[Solved] MIgrated Ubuntu VPS take 2 minutes to shut down [message #34036] |
Wed, 26 November 2008 17:07 |
tanteauguri
Messages: 8 Registered: November 2008
|
Junior Member |
|
|
I am migrating a lot of Ubuntu VE from Vserver to OpenVZ. I could solve all problems except one: the machines wont shut down. A vzctl VEID stop kills all daemons except the init process - after 2 Minutes it seemes the init will recieve a kill -9. A reboot inside the VE in not successful, init hangs forever.
In this forum I found, that it is likely that this behavior has to do with udev. Does someone know the difference to the templates, wich work fine? What changes do I have to make to shut down clean the init process of the VE?
Thx
[Updated on: Wed, 10 December 2008 15:37] Report message to a moderator
|
|
|
|
Re: MIgrated Ubuntu VPS take 2 minutes to shut down [message #34090 is a reply to message #34089] |
Mon, 01 December 2008 13:24 |
tanteauguri
Messages: 8 Registered: November 2008
|
Junior Member |
|
|
Thank you for the reply. I do not use a template. The templates are working. I draged a VE from vserver to openzv. Because they are about 300 servers I have no other chance. vzctl is able to shutdown the init process in the template VE's, so there must have been modifications made in the guest OS, which I was not able to find.
The versions:
Linux openvz01 2.6.24-21-openvz #1 SMP Wed Oct 22 02:50:53 UTC 2008 i686 GNU/Linux
root@openvz01:~# dpkg -l | grep vzctl
ii vzctl 3.0.22-5ubuntu1 server virtualization solution - control too
root@openvz01:~#
[Updated on: Mon, 01 December 2008 13:46] Report message to a moderator
|
|
|
|
Re: MIgrated Ubuntu VPS take 2 minutes to shut down [message #34097 is a reply to message #34093] |
Mon, 01 December 2008 17:54 |
tanteauguri
Messages: 8 Registered: November 2008
|
Junior Member |
|
|
This is what the mounts look like:
root@ppdev:/# mount
simfs on / type simfs (rw,relatime)
devpts on /dev/pts type devpts (rw,relatime)
tmpfs on /dev/shm type tmpfs (rw,relatime)
proc on /proc type proc (rw,relatime)
sysfs on /sys type sysfs (rw,relatime)
tmpfs on /var/run type tmpfs (rw,relatime)
tmpfs on /var/lock type tmpfs (rw,relatime)
I try to stop inside the VE ppdev:
root@ppdev:/# udevcontrol log_priority=debug
root@ppdev:/# halt
root@ppdev:/# got signal 15
exited from VE 103
All processes seem to be killed quick. From now on only one process ist running for ever:
~root@openvz01:~#vzlist 103
VEID NPROC STATUS IP_ADDR HOSTNAME
103 1 running 193.80.225.57 ppdev
root@openvz01:~#
It's the init process:
root 15617 0.0 0.0 1568 536 ? Ss Nov27 0:01 init [0]
Trying to kill:
root@openvz01:~# kill -9 15617
Now the VE is still running without any procs =8-|
root@openvz01:~# vzlist 103
VEID NPROC STATUS IP_ADDR HOSTNAME
103 - mounted 193.80.225.57 ppdev
Trying to stop from outside:
root@openvz01:~# vzctl stop 103
Unable to stop: VE is not running
There was no output in the syslog until I startet again
Dec 1 18:07:20 ppdev shutdown[11782]: shutting down for system halt
Dec 1 18:07:21 ppdev init: Switching to runlevel: 0
Dec 1 18:07:26 ppdev exiting on signal 15
Dec 1 18:19:44 ppdev syslogd 1.4.1#17ubuntu7.1: restart.
Dec 1 18:19:45 ppdev init: no more processes left in this runlevel
I also tried to unmount all bevor stopping the VE - no success
How can I trace when nothing ist running? I am not sure, if there is a possibility to trace init, because its the first porcess.
[Updated on: Mon, 01 December 2008 17:57] Report message to a moderator
|
|
|
|
|
|
|
|
|