systemd (Was Re: tmpfs for strategic directories)

[Lennart Poettering]

On Tue, 01.06.10 05:53, Mike Fedyk (mfedyk at mikefedyk.com) wrote:

> On Tue, Jun 1, 2010 at 4:22 AM, Lennart Poettering <mzerqung at 0pointer.de> wrote:
> > On Tue, 01.06.10 01:56, Mike Fedyk (mfedyk at mikefedyk.com) wrote:
> >
> >> > KillMode=control-group → the entire cgroup is shot down
> >> > KillMode=process-group → only the process group of the process we forked is shot down
> >> > KillMode=process → only the process we forked is shot down
> >> > KillMode=none → nothing is shot down
> >>
> >> And what is the default?
> >
> > KillMode=control-group for native services and KillMode=process-group
> > for SysV services.
> >
> > Lennart
> >
> 
> Ok, the defaults look good.
> 
> What type of cgroup are you using?  Does it impede the use of lxc for
> containers?  Ie, the cgroup type that systemd needs to be able to be
> nested inside a container in a whole OS virtualization (think VPS /
> Virtual Private Server where each VPS has root in its container).
> 
> I currently use openvz and lxc (which is similar but based on cgroups)
> is getting close to feature pairity.  systemd shouldn't get in the way
> of being used within a container based on lxc...

As Dhaval already pointed out we use our own, named hierarchy,
independent of any controller. On top of that we use it recursively:
i.e. if you run systemd it will create the service groups beneath the
group it itself is running in. 

That means that systemd should not interfere with any other system
(unless you tell it to do so, via some config item), and is nicely
recursively stackable.

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net
http://0pointer.net/lennart/           GnuPG 0x1A015CC4