[CRIU] [PATCH 1/2] Re-create cgroups if necessary

[Pavel Emelyanov]

On 06/26/2014 10:46 PM, Saied Kazemi wrote:
> Pavel,
> 
> Just a quick follow up:
> 
>> 1. cgroups sub-hierarchy
>> 2. configuration of cgroups
>> 3. populating cgroups with tasks
>> 4. creating cgroup mountpoints (if any)
> 
> Pertaining to the list above, one of our interns is working on #2 and will probably
> have a patch in a month.  Please let me know if someone is already working on this 
> to avoid duplication of effort.

Cool! From Parallels, only me was about to look into this, but since you have
started -- I will switch to other tasks :)

>> With a CLI option we tell CRIU:
>>
>> 1. Expect the cgroup to already exist, just put the process back in it.  If cgroup doesn't exist, fail.
>> 2. Expect the cgroup not to exist, create it and put the process in it.  If cgroup exists, fail.
> 
> I was wondering if you had a chance to think about this.  As explained before, there 
> are legitimate cases where the cgroups do already exist prior to criu restore, so their
> absence is a definite error.  Also, there are legitimate cases where the cgroups do
> not already exist prior to criu restore, so CRIU should create them.  CRIU on its own
> cannot determine which scenario it's dealing with.  We can make the default action to
> always create cgroups, but we still need a mechanism to tell it otherwise.

And, as Serge mentioned, there are cases when we would want to change the top-most
cgroup path on restore.

Right now I have this model in my head:

1. On dump get the paths to cgroups where the container's init process lives
   and dump all the other cgroup paths as relative to this one.

2. On restore introduce the --cgroup-root $path option that will prepend the $path
   to every cgroup being used -- for creation of them and for moving tasks into.

and taking into account the exist/not-exist issues

3. On restore introduce the --cgroup-create 'yes|no|may', where 'yes' allows to
   create missing cgroups and _don't_ use existing (abort restore if mkdir reports
   EEXISTS), 'no' denies creation of cgroups and abort the restore if smth is 
   missing and 'may' means create what is necessary or use existing if they exist.

What do you think about it? Serge, correct me if I've mixed up something in the 
items 1 and 2.

Thanks,
Pavel