You are not logged in.
I'm not entirely sure in what section to post this, please move if there is a better section than this...
After updating networkmanager last night, the shutdown process hangs for about 15 seconds after sending SIGTERM,then reports fail and shutdown continues. A shutdown example from messages.log seems to point to networkmanager/dhcp indeed, but I have no clue what would be wrong:
Jun 15 06:58:15 localhost shutdown[839]: shutting down for system halt
Jun 15 06:58:16 localhost init: Switching to runlevel: 0
Jun 15 06:58:17 localhost bumblebeed[478]: Received Terminated signal.
Jun 15 06:58:17 localhost kernel: [ 399.315999] bbswitch: enabling discrete graphics
Jun 15 06:58:18 localhost bumblebeed[478]: The PCI config space could not be written fully at offset 0x1; 0 bytes have been written; error: Interrupted system call
Jun 15 06:58:18 localhost kernel: [ 399.705390] pci 0000:01:00.0: power state changed by ACPI to D0
Jun 15 06:58:18 localhost kernel: [ 399.705398] pci 0000:01:00.0: power state changed by ACPI to D0
Jun 15 06:58:18 localhost kernel: [ 399.705444] pci 0000:01:00.0: power state changed by ACPI to D0
Jun 15 06:58:18 localhost kernel: [ 399.705448] pci 0000:01:00.0: power state changed by ACPI to D0
Jun 15 06:58:18 localhost kernel: [ 399.705453] pci 0000:01:00.0: enabling device (0000 -> 0003)
Jun 15 06:58:18 localhost NetworkManager[488]: <warn> disconnected by the system bus.
Jun 15 06:58:18 localhost dhcpcd[568]: received SIGTERM, stopping
Jun 15 06:58:18 localhost dhcpcd[568]: wlan0: removing interface
Jun 15 06:58:18 localhost NetworkManager[488]: <info> (wlan0): DHCPv4 client pid 568 exited with status 1
Jun 15 06:58:18 localhost NetworkManager[488]: <info> (wlan0): device state change: activated -> failed (reason 'ip-config-expired') [100 120 6]
EDIT: Marked solved, latest package update seems to have fixed the problem
Last edited by D. (2012-06-16 07:22:53)
Offline
I'm also getting this same issue after the update to NetworkManager last night. I haven't had a chance to look at my logs yet, but I bet it's similar. I tried stopping the NetworkManager daemon before shutting down, but SIGTERM still fails anyways.
And It doesn't matter if I use halt or 'shutdown -h now', and it also happens when I reboot as well.
Last edited by SolarBoyMatt (2012-06-15 13:20:59)
Offline
I've confirmed that the issue is with the NM update; after downgrading networkmanager (0.9.4.0-5 -> 0.9.4.0-4) SIGTERM does not fail anymore.
Offline
Have you filed a bug report?
meh
Offline
I have the same symptoms - and have in /var/log/errors/log:
Jun 15 10:58:22 localhost NetworkManager[1441]:
nm_supplicant_info_set_call: assertion `call != NULL' failed
during the bootup time.
Mike C
Offline
Reported as FS#30305 at https://bugs.archlinux.org
Mike C
Offline
Absolutely superb - Jan Steffens is a super-hero - the bug is fixed and a new version of networkmanager is out - I already have the new package networkmanager 0.9.4.0-6 is already released to the mirrors. This certainly fixes the issue for me.
Mike C
Offline
Yep, issue seems to be resolved after the latest update. Marked solved.
Offline
Looks like it happens again with Networkmanager 0.9.6.2. Can anyone confirm that?
SIGTERM can not longer terminates networkmanager on shutdown
Offline
Looks like it happens again with Networkmanager 0.9.6.2. Can anyone confirm that?
SIGTERM can not longer terminates networkmanager on shutdown
Yep. that is what led me to this. I put /bin/bash in my /etc/rc.d/functions: kill_all() ala ( https://bbs.archlinux.org/viewtopic.php?id=129023 ) and NetworkManager and dhcpd were still running at reboot.
Maybe reopen bug?
Offline
I have also the bug for sigterm at shutdown and reboot but I don't use networkmanager,
so the bug is not related with networkmanager
Offline
I have also the bug for sigterm at shutdown and reboot but I don't use networkmanager,
so the bug is not related with networkmanager
Follow the instructions above to insert /bin/bash into the kill_all() function. Reboot and type "ps ax" to see what process is still running for you.
For me, it's dhcpd (tied to networkmanager) and network manager itself.
edit: W.T.Grammer?
Last edited by chisleu (2012-09-12 23:14:12)
Offline
Potomac wrote:I have also the bug for sigterm at shutdown and reboot but I don't use networkmanager,
so the bug is not related with networkmanager
Follow the instructions above to insert /bin/bash into the kill_all() function. Reboot and type "ps ax" to see what process is still running for you.
For me, it's dhcpd (tied to networkmanager) and network manager itself.
edit: W.T.Grammer?
Per instructions above (and others):
insert
/bin/bash
into
/etc/rc.d/functions
(made backup in case)
powerdown
after
sending SIGTERM to processes...FAIL
enter command
ps ax
If I read the file right it seems that it is the networkmanager that is causing this. I also examined my my log files and this error started directly after updating networkmanager. Everything points towards networkmanager, but one thing doesn't make sense. Downgrading networkmanager did NOT fix this for me either, so I am at a bit of a loss.
version:
networkmanager 0.9.6.0-1 (current)
networkmanager-0.9.4.0-6 (previous)
Anyone finding similar results?
EDIT: listed current/previous versions of networkmanager; Can't spell
Last edited by brewforce (2012-09-15 02:26:42)
Offline
Well I have found myself in the same problem as many of you, and diagnosed networkmanger package as source of problem.
Tried yesterday to solve a problem on a machine on which upgrade to newest version of networkmanager package was allready done.
Downgrading a networkmanager package (and network-manager-applet package) to previous version (networkmanager-0.9.4.0-6, network-manager-applet-0.9.4.1-1) wasn't the solution which worked for me,
but step-by-step downgrading fixed this problem with version 0.9.4.0-3, and setting networkmanager (and network-manager-applet) to IgnorePkg list in pacman.conf.
Maybe this could help someone for a temporary solution ?
I say temporary, because on a another (virtual) machine with networkmanager-0.9.4.0-6 which wasn't upgraded at that time, setting networkmanager network-manager-applet on IgnorePkg list in pacman.conf, and after updating the rest of the packages, SIGTERM error didn't show at all, and still using networkmanager-0.9.4.0-6 without a problem.
Would be really nice to hear someone if there is some other possible "glitch" with my solution?
I also have found one thing which puzzled me a bit. Somehow networkmanger confused my DHCP server which is on another machine, "eating up" most of my free IP addresses in the network range, allthough this machine has a IP reservation.
After releasing all those IP's from the DHCP list, downgrading packages and rebooting Arch machine everything was back to normal.
Maybe this info can help somone.
Last edited by prometej (2012-09-15 10:09:40)
UTERE SED NE ABUTERE !!
Offline
same issue here, using networkmanager 0.9.6.0-1
Offline
Having the same issue. Using 0.9.6.0-1 as well. This thread should be unmarked, maybe that way someone could help.
Offline
It seems as though the new networkmanager update (networkmanager 0.9.6.0-2) has resolved the issue. A big thanks to Jan Steffens
Offline