[Lxc-users] what's the difference in lxc-attach

[C Anthony Risinger]

On Jul 18, 2011 2:54 PM, "Serge E. Hallyn" <serge.hallyn at canonical.com>
wrote:
>
> Quoting Joerg Gollnick (code4lxc+list at wurzelbenutzer.de):
> > Hello Serge,
> > I think that the main point is the initial setup of the cgroup
(directory)
> > structure.
> >
> > systemd
> > tmpfs on /sys/fs/cgroup type tmpfs
(rw,nosuid,nodev,noexec,relatime,mode=755)
> > cgroup on /sys/fs/cgroup/systemd type cgroup
> >
(rw,nosuid,nodev,noexec,relatime,release_agent=/lib/systemd/systemd-cgroups-
> > agent,clone_children,name=systemd)
> > cgroup on /sys/fs/cgroup/cpuset type cgroup
> > (rw,nosuid,nodev,noexec,relatime,cpuset,clone_children)
> > cgroup on /sys/fs/cgroup/ns type cgroup
(rw,nosuid,nodev,noexec,relatime,ns)
> > cgroup on /sys/fs/cgroup/cpu type cgroup
> > (rw,nosuid,nodev,noexec,relatime,cpu,clone_children)
> > cgroup on /sys/fs/cgroup/cpuacct type cgroup
> > (rw,nosuid,nodev,noexec,relatime,cpuacct,clone_children)
> > cgroup on /sys/fs/cgroup/memory type cgroup
> > (rw,nosuid,nodev,noexec,relatime,memory,clone_children)
> > cgroup on /sys/fs/cgroup/devices type cgroup
> > (rw,nosuid,nodev,noexec,relatime,devices,clone_children)
> > cgroup on /sys/fs/cgroup/freezer type cgroup
> > (rw,nosuid,nodev,noexec,relatime,freezer,clone_children)
> > cgroup on /sys/fs/cgroup/net_cls type cgroup
> > (rw,nosuid,nodev,noexec,relatime,net_cls,clone_children)
> > cgroup on /sys/fs/cgroup/blkio type cgroup
> > (rw,nosuid,nodev,noexec,relatime,blkio,clone_children)
>
> That looks an awful lot like the default setup with cgroup-bin installed
on
> a ubuntu oneiric upstart system.  Actually, I see ns cgroup is mounted
> (separately).  If you can find a way to not have that mounted, that may
> solve the issue.
>
> I wonder if systemd actually uses ns cgroup (perhaps to lock consoles into
a
> cgroup)?

I don't believe it does -- not certain though.

IIRC from some of Lennert's writings/posts they only mount them all so they
are available?  Perhaps in case a unit needs it ... if another process were
to mount ahead of time systemd might not be able to fulfill the unit's
request properly.

By default I don't think systemd even uses any of the cgroups except the
`name` group for process tracking, but I remember some rants about the fact
that you can't really discover new groups easily, and a bare mount of cgroup
blindly mounts all subsystems.

C Anthony [mobile]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20110718/909b46ab/attachment.html>