9d06a3c7e2
The nova_libvirt entry script nova_libvirt_launcher.sh starts libvirtd in the determenistic /system.slice/tripleo_nova_libvirt.service scope. That puts newly started qemu instances into the same cgroup scope: ... 2:pids:/system.slice/tripleo_nova_libvirt.service 1:name=systemd:/system.slice/tripleo_nova_libvirt.service Given the changed KillMode to control-cgroup, those must be started in another scope, in order to survive restarting of libvirt daemon or nova_libvirt container. Ensure that by starting libvirtd in a transient scope. Change-Id: Ib79bb77c62223cc2e54d92a19e02c1ad8cb8b12a Closes-bug: #1953503 Related: rhbz#2010135 Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com> |
||
---|---|---|
.. | ||
nova-libvirt-container-puppet.yaml |