https://bugzilla.redhat.com/show_bug.cgi?id=1330598
But how (and why) start any service by systemd from top node can
affect /machine.slice/machine-lxc subtree ?
And what is special in start of disabled service by systemd ?
We have ~ 90 LXC HW nodes with total 400 containers.
I write now simple crontab script for periodically fix 8Ptb problem,
but it's so ugly way.
Upgrade version of libvirt is heavy way for some reason, and any
trick/tune in systemd would be
more desirable for old HW nodes.
b.r.
Maxim Kozin