-
Bug
-
Resolution: Unresolved
-
Critical
-
rhel-9.3.0
-
None
-
Important
-
ZStream
-
rhel-sst-cs-plumbers
-
ssg_core_services
-
26
-
5
-
False
-
-
None
-
Red Hat Enterprise Linux
-
None
-
Approved Blocker
-
None
-
None
-
-
All
-
None
What were you trying to do that didn't work?
STIG and other organizations require implementing StopIdleSessionSec=xxx in /etc/systemd/logind.conf.
With latest RHEL8.9, it appears 2 issues arise when using the feature:
- The session disconnects even not when being idle
Executing anything spawning a pager will end up getting disconnected, e.g. with following the journal and hitting Enter every second, or through browsing a man page and use the arrow keys. - When specifying KillUserProcesses=no to let runaway processes execute in the background, processes are killed anyway, e.g.
[user@vm-logind8 ~]$ nohup sleep 1234 & [1] 1705 nohup: ignoring input and appending output to 'nohup.out' [user@vm-logind8 ~]$ <<< wait for idle timeout Connection to vm-logind8 closed by remote host.
The journal shows sleep got killed:
Feb 06 11:08:54 vm-logind8 systemd-logind[786]: Session "5" of user "user" is idle, stopping. Feb 06 11:08:54 vm-logind8 systemd[1]: session-5.scope: Killing process 1646 (sshd) with signal SIGTERM. Feb 06 11:08:54 vm-logind8 systemd[1]: session-5.scope: Killing process 1649 (sshd) with signal SIGTERM. Feb 06 11:08:54 vm-logind8 systemd[1]: session-5.scope: Killing process 1650 (bash) with signal SIGTERM. Feb 06 11:08:54 vm-logind8 systemd[1]: session-5.scope: Killing process 1705 (sleep) with signal SIGTERM.
This hence tends to indicate that KillUserProcesses=no is just ignored by StopIdleSessionSec=xx.
Please provide the package NVR for which bug is seen:
systemd-239-78.el8.x86_64
How reproducible:
Always, see above description.
Steps to reproduce
- Set StopIdleSessionSec=60 and restart logind
- Connect as a user, start browsing a manpage and use arrow keys to move in the manpage
- After 60 seconds the connection closes automatically despite moving in the manpage