[lxc-devel] [PATCH] Add support for checkpoint and restore via CRIU

[Tycho Andersen]

On Wed, Aug 20, 2014 at 09:50:28AM -0500, Stéphane Graber wrote:
> On Wed, Aug 20, 2014 at 07:32:34AM -0500, Tycho Andersen wrote:
> > Hi Stéphane,
> > 
> > On Tue, Aug 19, 2014 at 10:40:15PM -0500, Stéphane Graber wrote:
> > > On Tue, Aug 19, 2014 at 10:14:03PM -0500, Tycho Andersen wrote:
> > > > This patch adds support for checkpointing and restoring containers via CRIU.
> > > > It adds two api calls, ->checkpoint and ->restore, which are wrappers around
> > > > the CRIU CLI. CRIU has an RPC API, but reasons for preferring exec() are
> > > > discussed in [1].
> > > > 
> > > > To checkpoint, users specify a directory to dump the container metadata (CRIU
> > > > dump files, plus some additional information about veth pairs and which
> > > > bridges they are attached to) into this directory. On restore, this
> > > > information is read out of the directory, a CRIU command line is constructed,
> > > > and CRIU is exec()d. CRIU uses the lxc-restore-net callback (which in turn
> > > > inspects the image directory with the NIC data) to properly restore the
> > > > network.
> > > > 
> > > > This will only work with the current git master of CRIU; anything as of
> > > > a152c843 should work. There is a known bug where containers which have been
> > > > restored cannot be checkpointed [2].
> > > > 
> > > > [1]: http://lists.openvz.org/pipermail/criu/2014-July/015117.html
> > > > [2]: http://lists.openvz.org/pipermail/criu/2014-August/015876.html
> > > > 
> > > 
> > > Thanks for that work!
> > > 
> > > Just one comment before I dive in a bit closer, I'd really rather we
> > > just have one command, say lxc-checkpoint which does both the dumping
> > > and the restarting rather than have the extremely confusing lxc-restart
> > > (which to most user will sound like an equivalent to c->reboot(c).
> > 
> > It is lxc-restore (vs. -restart), not sure if that changes your mind
> > at all.
> 
> Oops, should have read a bit more carefully :)
> 
> It is a bit less confusin indeed, though restore may be confusing with
> snapshot related functions maybe?

Yeah, I wondered about confusion between "snapshot" and "checkpoint"
since (without the context) they both mean essentially the same thing.
I don't mind renaming things, I just couldn't come up with a better
name :)

> > 
> > > Would you mind merging lxc-restart into lxc-checkpoint?
> > 
> > Sure, what should it be called? I guess the CLI would be something
> > like:
> > 
> > lxc-foo checkpoint [checkpoint opts]
> > lxc-foo restore [restore opts]
> 
> So, I think we would have something like:
> 
> lxc-checkpoint -n <container> [-D DIR] [-s] [-v] => checkpoint
> lxc-checkpoint -n <container> -r [-D DIR] [-v] [-d] [-F] => restore
> 
> So basically merging the two commands, then making the following options
> conflict:
>  - -r with -s
>  - ! -r with -d and -F

Sounds good, I will make the changes.

Tycho

> > 
> > ?
> > 
> > Tycho
> > _______________________________________________
> > lxc-devel mailing list
> > lxc-devel at lists.linuxcontainers.org
> > http://lists.linuxcontainers.org/listinfo/lxc-devel
> 
> -- 
> Stéphane Graber
> Ubuntu developer
> http://www.ubuntu.com



> _______________________________________________
> lxc-devel mailing list
> lxc-devel at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-devel