Bug 432870
Summary: | telinit 3 does not shut down X | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Bill Nottingham <notting> |
Component: | event-compat-sysv | Assignee: | Casey Dahlin <cdahlin> |
Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | low | Docs Contact: | |
Priority: | low | ||
Version: | rawhide | CC: | notting, rvokal, scott, vanhoof |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2008-02-26 23:03:47 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | |||
Bug Blocks: | 431106 |
Description
Bill Nottingham
2008-02-14 20:55:32 UTC
SSIA? Subject Says It All :) Ah. This one is slightly interesting, are you running X from an event.d script? Could you attach that if you are. # prefdm - preferred display manager # # Starts gdm/xdm/etc by preference start on stopped rc5 stop on runlevel [!5] console output script exec /etc/X11/prefdm end script Does "stop prefdm" stop X? No. Well, it never returns. So it could be 'failing to stop it correctly.' Could you paste the output from that? Mine does return. I just get a "stop: Job not changed: prefdm" Random question, does prefdm stay in the foreground while X is running or does it go into the background? prefdm execs whatever the configured display manager is. As for the hang I was seeing, SELinux is preventing it from working; without SELinux I see the 'job not changed' that Casey does. I have a quick solution to this. Edit /etc/X11/prefdm and anywhere it runs /usr/sbin/gdm add the --no-daemon parameter. Haven't tested this issue with kdm/xdm/et cetera. gdm itself forks automatically, that's where the issue lies. Scott has a somewhat controversial but ultimately effective way of getting around forking daemons via ptrace (It will be interesting to see how it fares given that gdm forks the holy hell out of itself). It may be worth trying to get in early. There may also be a more upstarty way of replacing prefdm entirely with an event. KDM accepts a -nodaemon switch, it also accepts --nodaemon, but not --no-daemon. I don't think ptrace hackery is a good idea. That's more upstream's decision. I do think its probably the only available solution to the fork issue long term, and its done in a fairly benign way, so I intend to support them in it. If we set up daemons that don't fork, then it won't affect us at all. Whooops, this is simpler than we made it out to be. The old inittab event is: x:5:respawn:/etc/X11/prefdm -nodaemon Ergo, the upstart event should be -nodaemon as well - this should just be changed in the upstart event, not /etc/X11/prefdm. Package building now. Closed. |