Bug 1160378

Summary: systemd fails to start systemd-logind.service in rescue system
Product: [openSUSE] openSUSE Tumbleweed Reporter: Steffen Winterfeldt <snwint>
Component: InstallationAssignee: Steffen Winterfeldt <snwint>
Status: RESOLVED FIXED QA Contact: Jiri Srain <jsrain>
Severity: Normal    
Priority: P5 - None CC: fbui, jslaby, ub1x
Version: Current   
Target Milestone: ---   
Hardware: Other   
OS: Other   
URL: https://trello.com/c/PJ4RAk9f
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Steffen Winterfeldt 2020-01-08 09:29:45 UTC
It says:


0:vm2302:~ # systemctl status systemd-logind
● systemd-logind.service - Login Service
   Loaded: loaded (/mounts/mp_0000/usr/lib/systemd/system/systemd-logind.service; static; vendor preset: disabled)
  Drop-In: /mounts/mp_0000/usr/lib/systemd/system/systemd-logind.service.d
           └─nosandbox.conf
   Active: failed (Result: exit-code) since Wed 2020-01-08 09:16:23 UTC; 8min ago
     Docs: man:systemd-logind.service(8)
           man:logind.conf(5)
           https://www.freedesktop.org/wiki/Software/systemd/logind
           https://www.freedesktop.org/wiki/Software/systemd/multiseat
  Process: 2831 ExecStartPre=/sbin/modprobe -abq drm (code=exited, status=226/NAMESPACE)
  Process: 2832 ExecStart=/usr/lib/systemd/systemd-logind (code=exited, status=226/NAMESPACE)
 Main PID: 2832 (code=exited, status=226/NAMESPACE)

Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Main process exited, code=exited, status=226/NAMESPACE
Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Failed with result 'exit-code'.
Jan 08 09:16:23 vm2302 systemd[1]: Failed to start Login Service.
Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Service has no hold-off time (RestartSec=0), scheduling >
Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Scheduled restart job, restart counter is at 5.
Jan 08 09:16:23 vm2302 systemd[1]: Stopped Login Service.
Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Start request repeated too quickly.
Jan 08 09:16:23 vm2302 systemd[1]: systemd-logind.service: Failed with result 'exit-code'.
Jan 08 09:16:23 vm2302 systemd[1]: Failed to start Login Service.
Comment 1 Steffen Winterfeldt 2020-01-08 09:34:04 UTC
Franck, I'm sure this means something but I don't remember. You probably
know what's missing.
Comment 2 Franck Bui 2020-01-08 14:22:52 UTC
Hmm I can't remember anything related.

Something went wrong with your environment when PID1 tried to restrict the runtime environment of logind.

Maybe you can try to strace PID1 and see if you can spot which syscall failed ?
Comment 3 Steffen Winterfeldt 2020-01-08 14:28:40 UTC
I was thinking in the direction of bug 937237.
Comment 4 Franck Bui 2020-01-13 10:57:28 UTC
Can you provide a link so I can try to reproduce it locally ?
Comment 5 Steffen Winterfeldt 2020-01-13 11:07:29 UTC
Any (not too old) TW image will do.
Comment 6 Franck Bui 2020-01-13 14:50:37 UTC
Now that reminds me something related to the way '/' is mounted by the rescue system... I vaguely recall one service (maybe udev ?) failing to start because of that.
Comment 7 Franck Bui 2020-01-13 15:03:18 UTC
Or maybe '/' is not mounted at all and that was an issue with services creating their own mount namespace.
Comment 8 Steffen Winterfeldt 2020-01-13 16:31:58 UTC
Yes, that was what I meant in comment 3. But that was fixed long ago (or
at least worked for some time).

The current issue started some time after 15.1; there were no changes in the
way linuxrc sets up the rescue system.

If you have any idea what exactly is not as expected, that would be great.
Comment 9 Franck Bui 2020-01-13 16:39:02 UTC
The first thing to check is how linuxrc setup the rootfs of the rescue system before it "chroots" in it.
Comment 10 Franck Bui 2020-02-03 13:52:11 UTC
*** Bug 1162472 has been marked as a duplicate of this bug. ***
Comment 11 Steffen Winterfeldt 2020-02-04 09:26:53 UTC
Tracking in YaST Scrum board.
Comment 12 Steffen Winterfeldt 2020-02-07 14:40:21 UTC
I've found the place where the root fs entry gets lost in linuxrc while setting up the rescue system. With that fixed, systemd works fine.
Comment 13 Steffen Winterfeldt 2020-02-07 15:37:31 UTC
so far: https://github.com/openSUSE/linuxrc/pull/212
Comment 14 Franck Bui 2020-02-21 12:54:31 UTC
*** Bug 1163760 has been marked as a duplicate of this bug. ***
Comment 15 Steffen Winterfeldt 2020-03-05 14:19:46 UTC
finally fixed: https://github.com/openSUSE/linuxrc/pull/216