Re: EOPNOTSUPP while trying to enable memory on cgroup.subtree_control

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Dec 07, 2023 at 05:37:06PM +0100, Michal Koutný wrote:
> Hello.
> 
> It does. Memory controller cannot be passed down to a threaded subtree (i.e. you
> only get memcg for the whole `docker` subtree) because memory controller
> is not threaded (e.g. two threads that share memory space but could be
> in two different cgroups make no sense for such a controller).
> 
> > Any clue? Is there some bug here or could be some problem with kernel
> > build/config?
> 
> And why your cgroup subtree is threaded? I suspect because of some
> threaded controllers were passed down [1] the tree while there are also some
> internal node processes (see cgroups(7) paragraph beginning with: "The
> second way of creating a threaded subtree").
> 
> To get out of this, your docker(?) needs to migrate all processes down
> from `docker` cgroup before enabling the memory controller.
> 

Hi!

Thanks for the reply and excuse me the delay, all that I investigated
seemed correctly so I was still thinking that there was something wrong
in the kernel....

But finally the problem is not docker, is openrc who creates initially
'/sys/fs/cgroup/docker'. Avoiding that both openrc and docker use the
same path fixes the problem.




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]     [Monitors]

  Powered by Linux