[CRIU] Issues with restoring multiple instances of the same source

Pavel Emelyanov xemul at parallels.com
Fri Sep 4 03:39:06 PDT 2015


On 09/04/2015 05:17 AM, Ruslan Kuprieiev wrote:
> By sending SIGCONT, I guess.

That would be too annoying. I would go with freezing cgroup instead. We've recently
had patches from Andrey merged using freeze cgroup for dump, I'm sure we can do the
same (or similar) to restoring stage. I.e. -- restore tasks into freezing cgroup,
put it into frozen state right before final wakeup. Later the whole cgroup can be
thawed.

> So, my first question is whether you would expect things to start breaking in this scenario (they seem to
> work a lot of times, again, we see occasional failures over time in the form of stream breakages possibly, 
> or just getting “stuck” (I believe it gets stuck waiting on a pipe though)). 

I would expect the container itself to start behaving strangely after restore, e.g.
due to networking being the same or similar. As far as CRIU itself is concerned --
if all the tasks you work with are namespace-d then nothing should go wrong.

-- Pavel



More information about the CRIU mailing list