lxc-checkpoint failed on current newest version

[Wei Tsui]

The lxc-checkpoint works well under the following version:

>> Linux Image Version: 4.4.0.12.13
>> CRIU Version: 2.0-2ubuntu1
>> LXC Version: 2.0.0~rc10-0ubuntu2

---------------------------

But failed after ungrading:

>> Linux Image Version: 4.4.0.15.16
>> CRIU Version: 2.0-2ubuntu1
>> LXC Version: 2.0.0~rc12-0ubuntu1

When doing lxc-checkpoint, it replies:

lxc-checkpoint: criu.c: do_dump: 854 dump signaled with 11

Chechpointing xxx failed.

[Tycho Andersen (tycho-s)]

Hi,

What version of CRIU are you using? (`dpkg -l | grep criu`) There were some bugs with ubuntu1, so try 2.0-2ubuntu2.

[Tycho Andersen (tycho-s)]

Can you please try the updated CRIU?

[Wei Tsui (ghostplant)]

Just new I tried the updated CRIU:

>> Linux Image Version: 4.4.0.15.16
>> CRIU Version: 2.0-2ubuntu2
>> LXC Version: 2.0.0~rc12-0ubuntu1

The checkpoint step became success, but recovery step failed:

+ lxc-checkpoint -s -D /var/lib/lxc/cr-test/cp -n cr-test

+ lxc-checkpoint -r -D /var/lib/lxc/cr-test/cp -n cr-test
lxc-checkpoint: criu.c: do_restore: 683 criu process exited 1, output:
lxc-checkpoint: criu.c: do_restore: 631 Bad file descriptor - F_GETFD failed

lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/systemd//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/cpu//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/freezer//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/perf_event//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/cpuset//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/pids//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/hugetlb//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/net_cls//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/blkio//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/memory//lxc/cr-test
lxc-checkpoint: cgfsng.c: recursive_destroy: 999 Error destroying /sys/fs/cgroup/devices//lxc/cr-test
lxc-checkpoint: criu.c: restore: 935 restore process died
Restoring cr-test failed.

[Tycho Andersen (tycho-s)]

On Thu, Mar 24, 2016 at 06:00:53PM -0000, Cui Wei wrote:
> Just new I tried the updated CRIU:
>
> >> Linux Image Version: 4.4.0.15.16
> >> CRIU Version: 2.0-2ubuntu2
> >> LXC Version: 2.0.0~rc12-0ubuntu1
>
>
> The checkpoint step became success, but recovery step failed:
>
> + lxc-checkpoint -s -D /var/lib/lxc/cr-test/cp -n cr-test
>
> + lxc-checkpoint -r -D /var/lib/lxc/cr-test/cp -n cr-test
> lxc-checkpoint: criu.c: do_restore: 683 criu process exited 1, output:
> lxc-checkpoint: criu.c: do_restore: 631 Bad file descriptor - F_GETFD failed

I believe this is fixed in rc13.

[Wei Tsui (ghostplant)]

Yes, it's fixed in lxc-rc13.