat/lib
removed /usr/share/tomcat/logs
removed /usr/share/tomcat/temp
removed /usr/share/tomcat/webapps
removed /usr/share/tomcat/work
removed /var/cache/tomcat/temp
removed /var/cache/tomcat/work
removed /var/lib/tomcat
removed /var/lib/tomcats
It looks like this update re-introduces the big pile of changes that got tried and then reverted a while back, but without a very clear explanation, and autochangelog choked on the huge list of commits or something.
This seems to be broken because /usr/bin/pki-tomcat-start isn't there:
Mar 09 11:17:18 ipa002.test.openqa.fedoraproject.org systemd[1]: Starting pki-tomcatd@pki-tomcat.service - PKI Tomcat Server pki-tomcat...
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org systemd[1]: Started pki-tomcatd@pki-tomcat.service - PKI Tomcat Server pki-tomcat.
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=pki-tomcatd@pki-tomcat comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org (pki-tomcat-start)[7020]: pki-tomcatd@pki-tomcat.service: Unable to locate executable '/usr/bin/pki-tomcat-start': No such file or directory
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org (pki-tomcat-start)[7020]: pki-tomcatd@pki-tomcat.service: Failed at step EXEC spawning /usr/bin/pki-tomcat-start: No such file or directory
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org systemd[1]: pki-tomcatd@pki-tomcat.service: Main process exited, code=exited, status=203/EXEC
Mar 09 11:17:19 ipa002.test.openqa.fedoraproject.org systemd[1]: pki-tomcatd@pki-tomcat.service: Failed with result 'exit-code'.
Missing a rebuild of polymake:
Dependencies of other packages that would be BROKEN by the tested packages:
package: polymake-4.15-5.fc44.x86_64 from https://kojipkgs.fedoraproject.org/repos/f44-build/latest/x86_64
perl-interpreter = 4:5.42.0
Several things should have been rebuilt with this:
Dependencies of other packages that would be BROKEN by the tested packages:
package: perl-Devel-Cover-1.52-1.fc45.x86_64 from https://kojipkgs.fedoraproject.org/repos/f45-build/latest/x86_64
perl(:VERSION) = 5.42.0
package: perl-PAR-Packer-1.064-3.fc44.x86_64 from https://kojipkgs.fedoraproject.org/repos/f45-build/latest/x86_64
perl(:VERSION) = 5.42.0
package: polymake-4.15-6.fc45.x86_64 from https://kojipkgs.fedoraproject.org/repos/f45-build/latest/x86_64
perl-interpreter = 4:5.42.0
adamwill
commented & provided feedback
on american-fuzzy-lop-4.35c-5.20260206gitafbcb07.fc44, …
4 days ago
@tuliom ah ok, thanks. That's technically a bug in rmdepcheck, then, but it might be a bit difficult to have it check retirements...
Fine here too.
If I'm reading the PR discussion right, the bug may now have been fixed by FEDORA-2026-cc0ddcfb97 , so I've re-submitted this and re-triggered the tests.
Well, Rawhide testing doesn't show the same, and this was in Beta-1.2 which just got signed off, so we need to push it.
Ah, yeah, I think I see the problem. --enable-vt-switch was added between --wayland-display and constants.WAYLAND_SOCKET_NAME. It needs to be before --wayland-display or after constants.WAYLAND_SOCKET_NAME. I guess I'll send a fix.
anaconda still fails to start up. From the logs I suspect the patch to use the new argument was implemented wrongly:
Mar 05 21:56:05 fedora gnome-kiosk[2377]: Using Wayland display name '--enable-vt-switch'
Mar 05 21:56:05 fedora systemd[1]: Starting systemd-localed.service - Locale Service...
Mar 05 21:56:05 fedora systemd[1]: Started systemd-localed.service - Locale Service.
Mar 05 21:56:05 fedora run-in-new-session[2377]: Failed to run the command: Failed to execute child process “wl-sysinstall-0” (No such file or directory)
Mar 05 21:56:06 fedora anaconda[2104]: anaconda: display: Wayland startup failed: /usr/libexec/anaconda/run-in-new-session exited with status 1
note especially gnome-kiosk[2377]: Using Wayland display name '--enable-vt-switch', that looks wrong.
This causes rtkit-daemon.service to fail repeatedly:
Mar 05 12:11:37 fedora systemd[1]: Started rtkit-daemon.service - RealtimeKit Scheduling Policy Service.
Mar 05 12:11:37 fedora rtkit-daemon[1246]: Demoting known real-time threads.
Mar 05 12:11:37 fedora rtkit-daemon[1246]: Demoted 0 threads.
Mar 05 12:11:37 fedora audit[1246]: AVC avc: denied { write } for pid=1246 comm="rtkit-daemon" path="/run/systemd/inhibit/5.ref" dev="tmpfs" ino=2482 scontext=system_u:system_r:rtkit_daemon_t:s0 tcontext=system_u:object_r:systemd_logind_inhibit_var_run_t:s0 tclass=fifo_file permissive=0
Mar 05 12:11:37 fedora audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=rtkit-daemon comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 05 12:11:37 fedora systemd[1]: rtkit-daemon.service: Deactivated successfully.
Mar 05 12:11:37 fedora audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=rtkit-daemon comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 05 12:11:37 fedora systemd[1]: Starting rtkit-daemon.service - RealtimeKit Scheduling Policy Service...
Mar 05 12:11:37 fedora audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=rtkit-daemon comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 05 12:11:37 fedora systemd[1]: Started rtkit-daemon.service - RealtimeKit Scheduling Policy Service.
Mar 05 12:11:37 fedora audit[1258]: AVC avc: denied { write } for pid=1258 comm="rtkit-daemon" path="/run/systemd/inhibit/6.ref" dev="tmpfs" ino=2486 scontext=system_u:system_r:rtkit_daemon_t:s0 tcontext=system_u:object_r:systemd_logind_inhibit_var_run_t:s0 tclass=fifo_file permissive=0
Mar 05 12:11:37 fedora rtkit-daemon[1258]: Demoting known real-time threads.
Mar 05 12:11:37 fedora rtkit-daemon[1258]: Demoted 0 threads.
Mar 05 12:11:37 fedora wireplumber[1253]: mod.rt: RTKit error: org.freedesktop.DBus.Error.NoReply
Mar 05 12:11:37 fedora wireplumber[1253]: mod.rt: RTKit does not give us MaxRealtimePriority, using 1
Mar 05 12:11:37 fedora pipewire[1252]: mod.rt: RTKit error: org.freedesktop.DBus.Error.NoReply
Mar 05 12:11:37 fedora pipewire[1252]: mod.rt: RTKit does not give us MaxRealtimePriority, using 1
Mar 05 12:11:37 fedora audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=rtkit-daemon comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Mar 05 12:11:37 fedora systemd[1]: rtkit-daemon.service: Deactivated successfully.
This needs to go together with the gnome-kiosk update. They cannot go separately.
- If gnome-kiosk goes first, VT switch from anaconda is broken.
- If anaconda goes first, anconda is broken because gnome-kiosk doesn't have the option it's trying to use.
Because this update would cause "anaconda goes first", the tests are failing, anaconda on dedicated installer images doesn't start:
Mar 05 15:27:47 fedora run-in-new-session[2380]: Failed to run the command: Failed to execute child process “wl-sysinstall-0” (No such file or directory)
Mar 05 15:27:48 fedora anaconda[2109]: anaconda: display: Wayland startup failed: /usr/libexec/anaconda/run-in-new-session exited with status 1
I assume that really just means "oops, gnome-kiosk blew up because we passed it an invalid argument".
Yeah, VPN fix looks good here, no other issues noted.
This seems fine in Beta-1.2 testing so far.
We need this to go stable ASAP to fix this known bug, now PackageKit 1.3.4 landed in stable. Please karma this.