[CRIU] Trying to Checkpoint LXC

[Nipun Arora]

Hi Andrew and Pavel,

I've tried setting both lxc.tty=0, and lxc.console=none.
I end up being unable to start the lxc container itself.

Just as a background for my testing environment.
I am using a ubuntu virtualbox launched using vagrant (I don't think this
should make any difference as it's essentially running ubuntu 12.04 in a VM)

The linux container is an ubuntu 12.04 precise container.

I've tried starting the container using the following configuration

cat lxc.conf
>>
lxc.tty=0
lxc.console=none
>>

trying to start lxc both as a daemon or otherwise does not seem to work..

lxc-start -n lxc_test1 -t ubuntu -f lxc.conf

or with the -d option for daemon.

I tried ssh'ing into the container to see if the lxc-start was successful
but that did not work either.
Could you tell me what I need to modify?

Thanks
Nipun

P.S - I guess you guys are in both lists - lxc, and criu so I'm sorry about
the spam :)


On Sun, Dec 1, 2013 at 12:21 PM, Andrew Vagin <avagin at parallels.com> wrote:

> On Sun, Dec 01, 2013 at 08:36:32AM +0400, Pavel Emelyanov wrote:
> > On 11/29/2013 11:39 PM, Nipun Arora wrote:
> > > Hi all,
> > >
> > > Thanks for the assist, I still seem to be having problems. Is there
> any other specific configuration flags I need to be aware of ?
> > > Even with the lxc.console = none in the config still seem to have some
> issues.
> >
> > Below is the list of mounts found in you container's mount namespace:
> >
> > (00.153292) [/](68->40)
> > (00.153295)  [/run](44->68)
> > (00.153298)   [/run/user](47->44)
> > (00.153301)   <--
> > (00.153303)   [/run/lock](45->44)
> > (00.153306)   <--
> > (00.153308)   [/run/shm](46->44)
> > (00.153311)   <--
> > (00.153313)  <--
> > (00.153315)  [/proc](69->68)
> > (00.153318)  <--
> > (00.153320)  [/sys](70->68)
> > (00.153323)   [/sys/fs/cgroup](43->70)
> > (00.153325)   <--
> > (00.153328)   [/sys/fs/fuse/connections](71->70)
> > (00.153330)   <--
> > (00.153333)   [/sys/kernel/debug](72->70)
> > (00.153335)   <--
> > (00.153338)   [/sys/kernel/security](73->70)
> > (00.153340)   <--
> > (00.153342)   [/sys/fs/pstore](74->70)
> > (00.153345)   <--
> > (00.153347)  <--
> > (00.153350)  [/dev/lxc/tty1](75->68)
> > (00.153352)  <--
> > (00.153355)  [/dev/lxc/tty2](76->68)
> > (00.153357)  <--
> > (00.153359)  [/dev/lxc/tty3](77->68)
> > (00.153362)  <--
> > (00.153364)  [/dev/lxc/tty4](78->68)
> > (00.153367)  <--
> > (00.153369)  [/dev/pts](41->68)
> > (00.153372)  <--
> > (00.153374)  [/dev/ptmx](42->68)
> > (00.153377)  <--
> > (00.153379) <--
> > (00.153387) Error (mount.c:254): 78:/dev/lxc/tty4 doesn't have a proper
> root mount
> > (00.153819) Error (namespaces.c:442): Namespaces dumping finished with
> error 65280
> >
> > The Error mount.c:254 says, that you still have bind mounted tty-s
> > in your container and the list of mounts says you do. Maybe there's
> > some other stuff in the container's config file you should tweak?
>
> lxc.tty=0
>
> >
> > Optionally you can unmount the /dev/lxc/tty* by hands before running
> criu dump.
> >
> > > Here is the dump.log attached...
> > >
> > > Thanks
> > > Nipun
> >
> > Thanks,
> > Pavel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvz.org/pipermail/criu/attachments/20131206/f8d8e94c/attachment.html>