[CRIU] [PATCH 2/2] Make lacking cgroup properties non-fatal

[Pavel Emelyanov]

On 08/14/2014 08:37 PM, Garrison Bellack wrote:
> I am unfortunately leaving back for school at the end of this week so my ability to further contribute will be limited.

Thank you for the work done, hopefully we'll see you again :)

> If we want a failure in the cgroup property restore to stop the entire restoration, we need a patch further up
> the call chain. I agree the behavior should either be a complete stop of the restore or the behavior supplied
> by the above patch.
> 
> While it seems there are different cases where we want different behaviors when properties are missing on the 
> restore side, I think we are all in agreement that properties that are missing during dump should just be ignored.
> Would you like me to break this patch into 2 patches, handling missing property failure on the dump side and the
> restore side, so we can merge the dump patch?

Yes, please :)

> Getting this part merged is a priority for me as it blocking some of my colleagues.

Thanks,
Pavel