[CRIU] how to lively migration a vm container?

[Bing X]

I don't understand, under my /etc/ there's no directory named as hosts ....

I install docker.io through apt-get.

On Tue, May 13, 2014 at 11:35 AM, Pavel Emelyanov <xemul at parallels.com>wrote:

> On 05/13/2014 07:32 PM, Bing X wrote:
> > As it's clear in the hierarchy, I have 3 docker containers running under
> docker.io
> > <http://docker.io>; for 2 of them, the ssh connected, one is bash
> (local container)..
> >
> > All of these images stored under /var/lib/docker/graph/
> >
> > but why I couldn't successfully dump any of them? Please help me take a
> look.
>
> Check the criu's log files, the dump errors are written into it.
>
> > thanks!
> >
> >
> > On Mon, May 12, 2014 at 7:37 PM, Bing X <xiebingbing at gmail.com <mailto:
> xiebingbing at gmail.com>> wrote:
> >
> >     I am running two docker containers with ssh service under /root. in
> one of containers,
> >     I run a IOR benchmark, which is live, and which is the one I want to
> dump.
> >
> >     what are the mountpoints visible by it? --- what's this mean?
>
> Linux allows to create so called mount points in the FS tree. Different
> processes
> may see different mount points. In your case the docker container sees
> that the
> /etc/hosts file is a mount point with external source. What is it?
>
> >     here's the pid hierarchical structure:
> >     2569 ?        Sl     0:12  \_ /usr/bin/docker.io <http://docker.io>
> -d
> >      5984 ?        Ss     0:00      \_ /bin/sh -c /usr/sbin/sshd -D
> >      6016 ?        S      0:00      |   \_ /usr/sbin/sshd -D
> >     12551 ?        Ss     0:02      |       \_ sshd: root at notty
> >     12563 ?        Ss     0:00      |           \_ bash -c cd
> /root/io_run; bash ./run.sh
> >     12564 ?        S      0:00      |               \_ bash ./run.sh
> >     14105 ?        S      0:00      |                   \_ mpirun ./IOR
> -w -z -e -f ./parameter.script
> >     14106 ?        S      0:00      |                       \_
> /usr/bin/hydra_pmi_proxy --control-port 280bce53d979:54710 --rmk user
> --launcher ssh --demux poll --pgid 0 --retries 10 --proxy-id 0
> >     14107 ?        S      0:00      |                           \_ ./IOR
> -w -z -e -f ./parameter.script
> >      9017 ?        Ss     0:00      \_ /bin/sh -c /usr/sbin/sshd -D
> >      9034 ?        S      0:00          \_ /usr/sbin/sshd -D
> >
> >
> >
> >
> >     On Mon, May 12, 2014 at 7:29 PM, Pavel Emelyanov <
> xemul at parallels.com <mailto:xemul at parallels.com>> wrote:
> >
> >         On 05/13/2014 03:13 AM, Bing X wrote:
> >         > What does it mean: I run criu dump with a pid I assume it is a
> container process, but get an error as:
> >         >
> >         > Error (mount.c:356): 119:./etc/hosts doesn't have a proper
> root mount
> >         > Error (cr-dump.c:1870): Dumping FAILED.
> >         >
> >         > what I should change?
> >
> >         The /etc/hosts is bind-mounted file and source of it sits out of
> >         the scope of what you dump.
> >
> >         Please, describe your setup -- what app do you dump, in what
> namespaces
> >         does it live and what are the mountpoints visible by it?
> >
> >         Thanks,
> >         Pavel
> >
> >
> >
> >
> >     --
> >     Thanks
> >     Bing
> >
> >
> >
> >
> > --
> > Thanks
> > Bing
>
>
>


-- 
Thanks
Bing
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvz.org/pipermail/criu/attachments/20140513/e1873961/attachment.html>