Created attachment 12369 [details] journalctl -b Mageia 8 x86_64 Plasma, classic ISO. Online repositories and 32bits one activated for the following. After wine and lutris installed, I see in system journal: $ systemctl status wine.service ● wine.service - LSB: Allow users to run Windows(tm) applications Loaded: loaded (/etc/rc.d/init.d/wine; generated) Active: failed (Result: exit-code) since Tue 2021-02-23 15:17:19 CET; 2min 55s ago Docs: man:systemd-sysv-generator(8) Process: 1130 ExecStart=/etc/rc.d/init.d/wine start (code=exited, status=1/FAILURE) CPU: 10ms févr. 23 15:17:19 mageia.local wine[1130]: Lancement de wine : [ÉCHEC ] févr. 23 15:17:19 mageia.local wine[1130]: Enregistrement de Wine déjà activé ! févr. 23 15:17:19 mageia.local systemd[1]: Starting LSB: Allow users to run Windows(tm) applications... févr. 23 15:17:19 mageia.local systemd[1]: wine.service: Control process exited, code=exited, status=1/FAILURE févr. 23 15:17:19 mageia.local systemd[1]: wine.service: Failed with result 'exit-code'. févr. 23 15:17:19 mageia.local systemd[1]: Failed to start LSB: Allow users to run Windows(tm) applications.
Assigned to the package maintainer. (Please set the status to 'assigned' if you are working on it)
Source RPM: (none) => wine-6.0-3.mga8.src.rpmAssignee: bugsquad => rverschelde
I think this faild because /proc/sys/fs/binfmt_misc/windows already exists because of: proc-sys-fs-binfmt_misc.automount and qemu installed on this system for aarch build ISO tests... So the wine initscript should fail gracefully instead of report an error.
The initscript file, /etc/rc.d/init.d/wine should be removed from the package as registering the binary format is the sole purpose of the script duplicating the effect of /usr/lib/binfmt.d/wine.conf from the same package. Lowering the priority as the impact is cosmetic. Workaround is to run as root "chkconfig wine off" for those who do not want systemd status showing degraded as it's status. For now, I recommend removing the file /etc/rc.d/init.d/wine from the package in svn and waiting for the next wine update for it to be distributed.
Priority: Normal => LowCC: (none) => davidwhodgins
I use the solution posted at the end of Alex Gromnitsky's blog https://gromnitsky.blogspot.com/2017/03/how-to-prevent-wine-from-auto-running.html : # ln -s /dev/null /etc/binfmt.d/wine.conf # systemctl restart systemd-binfmt This seems to avoid all issues with wine service: # systemctl status wine.service ● wine.service - LSB: Allow users to run Windows(tm) applications Loaded: loaded (/etc/rc.d/init.d/wine; generated) Active: active (exited) since Wed 2021-06-30 12:51:39 EEST; 12h ago Docs: man:systemd-sysv-generator(8) Tasks: 0 (limit: 19093) Memory: 0B CPU: 0 CGroup: /system.slice/wine.service Jun 30 12:51:39 ryzen-ux systemd[1]: Starting LSB: Allow users to run Windows(tm) applications... Jun 30 12:51:39 ryzen-ux wine[2912]: Starting wine: [ OK ] Jun 30 12:51:39 ryzen-ux systemd[1]: Started LSB: Allow users to run Windows(tm) applications. # systemctl status systemd-binfmt.service ● systemd-binfmt.service - Set Up Additional Binary Formats Loaded: loaded (/usr/lib/systemd/system/systemd-binfmt.service; static) Active: active (exited) since Wed 2021-06-30 12:51:37 EEST; 12h ago Docs: man:systemd-binfmt.service(8) man:binfmt.d(5) https://www.kernel.org/doc/html/latest/admin-guide/binfmt-misc.html https://www.freedesktop.org/wiki/Software/systemd/APIFileSystems Main PID: 684 (code=exited, status=0/SUCCESS) Tasks: 0 (limit: 19093) Memory: 0B CPU: 0 CGroup: /system.slice/systemd-binfmt.service Regards.
CC: (none) => arusanu
The problem is caused by the wine package using both an init script and systemd-binfmt to register the the binary formats. One or the other should be used, not both. Note that systemd-binfmt is used by packages other then wine, so I think it's preferable to remove the init script.
(In reply to Dave Hodgins from comment #5) > The problem is caused by the wine package using both an init script and > systemd-binfmt to register the the binary formats. One or the other should > be used, not both. > > Note that systemd-binfmt is used by packages other then wine, so I think it's > preferable to remove the init script. Agree.
Priority: Low => NormalSummary: wine.service failed at boot time, results systemd status degraded => Wine package should not have init.d script and a systemd-binfmt: wine.service failed at boot timeStatus comment: (none) => Potential fix in Comment 5
Not directly related to this issue, but for the record: I picked up maintainership of the wine package when damsweb retired IIRC, but I've never been its most active maintainer (that's more tv). I re-assigned the wine package to `nobody` as I'm not active enough to fulfill the needs for this package. I might still poke at some bugs as time permits but can't make any promise. Re-setting assignee to pkg-bugs@, but feel free to snatch it up Thierry if you're up for it :)
CC: (none) => thierry.vignaudAssignee: rverschelde => pkg-bugs
CC: (none) => rverschelde
*** Bug 28570 has been marked as a duplicate of this bug. ***
CC: (none) => pealfa
Also agree with Dave. I guess care needs to be taken to have the updated package run 'chkconfig --del wine' so as not to leave old broken symlinks around, probably in a %triggerun scriplet.
*** Bug 29343 has been marked as a duplicate of this bug. ***
CC: (none) => egc
*** Bug 29430 has been marked as a duplicate of this bug. ***
CC: (none) => mageia
(In reply to David Walser from comment #9) > Also agree with Dave. I guess care needs to be taken to have the updated > package run 'chkconfig --del wine' so as not to leave old broken symlinks > around, probably in a %triggerun scriplet. I think this part was ignored in the currently "fixed" wine packages...
I think it's not needed because the package removes the init script in it's %preun script. So by dropping %post and %preun scripts in new version, the old bugged wine version removes the init script by itself. I upgraded to local build version and didn't get a problem, tell me if you have trailing wine stuff in /etc/rc?.d/*wine ? Best regards
For me the only remaining action is to pass the package from updates_testing to updates (and issue errata ?).
This script from previous version should handle init script removal : %preun -n %{wine} %_preun_service %{name}
We still need a list of the srpm/rpm packages, an advisory stating what the purpose of the update is, and the bug assigned to qa before it can be validated for moving from updates testing to updates.
I have uploaded a fixed package for Mageia 8. You can test this by installing and see that it don't have a failing sysv init script anymore. Suggested advisory: ======================== Updated wine packages fix failing sysv init script: Wine sysv init fail since registering binfmt windows and windowsPE is already done by systemd-binfmt.service with /usr/lib/binfmt.d/wine.conf config rile. This update just drop script and triggers, old package will drop the service itself. References: https://www.mageialinux-online.org/forum/topic-28733+wine.php ======================== Updated packages in core/updates_testing: ======================== In armv7hl and i586 arch: wine-6.0-4.1.mga8 wine-devel-6.0-4.1.mga8 wine32-6.0-4.1.mga8 In aarch64 and x86_64 arch: wine64-6.0-4.1.mga8 wine64-devel-6.0-4.1.mga8 Source RPMs: wine-6.0-4.1.mga8.src.rpm
Assignee: pkg-bugs => qa-bugs
MGA8-64 Plasma on Lenovo B50 This laptop did not have wine installed before. No installation issues, i.e. no feedback on any problem. Run winefile OK, so OK for me.
CC: (none) => herman.viaeneWhiteboard: (none) => MGA8-64-OK
Before installing the update ... [root@x3 ~]# chkconfig 2>/dev/null|grep wine wine 0:off 1:off 2:on 3:on 4:on 5:on 6:off [root@x3 ~]# ls -l /etc/rc.d/rc3.d/S99wine lrwxrwxrwx 1 root root 14 Sep 13 14:19 /etc/rc.d/rc3.d/S99wine -> ../init.d/wine* After installing the update ... [root@x3 ~]# chkconfig 2>/dev/null|grep wine [root@x3 ~]# ls -l /etc/rc.d/rc3.d/S99wine ls: cannot access '/etc/rc.d/rc3.d/S99wine': No such file or directory Validating the update.
CC: (none) => sysadmin-bugsKeywords: (none) => validated_update
Keywords: (none) => advisory
An update for this issue has been pushed to the Mageia Updates repository. https://advisories.mageia.org/MGAA-2021-0185.html
Status: NEW => RESOLVEDResolution: (none) => FIXED