| Re: veth and checkpoint and restore [message #11152 is a reply to message #11148] |
Wed, 14 March 2007 12:31   |
Andrey Mirkin
Messages: 193
Registered: May 2006
|
Senior Member |
|
|
Hello!
Can you give an access to node?
If not, then try following:
1. Increase log level of CPT modules with commands:
echo 3 > /proc/sys/debug/cpt
echo 3 > /proc/sys/debug/rst
2. Checkpoint and restore VE in question
After that send me kernel logs and image file. Image file can be huge, so
please do not send it to list.
Regards,
Andrey
On Wednesday 14 March 2007 14:44 Nils Domrose wrote:
> I installed the 028-18 debian package mentioned below,
>
> same effect.
>
> ovz2:~# vzctl chkpnt 151 --dumpfile /var/tmp/dump.151
> Setting up checkpoint...
> suspend...
> dump...
> kill...
> VE is unmounted
> Checkpointing completed succesfully
> ovz2:~#
>
> ovz2:~# vzctl restore 151 --dumpfile /var/tmp/dump.151
> Restoring VE ...
> Starting VE ...
> VE is mounted
> undump...
> Setting CPU units: 112005
> Configure veth devices: veth151.0
> Adding interface veth151.0 and route 10.215.127.151 for VE151 to VE0
> Error: undump failed: Invalid argument
> Restoring failed:
> add ifaddr err -22 for 5 eth0
> rst_restore_net: -22
> VE start failed
> Stopping VE ...
> VE was stopped
> VE is unmounted
> ovz2:~# uname -a
> Linux ovz2 2.6.18-openvz-amd64 #1 SMP Mon Mar 12 21:59:13 UTC 2007
> x86_64 GNU/Linux
>
> On Mar 14, 2007, at 11:07 AM, Nils Domrose wrote:
> > Hi Andrey,
> >
> > vzctl version:
> > ovz1:~# vzctl --version
> > vzctl version 3.0.14-1dso31.2
> >
> > for the kernel this is more problematic:
> >
> > i noticed that i cannot provide you with the version since i
> > installed this kernel
> > http://download.openvz.org/kernel/debian/etch/linux-headers- 2.6.18-
> > openvz-amd64_02_amd64.deb
> >
> > cat /proc/version
> > Linux version 2.6.18-openvz-amd64 (02) (ola@deb-build-
> > amd64.openvz.org) (gcc version 4.1.2 20061115 (prerelease) (Debian
> > 4.1.1-21)) #1 SMP Fri Jan 26 18:14:59 UTC 2007
> >
> > problem is that the debian packet and file name does not indicate
> > the version of the ovz patch - so your question was already
> > helpfull since no know that I do not know which version I use ;-)
> >
> > Do you think it makes sense to upgrade to the latest version 028.18
> > - I would try out:
> >
> > http://download.openvz.org/kernel/debian/etch/linux-headers- 2.6.18-
> > openvz-amd64_028.18-2.6.18-11-1_amd64.deb
> >
> > Nils
> >
> > On Mar 13, 2007, at 5:21 PM, Andrey Mirkin wrote:
> >> Hello,
> >>
> >> Can you, please, provide us what kernel and vzctl do you use.
> >>
> >> Regards,
> >> Andrey
> >>
> >> On Tuesday 13 March 2007 18:58 Nils Domrose wrote:
> >>> Hi,
> >>>
> >>> I have a Problem that I cannot restore a VE (and also not migrate)
> >>>
> >>> ovz2:/etc/vz/conf# vzctl --verbose restore 151 --dumpfile /var/tmp/
> >>> dump.151
> >>> Restoring VE ...
> >>> Starting VE ...
> >>> VE is mounted
> >>> undump...
> >>> Setting CPU units: 112005
> >>> Configure veth devices: veth151.0
> >>> Adding interface veth151.0 and route 10.215.127.151 for VE151 to VE0
> >>> Error: undump failed: Invalid argument
> >>> Restoring failed:
> >>> add ifaddr err -22 for 5 eth0
> >>> rst_restore_net: -22
> >>> VE start failed
> >>> Stopping VE ...
> >>> VE was stopped
> >>> VE is unmounted
> >>>
> >>>
> >>> vzctl start/stop works nicely in combination with the vznet.conf
> >>> script.
> >>>
> >>>
> >>> If i run strace the last scripts (vznet.conf) and if for testing - I
> >>> create a 151.start script return exit 0. Never the less the restore
> >>> fails and I have no more clue why.
> >>> It looks like some script or program is still trying to restore the
> >>> networking for vnet interfaces - I have no IP_ADDRESS set in my
> >>> 151.conf. Since I only use veth devices.
> >>>
> >>>
> >>>
> >>> Nils
|
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
