I use unprivileged lxc containers in Arch Linux. Here are the basic system infos:

[chb@conventiont ~]$ uname -a
Linux conventiont 3.17.4-Chb #1 SMP PREEMPT Fri Nov 28 12:39:54 UTC 2014 x86_64 GNU/Linux

It's a custom/compiled kernel with user namespace enabled:

[chb@conventiont ~]$ lxc-checkconfig 
--- Namespaces ---
Namespaces: enabled
Utsname namespace: enabled
Ipc namespace: enabled
Pid namespace: enabled
User namespace: enabled
Network namespace: enabled
Multiple /dev/pts instances: enabled

--- Control groups ---
Cgroup: enabled
Cgroup clone_children flag: enabled
Cgroup device: enabled
Cgroup sched: enabled
Cgroup cpu account: enabled
Cgroup memory controller: enabled
Cgroup cpuset: enabled

--- Misc ---
Veth pair device: enabled
Macvlan: enabled
Vlan: enabled
File capabilities: enabled

Note : Before booting a new kernel, you can check its configuration
usage : CONFIG=/path/to/config /usr/bin/lxc-checkconfig

[chb@conventiont ~]$ systemctl --version
systemd 217
+PAM -AUDIT -SELINUX -IMA -APPARMOR +SMACK -SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID -ELFUTILS +KMOD +IDN 

Unfortunately, systemd does not play well with lxc currently. Especially setting up cgroups for a non-root user seems to be working not well or I am just to unfamiliar how to do this. lxc will only start a container in unprivileged mode when it can create the necessary cgroups in /sys/fs/cgroup/XXX/*. This however is not possible for lxc because systemd mounts the root cgroup hierarchy in /sys/fs/cgroup/*. A workaround seems to be to do the following:

for d in /sys/fs/cgroup/*; do
        f=$(basename $d)
        echo "looking at $f"
        if [ "$f" = "cpuset" ]; then
                echo 1 | sudo tee -a $d/cgroup.clone_children;
        elif [ "$f" = "memory" ]; then
                echo 1 | sudo tee -a $d/memory.use_hierarchy;
        fi
        sudo mkdir -p $d/$USER
        sudo chown -R $USER $d/$USER
        echo $$ > $d/$USER/tasks
done

This code creates the corresponding cgroup directories in the cgroup hierarchy for an unprivileged user. However, something which I don't understand happens. Before executing the aforementioned I will see this:

[chb@conventiont ~]$ cat /proc/self/cgroup 
8:blkio:/
7:net_cls:/
6:freezer:/
5:devices:/
4:memory:/
3:cpu,cpuacct:/
2:cpuset:/
1:name=systemd:/user.slice/user-1000.slice/session-c1.scope

After executing the aforementioned code I see in the shell I ran it in:

[chb@conventiont ~]$ cat /proc/self/cgroup 
8:blkio:/chb
7:net_cls:/chb
6:freezer:/chb
5:devices:/chb
4:memory:/chb
3:cpu,cpuacct:/chb
2:cpuset:/chb
1:name=systemd:/chb

But in any other shell I still see:

[chb@conventiont ~]$ cat /proc/self/cgroup 
8:blkio:/
7:net_cls:/
6:freezer:/
5:devices:/
4:memory:/
3:cpu,cpuacct:/
2:cpuset:/
1:name=systemd:/user.slice/user-1000.slice/session-c1.scope

Hence, I can start my unprivileged lxc container in the shell I executed the code mentioned above but not in any other. It only works this one time and not any consecutive time:

1. Can someone explain this behaviour?

2. Has someone found a better way to set up the required cgroups with a current version of systemd (>= 217)?