Bug 642347

Summary:	Shutdown leaves machine on
Product:	[Fedora] Fedora	Reporter:	Horst H. von Brand <vonbrand>
Component:	systemd	Assignee:	Kernel Maintainer List <kernel-maint>
Status:	CLOSED RAWHIDE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	low
Version:	rawhide	CC:	bugzilla, dougsland, gansalmon, itamar, jcm, jonathan, kernel-maint, lpoetter, madhu.chinakonda, metherid, mschmidt, notting, plautrba
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:	systemd-13-1	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2010-11-20 00:43:30 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:

Description Horst H. von Brand 2010-10-12 16:32:00 UTC

Description of problem:
Using shutdown from the menus or even calling poweroff shuts all the system down (it seems it doesn't unmount some filesystems at the end), but the machine continues on. Have to power off by hand.

Version-Release number of selected component (if applicable):
kernel-2.6.36-0.35.rc7.git0.fc15.x86_64

How reproducible:
Always

Steps to Reproduce:
1. Shut down...
2.
3.
  
Actual results:
System shuts down, machine isn't powered off

Expected results:
Poweroff

Additional info:
Has happened on several notebooks (two Toshiba, a Samsung netbook). Several last kernels are affected.

Comment 1 Horst H. von Brand 2010-10-20 14:35:53 UTC

Still the same. Maybe systemd gets stuck, not the kernel's fault at all?

Comment 2 Horst H. von Brand 2010-10-25 16:42:56 UTC

kernel-2.6.36-1.fc15.x86_64
systemd-11-1.fc15.x86_64

Still the same. I saw a final "can't write bytes" scroll by before the system shut down.

Comment 3 Horst H. von Brand 2010-10-25 17:01:10 UTC

Under /sbin "can not write bytes" appears only in plymouthd, here from plymouth-0.8.4-0.20100821.1.fc15.x86_64. It does appear then that this is really systemd's fault (it is called e.g. from /lib/systemd/system/plymouth-halt.service)... perhaps a filesystem is unmounted too soon?

Changing to systemd, here systemd-11-1.fc15.x86_64

Comment 4 Lennart Poettering 2010-10-26 17:55:13 UTC

Hmm, maybe you are using "halt" instead of "poweroff"?

In systemd "halt" actually does what the name suggests: it halts the machine. You need to use "poweroff" to actually poweroff.

Comment 5 Horst H. von Brand 2010-10-27 19:14:22 UTC

Nope, I'm using poweroff (by hand and via System --> Shut down --> Poweroff).

Strangely this happens with all my machines here, all x86_64 SMP (Toshiba notebook, Samsung netbook, and a no-name oldish dual-core AMD64 box (but that one has an idiotic motherboard which acts up and has to be reset by the big button sometimes as it doesn't shut off or start up properly, so I won't count that one in fully)). Because of the above I suspect this is more widespread...

Comment 6 Horst H. von Brand 2010-11-02 13:27:54 UTC

Perhaps this is due to having /usr on a separate partition (really LVM2 logical volume)?

Comment 7 Horst H. von Brand 2010-11-15 14:17:49 UTC

Now systemd-11-2.fc15.x86_64, still the same.

Any way to narrow this down further?

Comment 8 Horst H. von Brand 2010-11-18 18:42:33 UTC

Lately, the Plymouth splash screen stays on (it used to be replaced/overwritten with black-on-black text), the machine reacts to CapsLock and alt-F?, but that's it.

I now see a "plymouthd: could not log bootup: Address already in use" on tty1 (from boot?)

Comment 9 Lennart Poettering 2010-11-18 23:39:11 UTC

Could you retry with v13? This is most likely because of the split /usr. I have added some fixes to deal with that properly.

Comment 10 Horst H. von Brand 2010-11-20 00:43:30 UTC

Just saw this comment, sorry.

Yes, systemd-13-1 fixes this. Thanks!

Comment 11 Lennart Poettering 2010-11-21 14:53:24 UTC

Closing then.